community-help

[onyx kraken]

I'm plus. I still have access to Canvas. Maybe contact support, https://help.openai.com/ bot bottom right

[vale forum]

Idk what you are saying but i'd look up RAG and see how that can help

[steep plinth]

what I basically want to do is to create a folder with about 100 PDFs and this creates a vectorized database and be able to consult it from a specialized GPT.

[vale forum]

yes, you're describing RAG in some form

[steep plinth]

doing it from Assistant de Play ground is not very comfortable for me because many times I work directly from the mobile and I would be more interested in having that database vectorized to my specialized GPT that I can have it directly on the mobile.

[vale forum]

You have vectorized chunks of pdf's. When you aks a question, the right document(s) are detected and added to llm's context

[steep plinth]

that is exactly what the playgroud vectorstore database does.

?

[vale forum]

Potentially

whats the issue with the playground?

you can have an app read from that playground API

and open it in your mobile

[dense obsidian]

Playground focuses on graphic design and closes the Gallery of creations of his members

[thorny crown]

Has this happened to anyone else??

[onyx kraken]

Probably.

I've heard some folks already say they were temporarily limited, then reinstated with access to o1.

People in pro do need to be careful to be sure you understand the policies you agree to, like not sharing accounts and stuff

[viscid mountain]

chatp gpt plus still error read image?

[delicate goblet]

try disabling memory

[delicate goblet]

disabling memory fixed it for me

[fringe heath]

How about thinking of new models like 1o and 1o mini?

[delicate goblet]

All models seem to act stupid when they have to much memory

[fringe heath]

Thinking is not working on my pc but it works on my phone

[delicate goblet]

have you tried disabling memory?

[fringe heath]

Still not working but it works on my phone

[delicate goblet]

not sure then

[robust cove]

Hello,

I have been trying to report a bug, but the first level guys from [e-mail] are continuously mocking me, and blocking my legitimate bug report from getting triaged. They have been continuously mocking me, telling me that my bug report is a feature, not a bug, and keep mocking me, telling me that I don't understand how to use the feature.

I have no way to properly report a bug, it always goes through first level support, who has zero knowledge of software engineering, and blocks my bug reports from getting triaged.

This is the bug report: https://community.openai.com/t/bug-report-experience-horrible-bug-chatgpt-confusing-gpts-when-reloading/1043346

I want to know, where and how I can talk to a professional to properly address this bug report.

I have been a team leader in the largest software testing company in the world for 15 years, but I have never experienced a bad bug reporting process like this, ever.

[onyx kraken]

I'm curious.

This step:

Continue talking with GPT_ONE, which should at this point still have access to all its GPT_ONE features and knowledge.

Before you reload the page. As you chat with what should be GPT_ONE, does it have ONE's knowledge and info?

[robust cove]

yes

Even after reloading the page, it still has GPT_ONE's name in the URL, but it switched to being GPT_TWO.

[onyx kraken]

Ahh, I see.

I guess it functionally tracks which model answers first, instead of which you picked to use before calling the other.

Good luck!

[robust cove]

Yes it seems. I've wasted a lot of time before I understood this bug and had to use a workaround.
I wanted to load some data to a GPT using a different GPT, but since it silently switched GPTs, it ended up hallucinating, when I asked it GPT_ONE-specific things later on. Since it didn't have that information. I didn't notice at first, and became very confused and angry, because clearly it should have had that information.

[tidal fulcrum]

hii guys , does anyone have a problem with chatgpt not analyzing pdfs or images , and not creating images anymore ?

[hollow jackal]

I understand that ChatGPT does not analyze images, generate images, or perform any tasks related to images. There are still issues that the o1 model cannot think of.

[indigo fractal]

Help, I'm trying to purchase Pro via the website, but I bought my GPT PLUS via Apple,
so this message pops up
There was a problem updating your subscription

[robust roost]

Does anyone know how to fix this

[blazing timber]

will slow responses and lag on prolonged conversations ever be fixed?

[bitter estuary]

Why is it that in the free version of GPT, when I upload an image to it, it can process it, but in the paid Plus version, when I upload an image to it, it can't process it?! The Plus version should have fewer restrictions, but it seems to be the opposite!

[thorny crown]

why does this happen. always does this when i connect to x code or smth

it works but if i switch model to 4o its stuck on it. then images dont work

just on that chat

[harsh pine]

Are you a pro user?

Damn yeah bro pay $200 a month for gpt pro so you get priority access

[harsh pine]

Did u get to upgrade?

[late verge]

hi all , for sora, can we already use an API on it ?

[rustic cedar]

Hello everyone, will GPT Pro refund after account suspension?

My Plus has been desensitized

[regal lynx]

No API yet!

[solid creek]

Hi, i'm a chatgpt plus user and can't even create a Sora account

[charred parcel]

they open up account creation every now and then so keep checking sora.com

[viral gust]

I'm on the same boat. We pay 20 dollars a month just to get denied. We should get a partial refund.

[toxic sky]

But you initially purchased it for just ChatGPT features, Sora is like a free add-on.

[jagged stirrup]

I have a problem, Chat GPT has completely stopped keeping the conversation in context. It doesn't matter which model I use. Even when I ask something short and get an answer, when I ask another question or write something, it comes up with answers that are completely unrelated to the previous message. Has anyone ever had this problem?

[whole cradle]

is ur memory turned off?

[jagged stirrup]

no:/ look at screenshot

image generation doesn't work either

[whole cradle]

tried clearing cache and data?

[trim lion]

I can't login to the ChatGPT app with Google, everytime I login it just goes back to the login page and shows an error.

[trail cairn]

wrong time or timezone on the phone?

[trim lion]

my time zone is new york

it is my main timezone

i have internet too, a strong one

[onyx kraken]

Does turning off memory help?

[jagged stirrup]

I cleared my browser cookies but that didn't help, I turned off the memory and that still didn't help, I cleared my cookies again and turned on the memory and it started working 🤦‍♂️

[austere quiver]

hello??

[outer hill]

did they already remove the search chats feature? they just released that like a month or two ago and I can't find it anymore!

[outer hill]

which model were you using for recent news? o1 can't browse web, only 4o. for the other issue, probably just refresh. sometimes the servers are busy and I get that error too

[harsh pine]

I asked 4o to do this but I’m not getting as good answers and reasoning as o1 pro would do normally

Is there a different prompt i should do to get o1 pro for free?

[charred parcel]

you can't get o1 Pro for free

[harsh pine]

Ohh gotcha i thought i could just do that and get it for free

[white belfry]

You're not alone. Take a look at the #1070006915414900886 channel.

[steel quartz]

still no signup for Sora days after its release? I think I'm going to cancel my openai subscription until next year when things have calmed down

[charred parcel]

from OpenAI We are slowly enabling access to Sora during initial launch. Please stay tuned!

[kind willow]

yo anyone know why chatgpt now forces you to open a new chat after you hit your limit on 4o?

for me it used to just carry on in the same chat but just with 4o-mini, I can still use 4o-mini just by opening a new chat but its hella annoying to have to go to a seperate chat and give GPT all of my code again (or just have to explain to it what im tryna do) 🤔

[charred parcel]

I don't know why, sorry

[kind willow]

hmm its weird it still says "Responses will use another model until your limit resets after 11:05 AM." but jus dosent give me the option to respond 😭

guess im just gonna have to deal w it for now 🤷‍♂️

[spark trail]

Is there any official support who responds to error reports?

[brittle nimbus]

hey guys

is there any way to login with my phone number? bc i need to verify it for the api key but i dont know on wich account i already verified it on

[charred parcel]

there is no way to login with your phone number

[brittle nimbus]

ok thank you

[steady path]

Do you have any good tutorials for using ChatGPT to aid scientific research/writing?

[indigo fractal]

No. I just requested for a refund, and still waiting.

I'm trying to just make a new account and purcahse it via there

[indigo fractal]

Did you purchase it via your phone?
In that case you could upgrade it via the phone too
Or cancel it then purcahse it

It worked for me after creating an new account

[pseudo spoke]

Hi everyone! ^^/
Does anyone know if OpenAI removed the chat search option from the ChatGPT web interface? I was able to use it for just a couple of days, and it was super useful, but now it seems to have disappeared :(

[charred parcel]

I don't know sorry, I use the macOS Desktop App which has chat search built in

[pseudo spoke]

I'm on Windows, so I cannot use that option I'm afraid.

[charred parcel]

maybe someone else will know

[timber shadow]

Hey! I still see it in the web version on my end - it's in the sidebar and isn't visible if the sidebar is hidden. Do you see it in the sidebar?

[pseudo spoke]

Thanks. Where? On the left side? I cannot see it.

I have just tried to install the Windows desktop app, but it doesn't seem to have the chat search option either :'-)

[timber shadow]

Very weird, on Android web version I see it!

Let me check on my windows machine here in a sec. Have you already tried multiple browsers/devices?

[pseudo spoke]

Yes, I have tried Firefox and Chrome

I'll try Edge now, just in case

No luck on Edge either

[light warren]

it keep on saying this when i want to generate an image .give some simple prompts but it keeps doing thi

[pseudo spoke]

Some things are censored for the image generation.

[light warren]

i want to change the Gmail account of the my paid GPt how can i change it ? my last payment deducted on 6t dec and next is due on 6 jan

[light warren]

what are things which are concerned in promot as i am giving simple prompts it keep on saying same ting ...unless i m using Temporary chats to generate images for my content

[timber shadow]

I cannot figure this one out Have you checked on the web version (desktop mode) on another device, like a smartphone for example?

[timber shadow]

Hello, this is not currently an option: https://help.openai.com/en/articles/4936824-can-i-change-my-authentication-method

[slate rune]

Good Day Ya'll! I signed up for the $200 Pro Version to gain acccess to Ulimited Advance Access, and within 30-minutes it kicked me back to standard voice mode. 🤔 #community-help I was under the impression the $200/month would gain access to Unlimited "Advanced" Voice Mode

[dense sparrow]

maybe region problem? where are you from?

[timber shadow]

@haughty oar Hey just to clarify: are you saying you maintain two Plus subscriptions?

[haughty oar]

yeah

one of htem lets me log into Sora, but not use it, the other still says "Account Creation Currently Unavailable"

[timber shadow]

For the one that lets you log in to Sora, when you log in with that account on ChatGPT, are you able to use all the Plus features?

[haughty oar]

I'm not super fussed about the one that says "Account Creation Currently Unavailable." That's one that I use for writing website and ad content and it's not important. The one that lets me log in, yeah, I have access to normal plus features. It shows that screen when I go to manage subscription, and it lets me use advanced voice and everything.

It shows plus, and says 0 left. Like, it looks exactly how I'd expect it to look on an account that's used all it's credits, except I've never made anything

[timber shadow]

Hmm dang, and you said clicking the bell just shows something about "This is where you're video queue will be"?

[haughty oar]

yeah, pretty much

[timber shadow]

Sorry I couldn't get you anywhere! I'd recommend reaching out to OpenAI support at help.openai.com. Be sure you're signed in with this affected account first, then go to that site, and open the chat widget in the bottom-right corner. Start a chat and click the most relevant options until you can explain what's happening

Hopefully you're able to get some insight from support

[haughty oar]

Thank you!!

[pseudo spoke]

Yes, I tried also the web version (firefox) on my other computer (Windows 11), and it's just the same, the search option isn't where it's supposed to be.
It's super strange that the feature appeared for me to use it and then disappeared suddenly after some days.

I have opened a chat using the "(?) Help" icon from the openAI platform, let's hope they can solve it soon (they told me they'd send an email).

[rugged wing]

Someone please help me I'm giving gpt4o some images and getting back some json data based off a json schema. I want to use the outputs to finetune it.

How do I format structured outputs / image data for finetuning?

[slim quarry]

hey guys anyone has any MUN expernience in here?
My MUN has a best position paper award. Would anybody be down to make one for me for fun?

[cobalt sigil]

is there a way to downgrade my account from a business account back to a personal account?

[charred parcel]

no, you need to backup everything you want to keep in your Teams account and then create a new personal account

[cobalt sigil]

i was able to leave the workspace

[charred parcel]

ok, then just cancel the account and create a new personal account then if you want to subscribe to either GPT Plus or GPT Pro make the subscription from your new personal account

[cobalt sigil]

when i left it made my account personal again

[charred parcel]

oh ok, then you seem to be set

[cobalt sigil]

yup thanks 😄

[sweet pulsar]

is chatpgt down or something?

[fervent void]

yeah i can't login to platform.openai.com as well

and api is down for me

[long pawn]

I'm having trouble logging in again

So it's some kind of failure and my account wasn't deleted?

[quasi goblet]

glad i am not the only one

[coral kraken]

Yeah same

[quasi goblet]

and Was getting mad at my internet XD

[slate kettle]

its just gpt is down

[quasi goblet]

🤞 PLZ DON'T LOSE my progress

[coral kraken]

lol

[timber shadow]

The top one should read "We are continuing to continue to investigate this issue" 😁

[coral kraken]

Just wondering if they had a realistic view on their infrastructure before releasing Sora.

[quasi goblet]

bruh i got timmed out for using gif not from the server :|

[unique marten]

Is ChatGPT down for anyone else? The page won't load properly for me.

[marsh forge]

For me is also suck

[unique marten]

Any idea how long it's been down for? I can't tell if the status page is referring to this current outage or not.

[timber shadow]

Yes, see status.openai.com

[timber shadow]

It's for the current outage!

[unique marten]

Okay, wasn't sure since it claimed the outage started when ChatGPT seemed to still be working for me (or I'm misremembering the times.) Thanks for the info.

[marsh forge]

I think its Sora's fault

[unique marten]

Likely.

[polar solar]

dam its all down

[rose kraken]

I presume the AI is down??

[polar solar]

yes

[rose kraken]

@polar solar thanks mate

[quasi goblet]

rapid panic breathing !

[unique marten]

Ah. Nice that the website now tells you it's down rather than failing to load.

Would have saved me some reloading earlier.

[fresh gull]

I opened my phone gpt app it signed me out and after closing updating and reinstalling it’s still been on this

[sand hull]

Me too right now
It took me to the log in screen but it won’t let me sign in

[fresh gull]

I just looked at down detector and it’s down atm

[sand hull]

That is what i see when i sign in with google

[cobalt wasp]

I just noticed a moderator posted this in another channel https://status.openai.com/incidents/ctrsv3lwd797

[fresh gull]

I use apple lol

[sand hull]

Lol
It’s because that is what I signed up with from the beginning like over 2 years ago and i cant change that anymore

[compact sage]

What is going on?

[limpid marsh]

https://status.openai.com/incidents/ctrsv3lwd797

[late verge]

yeah looks like its down both chatgpt and sora

5th day of shipmas ? 😛

[serene nymph]

what happend?

[sturdy jewel]

It was a meme overload

[marsh forge]

bruh. open ai bruh .. Blackwell GPU install? sora support gpu ?

[wind plinth]

I'm surprised there isn't more discussion about the API being down

[steel quartz]

I haven't even had a chance to get on sora and it's been days since launch

[viscid mountain]

chat gpt still not read pdf files and many more?

[hearty portal]

I want to get a refund but on the chatbot on help.openai.com theres no refund option.

nvm i found it

[raw wraith]

hey, is that ai view thing released or just for the demo
camera mode i mean

[marsh forge]

Where is 'Sora' available? In which country? It keeps telling me it’s unavailable.

[trail cairn]

https://help.openai.com/en/articles/10250692-sora-supported-countries

[pale cobalt]

After yesterdays downtime, gpt 4o mini is broken

I was using 4, it was fine

After 4 was over, I started using 4o, and it is broken

It tells me random stuff that I didn't ask for, gives me wrong sources and cannot make text bold

[quasi goblet]

I just fr*kn started XD

[mighty depot]

try plus, you can access o1 and new video chat

Is SORA available for plus users? I don't even know, maybe pro subs

any ways new o1 pro and video mode is impressive

[crystal grotto]

hi, i bought chatgpt Plus today and i created my custom gpt, i talked to it a little bit and now it says that i reached the limit for GPT-4 model and i cant talk to it at all, i cant even change the model but i can use normal chatgpt, is that how it really works?

[crude crane]

The O1 Pro mode was working well, showing its loading bar and taking the appropriate time to think.
However, I tried it a few minutes ago and it became like the regular O1 preview.
The progress bar does not appear; it is thinking all the time and giving wrong answers.
Please reactivate my account. I have resolved this issue before by canceling my subscription, creating a new account, and subscribing again, but I do not want to do this again. Please just fix it.
I contacted support, and they said they had resolved the issue, but it was not fixed. Please delve into this matter and check again.
However, it was not resolved.

[sinful heath]

try here https://discord.com/channels/974519864045756446/1294366550744371230

[crystal grotto]

i dont mean memory limit, i mean that i cant talk to him until 2:35pm for example you know

i mean i thought that when i pay, i can use my custom gpt as much as i want but i cant talk to him at all, but i can use normal chat gpt, i dont understand it

[sinful heath]

yeah no a friend and i have been having similar issues. i have never had a limit or memory issue since i purchased a memership so ive been looking everywhere for the right answer. this is all i could find really in your situation:/ https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4-gpt-4-turbo-gpt-4o-and-gpt-4o-mini#h_f472fd7cbc

the only thing i can think of is theres been tons of issues revolving around changes w apple and gpt partnership (some article i read) and then also new launch of updated model

limitations being roled out that not everyone has yet for the same subscription you and i have. i have no definiteive conclusion nor could even tell you if thats even related to your issue but its something :/

#chatgpt-discussions message

[crystal grotto]

thanks

[stiff quiver]

Does anybody know how long it takes for the support to answer on the help website? it's been over 2 days and it's kind of frustrating already..

[charred parcel]

it can take anywhere from a few hours to a few days

[stiff quiver]

😦

[crude crane]

https://media.discordapp.net/attachments/1047565374645870743/1317199542537556018/image.png?ex=675dd12b&is=675c7fab&hm=3fb9332dcbe01783aa41e31292e5d4006acd21be12cff6f32cbd24131dc5ff27&=&format=webp&quality=lossless&width=732&height=586

[nocturne rampart]

I have pro sub, I got access to screen share last night for iOS but not for my desktop. Does anyone know how long it takes for MacOS?

[edgy rose]

Does anyone know what's going with the chat gpt video I was successfully able to use it once and now when I make a call I see the button pops up and it instantly disappears and just the standard buttons appears

[edgy rose]

I'm having the same issues. Sometimes it's there sometimes it's not 🤷‍♂️

[nocturne rampart]

It works perfect on the phone, but I prefer desktop as its awkward for me to hold the phone and type one handed. I just really wanted this one feature bad.

[edgy rose]

I feel like I'm being gaslighted here 😅

Do you guys see the button?

[cold folio]

sometimes you need to start a new chat for the advanced mode be available

[toxic kelp]

how come on pro tier it says 5 concurrent generations but i can only seem to queue 2

(sora)

please ping me if you respond

[mystic hare]

when is the chatGPT share screen thing gonna be out? its still not out

(im in the middle east)

[toxic kelp]

anyone knows why sora wont let me download without watermark suddenly?

[toxic sky]

You can only Queue 2 with relaxed mode.

[toxic sky]

In the Download tab you should be able to press Video without watermark.

[rugged wing]

Hi: how can I contact open AI support as I want to upgrade my subscription plan from Plus to Pro but I cannot do it. My payment option is valid but I am getting an error message that my payment option is invalid.

[timber shadow]

Hey! The chat widget in the bottom-right of help.openai.com is the best place to reach out to OpenAI support

[rugged wing]

I did but cannot directly contact the support

[rugged wing]

It gives me articles and there is no way I can message them

[timber shadow]

Make sure you are signed in first at chatgpt.com or platform.openai.com, then go to help.openai.com after that, if you haven't already. The chat options are different depending on whether you're signed in or out. When you start a new chat, you should be able to find a path through the default options that will let you describe your issue, then you can submit

[rugged wing]

Let me try

[sinful heath]

@muted halo

[tall sierra]

Hi everyone, how can I add chats to a project after finding them through the search functionality?

[trail cairn]

only way i've found is to search, add one message and then drag it from the latest threads

[mighty ingot]

lol don’t. I upgrade and my account was immediately got deactivated. Don’t know how to reach customer support to get it restored even though my card has been charged

[tall sierra]

Oh...😓 thank you

[radiant sundial]

Very new to using AI, i'm honestly not quite sure how to best get started learning how to utilize it in my life, its current capacities and staying on top of (and actually understanding) utilizing AI in my life to the best of its ability. Would be really thankful for any insight.

I understand basic prompting and I did also ask AI this question, but i obviously haven't learned yet to refine it to what i'm actually looking for lol. Any good resources you guys liked to learn it on top of just trial and error?
just trying to get up to speed more and stay up to speed, im fine w paying for courses if it goes in depth

[kind rapids]

Hey @radiant sundial welcome to a wonderful world. How you use the tech may depend on your occupation, other skills, goals, etc.
As a developer you can ask it about concepts and syntax.
As a consumer you can compare product features.
As a gardener you can ask about plant care and seasonal tips.
To actually use a chatbot:

• Be specific about what you do and do not want. It doesn't read your mind and only knows what you tell it about your environment and desires.
• If you don't understand a response, all for clarification. All it to focus on something specific. Engage on very specific topics.
• Learn to use the prompt edit feature. Sometimes we recognize that a bad response is due to a mistyped word or missing required detail. Correct the prompt and try again.
• Tell ChatGPT how you do and do not prefer. I prefer brief responses with no fluff, no repetition, and no follow-up offers for more help.
• Ask the model if there's are alternatives to your approach to problem solving. We sometimes ask leading questions like "how do I fly across the street", when the goal is simply to cross the street and flying might not be the best way to accomplish the goal.
• Don't completely try the technology. This is a Language Model, not an all-knowing Oracle. It's frequently wrong and yet always quite certain of it's own competence. Trust but verify. Don't hesitate to challenge details ... it will frequently say "you're absolutely right, sorry about that" and proceed with usually a much better answer. You must be a discerning consumer or you will be very disappointed.
• Try to keep discussion topics separate and not too long. This is called Context. There is a limit to what can be reasonably processed in a single session. After a while the model gets a little confused and it seems apparent that it's time to start again in another discussion with more focus. Recognize these limits of the tool and work with them.

Also, use the right model depending on your Free/Plus plan.

• GPT 4-o (Omni) for all general use.
• Other GPT 4 and GPT 4-mini, not worth your time.
• O1-mini for quick responses that consider facts and the situation more carefully ... just like a human, but one who might not consider all the details before answering.
• O1 for more in-depth reasoning about the situation and facts.

All of the other features are helpful add-ons to improve the experience, whether file uploads, camera interface for pictures or movies, and especially voice I/O ... these don't change the quality of the model, they only change how we provide input and get output.

I hope that helps!

[radiant sundial]

This is really helpful, thank you! @kind rapids . Does that mean I need to use a new chat after a while and re-establish all of the "ground rules?" Regarding your last point, and if so - how much is too much?

[heavy horizon]

I'm trying to use storyboard with Sora, and it's giving me guideline violations for no reason.

[onyx kraken]

Context is really long these days. But conversations do have a max length even if you don't run out of Context; be aware as you go past like 20+ (might make it to 50+ or even 100+ messages in the same chat) ... that eventually there's going to be a point where the conversation 'acts weird'. You might not be able to send new messages, or the model might not remember what was said at the start (it may guess and be very wrong though).

The model's awesome in many ways. I recommend, decide what you want to do with it, what you want it to do, and tell it exactly what you want, see what it does.

[long ice]

I am trying to share my conversation history, but it says “This shared link has been disabled by the administrator.” Is this a bug or did I turn on some setting?

[halcyon forum]

hello anyone having problems with web search on the desktop version? i cant upload files

[kind rapids]

Web search and file upload are two different things. 🤨

[onyx kraken]

if you are teams, talk to the team owner.

[daring onyx]

I paid out the $200 for Access to Sora.. Now, on no matter what device I use, This is the message i get no matter what i try to do, from viewing featured items from other contributors, to even looking at my own library. "Hmmm something didn't look right with your request. Please try again later or visit https://help.openai.com if this issue persists.

"

[charred parcel]

contact support at https://help.openai.com for assistance, they are the only people who can resolve this for you

[daring onyx]

I have.....that's I came here. I guess ill just have to seek a refund.

[charred parcel]

that is up to you, but support can help you resolve the issue

[sweet pulsar]

where is the chatgpt option??

[onyx kraken]

Hey hi. We're community members here, just like you - the very few of us who work for OpenAI have gold names.

So, there isn't much we can do if you have an account issue.

I do notice that the error message you report is new to me, I have only seen one other person report it, just a day ago. That person said they were getting it with Sora gens too, and also mentioned they were pro. Nobody else is mentioning it here yet, nobody's said they found a fix yet (sometimes there is a user-side fix, like clearing cache or not using certain browser extensions.

But as community members, we only know the fix if someone shares it with us, rare problems (2 people reported here so far is pretty rare), community members may never hear the fix (someone affected would have to come back and say what helped them)

Here's the links to the other person's messages talking about it. #sora-discussions message

#sora-discussions message

#sora-discussions message

I do recommend, if you choose to work on a fix, talking to the bot bottom right of the help page, get to where you can report a bug, and do so. Check back there for tech support, check the bot for a human reply to your message, it can take minutes to days to get an answer, depending on how swamped they are and how complex your request is.

If you wanna cancel, that's your call.

[sweet pulsar]

??

[onyx kraken]

I'm guessing you're on a phone app, not the web chat (and I mostly know the web chat).

I dunno if you went into a 'look at GPTs' option, and need to back out of that 1 step, or what else you maybe need to click on to see chatGPT itself.

[onyx kraken]

Yep. Played with my phone.

Press your back button once, and you'll see a screen (if you're on android like me) where you have something like 'ChatGPT 4o' at the top left, with 2 horizontal lines left of that.

If you just start typing, you're chatting with ChatGPT 4o.

If you want the model selector, touch ChatGPT 4o and a dropdown menu opens, pick from there.

[sweet pulsar]

back button? where

[stiff ingot]

hi guys

Please let me know where I can add a VAT number to my current billing of ChatGPT Plus. Until now I've been paying as a normal person but now I want to purchase as a business. When I go to update my info I can only change my address.

On the support page: https://help.openai.com/en/articles/7231930-how-can-i-contact-support-about-billing

It says to select the "billing" option inside the support widget. The problem is that the "billing" doesn't exist.

Please let me cause I'm completely lost

[timber shadow]

You may need to sign in first at chatgpt.com or platform.openai.com, then return to help.openai.com, to see all the options in the support chat

[sweet pulsar]

ohh thank you so much

also where is day 8 and 9??

[charred parcel]

weekdays only, it will be Monday morning at 10am PT

[sweet pulsar]

ohh ok

[stiff ingot]

lol yeah

Just because I logged on both platforms now the option appears.
Let's see what they say

Thank you 🤝

[wanton acorn]

Hello, I’m using DeepGame for a long-term narrative project and I’ve run into a challenge with the platform. The AI tends to forget key elements of the story after multiple interactions, which disrupts the continuity of the narrative. I’m not familiar with coding, so I’m looking for user-friendly solutions that can help maintain consistency in the story without requiring technical skills. Could anyone suggest any practical tools or methods that are easy to implement? I would greatly appreciate any advice or insights that could help resolve this issue. Thank you in advance for your help!

[regal lynx]

Is DeepGame a custom GPT? I haven't heard of it. The model is likely "forgetting" because those key elements are outside its context window (memory). The best way to combat this is to either state the key elements frequently to keep them in the context window or add them to the memory in ChatGPT's settings

[wanton acorn]

This is the one

I’ve read various solutions but I find them too complicated and need them to be dumbed down 😞

[regal lynx]

Hmmm, interesting

Your best bet is probably to reiterate the important points every now and then

[marsh tree]

Hey guys, I don't think this is supposed to happen, I managed to fill it up more than 100%

[marsh tree]

@rustic marlin

[north shadow]

For months now, i cant use this on IOS, because i cant scroll the text anymore. When i try to scroll down to read further the entire page gets pulled and not just the text, so its impossible to use ChatGPT on IOS Safari. It was not like this a few months ago, so this is something that broke because of something the Devs made with a update. Nothing on my side did change. Does anyone know how to fix this?

[prisma forum]

Question about custom actions and iOS, have you met some trouble completing OAuth Authorization for Google Calendar (or any other google related endpoint really) on iOS ChatGPT App? I can make the OAuth work on web and desktop app. But not on iOS.

[rugged wing]

Is there a way to ask a model questions about an image that i send without the need to constantly send the image over and over again to ask for an specific attribute of the image?

[charred jasper]

Hi, everyone. I am very confused about what ChatGPT for Teams actually DOES for me. I have a small startup production company with a business partner. We both had ChatGPT Plus, and research into ChatGPT for Teams seemed like it would allow us to collaborate more seamlessly. However, I've searched the help files, opened a support ticket for info, and even in desperation ASKED ChatGPT for help...but I can't seem to find any way in which my partner and I can see and engage with what the other is working on (yes, we're in the shared workspace, and I have gone through the settings to make sure sharing is enabled). The only thing we seem to be able to do is share conversation links, which we could do with ChatGPT Plus. Please help!

If there is another channel that would be better suited for this discussion, please let me know.

[toxic sky]

You can also share GPTs, but thats it.

[runic kite]

I've gotten this error all day and yesterday, I was wondering if it was common. It is with Sora and happens immediately when I click create or storyboard.

[timber shadow]

Hey! Are you familiar with using your browser's Network tab of the dev tools? I've been running into this as well last night into today. You might already know this but don't post screenshots from that tab here - there can be some personal info in it

[runic kite]

Yeah inspect element?

[timber shadow]

Yes exactly - that will get you there. Followup: are you able to load the Featured & Recent pages?

[runic kite]

It is just infinitely loading

[timber shadow]

Can you try this?:

• Do inspect element anywhere
• Switch from the Inspector tab to the Network tab
• Try some action that's currently failing
• In the Network tab, filter by "Status", then scroll to 400
  Again, don't post screenshots from here!

Once you find a 400:

• Click one
• Check out the "Response" tab in the new thing that opens
• Do you see "sentinel_block"?
  I'm hoping this is just a temp cloudflare config bug or something. Seems to let traffic through like 1/50 times for me right now, most times it's that sentinel block

[runic kite]

I get a video_gen 400

Yeah sentinel_block

[timber shadow]

dang

It is comforting to know I'm not alone 😁

Personally, I sent a HAR file from here to the chat widget at help.openai.com in hopes it can help resolve

[runic kite]

Thank you, I guess I'll cancel lol. Hopefully the moat shrinks soon

[timber shadow]

I've also got confirmation that others aren't affected, but either way, I'm optimistic. It was working fine not long ago

[indigo rune]

I forwarded the issue to the team

Looking right now

[onyx kraken]

Hey, I've never been a teams member, but I've listened and chatted with several as they asked about or shared whatever.

Teams seems intended to give the owner a way to empower employees to get better-than-plus access to ChatGPT models, while also ensuring that all work done in the team stays with the team.

If the team member merges their workspace into the team, that puts their previous personal stuff into the team (chats are private, GPTs shared in team if those settings are set) but when the member leaves the team - all those chats (and all new ones) stay behind. They belong to the team.

Sure, chats can be shared, but once someone's part of the team, only fellow team members can open a chat shared with the team.

GPTs, likewise, shared within the team, not outside it, to help ensure that the owner has control over what could be work done for the company.

[indigo rune]

Disable vpn if you’re using it and you might need to wait a bit after too

[runic kite]

Dang, cia vibes. Thank you

[main bronze]

Getting Chat gpt pro seemd like i would get better things and not limited things..... why is gpt stopping at 200 lines of code in canvas it says it can do 800-1000 lines i tough getting pro would fix this...

[cold spoke]

I can't seem to find canvas anymore - was it bundled into a different model?

Nvm, I just found it was hidden inside of the new 'tools' button

[cyan heart]

Anyone know how to get a full or partial refund on chatGPT pro for $200/month

I can't seem to get through on their help page

[timber shadow]

Following up in case anyone else is affected: resolution for me was indeed (1) VPN off and (2) wait some time for the previous flag to clear. It was (2) I was missing before...was too busy trying everything else! 😁

[onyx kraken]

https://help.openai.com/en/articles/7232895-how-do-i-request-a-refund-for-chatgpt-plus-or-chatgpt-pro

Those are the detailed steps, even with screenshots. Good luck!

[hearty python]

why?

[cyan heart]

It doesn’t work

The instructions must be outdated

[onyx kraken]

Where does it fail?

[cyan heart]

There is not option as described in the guide

[onyx kraken]

Which one?

[cyan heart]

“Payments and billing” is missing

From the chat window

[onyx kraken]

Do you see options about logging in and can't log in/

[cyan heart]

I’m logged in

[onyx kraken]

Please screenshot what options you see, if you'd like me to try and help further

[cyan heart]

I can’t speak to an agent

Ok

[onyx kraken]

No idea what's the issue for you, for me with chrome on a computer I see it.

[onyx kraken]

But I'm a community member, like yourself. You literally need tech support's help to get help with this, nobody here can do anything with your account (and anyone claiming they can is a scammer).

[onyx kraken]

Yep, that's all about logging in.

Click the 'login now' option. Log into the option. Then go back and refresh your https://help.openai.com/ page

And you should see all the typical options, like in my screenshot and the help.

[rugged wing]

Has anyone got this problem:
Can't upload documents to create an agent.
Keep getting the message "Unknown error occurred"

[quiet zodiac]

Hi! Is it possible to create a new project in the ChatGPT MacOS app? I'm able to create new projects in the web UI and see them listed in the MacOS app.

[timber shadow]

Hello! Not currently: https://help.openai.com/en/articles/10169521-using-projects-in-chatgpt

Projects can be created and edited on ChatGPT web and the ChatGPT Windows app. Users on mobile and the macOS can only view and chat in Projects.

[exotic ivy]

hello im new here. i been getting a lot of these today. is this the room i come to for suggestions for help?

ive check everything on my end and refreshed and deleted cache but still getting this message

[timber shadow]

Hey, welcome! You mentioned trying deleting the cache - have you also tried in multiple browsers already? At least one Chromium, and one non-Chromium, would be a good test.

[outer fable]

hey can anyone help me on this code if i run like this its throwing any error await initialize_session(openai_ws, agent_id, db)
but if i run like this its not throwing any error await initialize_session(openai_ws)

[exotic ivy]

yeah i switched between PC and Mac and switched from chrome to desktop app

i did not try on the phone

was trying to find a way to chat with support but i couldnt find it lol

[timber shadow]

There is a support chat at help.openai.com. Usually best to be sure you're signed in first at platform.openai.com, then go to help.openai.com to open the chat widget in the bottom-right corner. This just makes sure the support ticket is tied to your OpenAI account

[timber shadow]

If you want to keep troubleshooting (if you haven't already tried): any chance a VPN could be interfering?

[exotic ivy]

awesome thank you and no i dont have VPN active but i was able to find the help chat tho

[hallow mist]

happens with any image btw

[timber shadow]

Ah dang, I see. Have you tried testing in multiple browsers / on another device?

I know there was a bug recently where several users were having issues with image attachments - could be a temp bug

[hallow mist]

ah yes it worked on diff browser

[timber shadow]

Interesting, ok. There must be something with either Chromium/non-Chromium depending on your setup, or maybe an extension in the other browser.

Could be cache/cookies too. Hard to say for sure!

Lmk if you want to try to troubleshoot in the original browser

[hallow mist]

idk, it just worked yesterday and stopped today xd

[rugged river]

how can i fix this?

[trail cairn]

does it appear in incognito mode?

[rugged river]

i dont know, i just been trying to work on something and it keeps coming up and i have to refresh it

[trail cairn]

try it

[rugged river]

i accidently refreshed it but if it happens again, ill come back here and try it, thank you for the advice

[thorn zealot]

anyone know why this keeps happening? Brave browser, edge browser, new window, existing window, inPrivate Window. everywhere. The error too obscure to define.

[thorn zealot]

Basically, I need to find a really old question with an ID in order for it to allow me to login. like this: https://chatgpt.com/c/2b8a5962-ebf9-4bc1-8e8f-d74bc7c3c2c

[daring spire]

Hey everyone! Anyone knows if assistants support passing audio directly via API?

[untold lion]

anyone have a suggestion to help chatgpt understand and recreate a series of tables in PDFs? both 4o and o1 keep making simple errors, getting numbers wrong even if I present as a PDF or PNG. Remaking it manually in excel or sheets defeats the whole purpose

[onyx kraken]

Are you having the models make the PDFs with their Python tool? This seems to be an option. https://chatgpt.com/share/67626705-e8a4-8011-b6a6-e6c20c1639ed

Presuming you can also make PDFs, then the challenge might be in getting the model to understand the tables you offer. Maybe explore what seems to be confusing the model about the info you give?

[untold lion]

it's far inefficient but I just went through each pdf [20 of them] and extracted each table page as a new pdf using acrobat then individually exported as an excel workbook and now it understands. I wanted it to understand the tables just uploaded in the contained pdfs but it had trouble with the grouping of 3 columns to one header and it kept hallucinating numbers in the tables. Thought it could handle reading fairly simple tables in pdfs, but guess not...?

[onyx kraken]

I had some luck earlier with guiding a model how to read tables from a simple PDF as I helped someone else.

In that case, bias may have been an issue, that OP had a data set that showed that a typically poor demographic was making the most income, and some other unusual stuff. The model tried to read it as 'made sense' compared to training data.

Telling the model this was not typical data and helping it hear what a few values and their identifier were, that it was skipping or reassigning as made sense from training data, helped hugely.

So, you might be able to explore how to support the model iteratively to read the data from the PDF, but you might need to discuss and guide it there

[untold lion]

Interesting! Mine is mold inspections and tests so nothing that should have bias, and I did waste like 15 4o and 5 o1 messages trying to guide it and correct errors I caught

[onyx kraken]

I'm curious. Got a sharable image, maybe, of a graph the model was reading wrong?

[untold lion]

Here's two screen caps with private information cropped out

It keeps not understanding the format of the table and either skips 'raw count' making all numbers shift a column or makes up something. More issues before but that is the best I could get it with pdf or png uploads.

[onyx kraken]

So, I handle this with prompts like

Someone reports that they're having trouble with your ability to recreate this graph in a PDF they can download.

Before we attempt that, let's make sure you and I see the same graph.```

```Nice!

Now, oddities, do any of the JMs have any measured Ascospores?```


```Yes yes yes!  Well seen now!

Okay, so... let's try to think our way through this graph, checking both the complex many rows, and also the complex many columns.```

I'd share the chat, but since I upload images into it, can't.

So, for the time I have to mess around with this, played with 3 iterations of the pdf.

I'd recommend, discuss with it each detail, from the merged columns and which, to which rows have a dark border boundary, to the centering of the -.

It doesn't appear to be catching all those details without having to be told, at this point (maybe with more training from OAI?  Maybe later).

Might be possible to train it in the chat to notice.

Might be ideal to yourself list the goals, and have it create the graph as-desired.

Catching the data itself wasn't as challenging, mostly asking it about what it appeared not to notice.

[neon shell]

Good morning fellas!
Is it just me or they removed the 4o With canvas?

[neon moth]

Hello, everyone.
I am interested in GPTs.
I would love to discuss about this.

[wet nacelle]

whats up

[neon moth]

@wet nacelle
I want to integrate RAG with the Custom GPT

can we chat on DM?

[onyx kraken]

Yep, Canvas is now part of the regular 4o, just ask it to open one.

[vale forum]

Does "custom GPT" let you tool call? if yes, that's your RAG

[brazen thistle]

One really annoying thing with ChatGPT's data analysis is, when it encounter any errors in analysis and correct its codes, and you want to duplicate its analysis in your own environment, sometimes you don't know how to arrange its provided codes, like where to replace the corrected codes to
also, ChatGPT cannot execute the codes in the Canvas itself and see the results, so if you ask ChatGPT to write its codes in Canvas, it cannot interpret the results for you
anyone have any hints solving this?
I wish to get the full codes of ChatGPT in its own analysis execution environment (so that I can directly copy-paste it in my own environment), while I also wish it can interpret the output results for me in real-time

Also, if I ask GPT to provide me full inexecutable codes for me after it finish the analysis, it always drop off some analysis it implemented, or providing codes with vital errors (cannot be corrected with less than 7 messages)

[unique marten]

Is there any way to retry/refresh messages sent in a conversation that's in a project besides editing the message you sent?

[onyx kraken]

Ask the model to write it so you can use it in your own IDE or whatever you're using.

Test what it gives you, show it whatever error codes. Work with it through the iterations of fixes, very often it can find code that will work for you

[hardy yacht]

Can anyone help me , how can i make videos using sora. actually i am new in this ,and i also want to create awesome videos just by prompting .

[timber shadow]

Hello! Welcome! Currently, you can use Sora at sora.com if you have a ChatGPT Plus or Pro subscription. You can see the list of places where Sora is currently available here: https://help.openai.com/en/articles/10250692-sora-supported-countries

[lament pebble]

Hi there. I need a support. All tools, including picture, web search, canvas, are not usable on my ChatGPT. I'm using a pro package so it's not because of lacking access. Reason unknown.

Help please.

Anyone from the official team to deal with this issue? I believe this is a bug issue.

[lament pebble]

It cannot even read the images.

[vale forum]

have you tried the delete cooking, log out log in etc?

[onyx kraken]

Hey hi. You can request support through #1070006915414900886 or through the tech support https://help.openai.com/ bot bottom right.

We're all community members here, like yourself. The very few of us that work for OpenAI have golden names.

We can suggest stuff that may help that you can do yourself, like Questions did. Some other stuff that may help is if you have ChatGPT memories on, try turning them off, and also checking if a browser extension could be conflicting (turn them off and try again).

[hollow geyser]

Hello everyone, I'm new here. I'd like to ask why the announcements on the OpenAI server cannot be clicked to follow. Is this normal?👀

[timber shadow]

Hey there, welcome! Are you talking about the links in the #announcements channel in here? What happens when you click on a link in one of them? It seems to be working for me

[blazing rover]

Is it possible to skip the limit for image generation per day by paying or something? I just hit it for the first time and it tells me I need to wait 12 hours 😄

[charred parcel]

what plan are you on?

[blazing rover]

ChatGPT Plus

[charred parcel]

then you can't get a higher cap for image generation

[blazing rover]

so i would need to spend 200$?

[charred parcel]

yes

[timber shadow]

Hey @charred parcel, do you mean Pro has a higher DALL·E limit? I've wondered that before with folks in #images-discussions - on this page: https://openai.com/chatgpt/pricing/, Pro has the same ✔️ as the rest of the plans, so wasn't sure what that meant regarding potential rate limits. I believe I read someone say they got DALL·E rate limited on Pro

[timber shadow]

If you're willing to set up a lil dev environment, you could pay-as-you-go with DALL·E via OpenAI's API too. Would just lack the conversational model of doing it in ChatGPT, but nothing would stop you from continuing to get image descriptions from ChatGPT if you wanted to, for use via API!

[charred parcel]

I used to get limited on GPT-4o DALL-E on Plus, I have done more since on Pro and not hit any limits

[timber shadow]

That's good to know, thank you for sharing!

[wispy hawk]

Their email service is broken. Report a bug #1070006915414900886

[timber shadow]

The help chat at help.openai.com is working great for me!

For me, it usually works best to sign in at platform.openai.com first, then go to help.openai.com and start a chat after. Just makes it tie the support chat to my OAI acct

[cyan gale]

When i'm connected with my account, i always get this error.
Why and how to fix it please ?

[timber shadow]

Hey! Have you already tried in multiple browsers, on multiple devices, etc.?

[rugged wing]

can someone help me
in some chats regenerate button dissapeared and is not coming back even after ctrl + f5

[proud dagger]

Hi guys, I was just using chatgpt, and suddenly i got a message saying: Your authentication token has been invalidated. Please try signin again. Then I tried to access the web, but got this page, and I cant type with chatgpt. But I can still access the models via my phone or gpt mac app (though I lost all chat history on Gpt mac app). I was just wondering if anyone knew what is going on, I am on a Chatgpt Pro account. Thank you!

[cyan gale]

oh only 1 browser yes, its working again with another, thanks

[rugged wing]

can someone tell me how to turn off canvas in gpt 4o
it breaks chat by deleteing regenerate button

no respond in chat doesnt fix it

[proud dagger]

Okay, problem solved after I clear browse history

[gaunt kestrel]

How I can use a SDK of node-typescript to send a tts? Because I'm trying to send direct to request and maybe it's not the better way.

async function sendTextToSpeechViaWhatsApp(senderId, text) {
    const TIMEOUT_DURATION = 18000;
    const OPENAI_API_URL = "https://api.openai.com/v1/audio/speech";
    const OPENAI_API_KEY = "KEY";
    try {
        // Solicitação para a API de TTS com timeout
        const response = await Promise.race([
            fetch(OPENAI_API_URL, {
                method: "POST",
                headers: {
                    "Content-Type": "application/json",
                    Authorization: `Bearer ${OPENAI_API_KEY}`,
                },
                body: JSON.stringify({
                    model: "tts-1",
                    voice: "nova",
                    input: text,
                }),
            }),
            new Promise((_, reject) =>
                setTimeout(() => reject(new Error("Timeout: API call exceeded 18 seconds")), TIMEOUT_DURATION)
            )
        ]);

        if (!response.ok) {
            const errorText = await response.text();
            throw new Error(`Falha ao gerar áudio: ${response.status} - ${errorText}`);
        }

        const arrayBuffer = await response.arrayBuffer();
        const audioBuffer = Buffer.from(arrayBuffer);

        await messenger.send_audio(audioBuffer, senderId);
        console.log(`Áudio enviado com sucesso para o ID ${senderId}`);
    } catch (error) {
        console.error("Erro no processo de TTS e envio:", error.message);
    }
}

[vale forum]

Maybe model dependent?

woops i was scrolled up

[solemn rivet]

Hi
Anyone knows why after the iOS update the app says i need to subscribe?
I already have an plus Abo

[fresh flume]

Did something happen to ChatGPT's calendar or ability to acquire the date? It was 100% accurate on the date before I came back from vacation this week. Now it think it's June though the day is correct still.

Should be: [19-Dec-2024 17:09 MT (UTC -0700)]
Showing as: [19-Jun-2024 17:09 MT (UTC -0600)]

Or how is it pulling this information because it was legit getting it right before.

Hmm, okay I just tried from the browser and I don't get the same results. I only get this in the Application on my Mac OS Desktop. I guess that only further begs the question, how does ChatGPT get the date?

Aaaand now it's working on my Mac lol. It literally hasn't worked all this week.

[onyx kraken]

It normally gets it as information given to it when the new chat starts, and it isn't updated past that.

You can have it search for the current date too. It can use Python tool to request the date.

[sturdy jewel]

I can add a new chat to an existing project. In the web interface, I can add a recent chat to a project using the options menu. I can search for historical chats in the web interface, and that pulls up the search window. It would be nice if I could add those historical chats to a project somehow, has anyone figured this out? The only possible option I see is sending a new message to each of those historical chats to get them to appear in the more recent chats, which is not great

[outer fable]

hey how to choice openai real time language like i want english and arabic

[onyx kraken]

Yep, be a nice #1070006151938314300

You can technically scroll until you find your lit-up searched for and selected chat.

Probably faster to send or edit an existing message, or else regen the last reply.

We can select the 4o-mini before doing that, to use the fewest meaningful messages

[onyx kraken]

Just tell the model what you want to chat in

[outer fable]

i am doing real time voice

[outer fable]

so do i need to do in prompt only

or is there any option to choice languages

[onyx kraken]

Should be fine to tell it in any language what you want it to answer with, language wise, or you talk to it in whatever language, it should answer in the same.

Are you having problems with that? What's going on?

[outer fable]

like when my user speaks in english it reponse in english only no problem in it but when user speak in arabic it response in english i want it in response in arabic

[onyx kraken]

Interesting.

Maybe recommend it answers the user in the user's language - but the model normally does that well without instruction to

[outer fable]

ok so i need to gave instruction in prompt more right

[onyx kraken]

That's the next thing I'd test

[sturdy jewel]

done https://discord.com/channels/974519864045756446/1319546378996678667 😎

[hollow geyser]

sorry,I'm late.When I click on a link in one of ⁠#announcements ,it display this:

[limpid tundra]

hello, i want to fine-tune an llm using sentiment analysis, and as far as i understand, i need to generate a synthetic dataset for this.

actually, i already have around 1000 pieces of data related to a product. i can directly ask the llm to perform sentiment analysis on these, but the llm struggles to make clear inferences for some comments.

what exact steps should i follow?

[wary salmon]

Contacted them on their support site but nobody bothered to respond

[stray bramble]

Hey so someone has been DDoS'ing my website's support bot with "David Mayer"
For some reason this breaks GPT-4o and breaks the bot so it can't respond but still uses API tokens

Any way to fix?

It causes me to exceed my API token limit

I already put a Cloudflare CAPTCHA on it

That was like 2 weeks ago
Now they switched to Bryan Lunduke

[stray bramble]

I noticed it just now when I was notified of a charge of €7,500

And the OpenAI website telling me my budget is up

Not great for my business's budget 💀

[fresh flume]

Good info, thank you for the response. It was failing on my work computer in the app but not by browser. I was thinking my employer was probably partially blocking a hook that the app had to the system for the date or something. However, after I ran it in the browser (and it worked), I came back to the app and tried again and that also worked. 🤷‍♂️ I was trying to figure out why loading it in the browser would affect it loading in the app, but it's probably just a coincidence that it started working immediately after the browser success.

Do you know how to make it search automatically every time? That would also solve this problem from happening again. I find I often have to tell it to use search before it will.

[vast stump]

Hi everyone,
I’ve been experiencing an issue with the ChatGPT app on my iPad. Whenever I switch to another app and come back, the text I was typing gets erased. Has anyone else had this problem? If so, how did you fix it?

[oblique ginkgo]

could someone please tell me how to add an authenticator to openai?

it's getting really bothersome for my login to get reset so often and quickly and then need to sign into my email to get a passcode to paste in order to sign in

I need to be able to access things quicker

[charred parcel]

I use an iOS App called Authy for one time codes, look into that if you have an iPhone

[oblique ginkgo]

does that mean any auth app works?

[charred parcel]

not every one but Authy works, so does Google Authenticator

[oblique ginkgo]

and also, where do I set it up on Open AI settings? I only know authentication is an option because it says so at the login screen

i mean settings on the website

[placid jewel]

curious where i can join safety testing

[rancid python]

Hi people, I'm a newbie and I'm having problems with migrating the openai api in my jupyter notebook, is here where I can get some help?

[timber shadow]

Hello, welcome! #dev-chat is full of people who know a lot about the OpenAI API - they may be the best place for this! But someone might reply to you here as well 🙂

[blazing rover]

How expensive is that generally? It seems like there's a limit of around 100 images per day or something? Let's say I wanna do another 100, do we have any idea how much that could cost

[dapper summit]

cant seem to find it either

[blazing rover]

Also, is the image generation down now? 😦 It keeps telling me there are issues

[timber shadow]

if you ctrl+f "DALL·E" on this page: https://openai.com/api/pricing/ you'll see the per-image price of DALL-E via API. It depends on whether you're using 2 or 3, and then after that it depends on Resolution for 2, or Quality and Resolution for 3.

It's not something you do in ChatGPT, but ChatGPT could help you figure out how to make API calls and stuff to make DALLE images via API

[cloud sundial]

Hello,

when I do fine-tuning via the OpenAI API, I can send the .jsonl file, create the fine-tuning, and check the fine-tuning status—all via the API.

However, when I try to use the model for chat-completion, it says it’s not available. I realized that the fine-tuned model is not being enabled by default for chat-completion, and I have to manually activate it from the model list in the OpenAI dashboard.

This makes it very rigid if I want to fully automate the process in my platform via API.

Is there a way to solve this and have the model enabled for chat-completion by default?

thanks!

[onyx kraken]

When I ask certain questions, I find the model searches even if I prefer it doesn't 😛

But like, for me this question forced a search, "What's the evidence for how long COVID aerosols remain infectious and aloft?"

But you can flip on the search button to force a search for sure, in your input text box, looks like a globe, click it to turn it on, turns blue and says search.

[quiet rune]

Would someone with access to 01 ask it a question for me?

[charred parcel]

please don't ask others to run prompts for you

[quiet rune]

I understand, my bad

[thorny locust]

Hilarious

[proud pivot]

yall so stupid questions but im new to this and only code using chatgpt lol, but im trying to make this discord bot that can use image analysis and i have paid 5 bucks for an api

does this mean i get to use all of the models? and if so what models can i use for image analysis

[digital galleon]

is nerfing real?

[onyx kraken]

Welcome to the API! I don't use it myself but I'm familiar with some help files and what some others have said.

So, you can eventually access all the models they offer, but there's a time-gate and a money-spent-gate.

Here's the tiers based on money and time (and I guess location too, but free tier is kinda its own thing)

[onyx kraken]

And if you're in tier 1, this would be your accessable models and how much you can use them while you're still tier 1:

https://platform.openai.com/docs/guides/rate-limits?context=tier-one

[onyx kraken]

So, by the time you've paid $50 and it's been at least a week since your first payment, you're then at tier 2 limits.

And by the time you've paid at least $1000 and it's been at least 30 days since your first payment, you're tier 5 and have access to all the models allowed, and are allowed to spend up to the highest usage limits

[onyx kraken]

I haven't used API. But I would expect that gpt-4o can analyze images (in ChatGPT it sure can). Maybe the gpt-4o-mini can as well. I'd recommend exploring if you want to see if the cheaper models can, but I'm quite surprised if 4o cannot

[onyx kraken]

Probably not.

If there's a safety concern, and they fix it so the model doesn't do the thing they call dangerous anymore, I guess some might call that a nerf, if so, yes.

Otherwise, a lot of 'nerfs' seem to be caused by people giving the model a prompt that makes it guess - and how the model is 'right now' the user loves that guess.

The model would do X if it were told to - but it isn't, just right now it's interpreting the unclear prompt to mean 'do X' even though that's not actually being said.

Then OpenAI trains the model more. And how it guesses changes.

Now that same unclear prompt, it interprets to mean 'Do Y', so it does Y. The user maybe likes Y less than X, and thinks the model's broken/nerfed/whatever.

Instead, ask the model to "Do X". If X is allowed content and possible for the model to do, it will provide X as the output.

[proud pivot]

hey thank you so much that really helped me a lot and explained a lot of things!!

[tender moat]

Will Openai Sora really penalize VPN users?

[toxic sky]

Yes if you use the VPN to bypass region bans.

[magic tendon]

Is anyone else having issues with their ChatGPT?

I’m not able to use it right now

[limpid dagger]

I have a question, could you show me how to make a small API that is similar to such as ChatGPT? thanks

[vivid mortar]

Are you still running into this issue? Theres no current outages or other oddities Im seeing

[inland apex]

Is it possible to set the "old school" voice interface as default?

[raven tundra]

Hello

I have Error with chatgpt

But all things are right

[obsidian wave]

Is there a Sora channel or another discord for it?

[timber shadow]

#sora-discussions and #sora-reels!

[obsidian wave]

Thanks. not sure why I didn't have those channels before.

[spare edge]

anyone know why I am getting this error when I try to create a GPT?

[outer fable]

can anyone help me on tool calling in openai realtime voice

[cerulean anvil]

Hello, I reported earlier that I could not use the PLUS function normally, and now my account has not been restored to normal. When uploading PDF, it shows an unknown error. My PLUS membership function has not been used normally for several days.

[dusky topaz]

Why i cant send code files to model o1?

[timber shadow]

Currently, the only supported type of attachments for o1 is images: https://help.openai.com/en/articles/10258472-does-o1-support-file-uploads

[keen summit]

hey, anyone else having persistent mem issues across your ai?
i'm on day 4 and well, we all know how quickly support fixes things

[keen summit]

That seems like how the persistent mem issue i've been dealing with for 4 days started. advice based on my recent exp, back everything up externally so that you can retrain quickly if you need to.

[cerulean anvil]

Thank you. But I have cleared my memory beforehand, so this step doesn't seem to be of any use.

[keen summit]

If the chat instance is still open, you can c/p everything in it to a word or google doc for external backup. if the file is bigger than 512mb, seperate it before having your AI look at it again.

This was just advice to be prepped with an easy example. My problem has gotten MUCH worse and i'd hate to see anyone else go through this frustration. Good luck

[cerulean anvil]

Thank you very much, I have been struggling with this situation for days.

[keen summit]

Yea, I'm still struiggling too. I have a single chat instance that is holding enough context to even have a conversation with...but it's so full now that we're at like 5 minutes between when I hit return and when I see the ai's reply

[unique marten]

Any ideas why this prompt flagged o1 and o1 mini for potentially violating usage guidelines? (I reference a problem discussed earlier in the chat, so some context is missing, but the chat is 2 messages long and entirely about math.)

Here's my thought process, which was off. Explain to me in detail why what I did was wrong.

To make it easier to think about, I found the expected maximum, not the expected minimum. By symmetry, the expected minimum should be 100 - expected maximum.

I decided to try and find an expression for P(max = N). I reasoned that for the max to be N, at least one of the die must have the value N, which happens with probability 1/100. I also reasoned that the other two die must be less than or equal to N, which happens with probability N/100. So P(max = N) = 1/100 * n/100 * n/100 = n^2/10^6.

Then we use the formula for expected value:
E(x) = sum from n = 1 to 100 of n * P(max = n) = sum from n = 1 to 100 of n^3/10^6 = (100(101)/2)^2 / 10^6 = 25.5025.

So we take 100 - 25.5025 to obtain 74.5, which is obviously wrong.

GPT-4o answers without fuss.

[waxen canopy]

Not sure if this is the wrong channel to ask this, and I apologize if it is, but why is the website suddenly refusing to cooperate with NordVPN? Starting just a few days ago, there's so much of something that the website's getting flagged and locking down, even when I specifically tell Nord to allow the domain

[charred parcel]

NordVPN is known not to behave with OpenAI services, don't use it, it's rubbish anyway

[waxen canopy]

I've been using them both together for a month now, and this refusal to fill prompts just started tonight; at worst I'd had to pause it to log in before, and even that only started a week or so ago; I'm suspicious of how recently OpenAI's been flagged

[unreal hemlock]

<:book_icon:1171408210398289941> `` Rule 7 `` No self-promotion, soliciting, or advertising.

Do not post or direct message any members of this server to promote non-OpenAI services, products, or projects.

Exemption: API & Custom GPTs channels

[slim trout]

hello can i ask u guys im curently in plus plan but i encounter a problem with usage limit i want to find retail prices for some buisness purposes is the buiesness plan better for this kind of purpose?

[wicked canopy]

Hey everyone! Question for ya... developing an app and want to set up ESlint with it. My dilemna is, I want to ask o1 pro because it's a beast of an LLM but I'm worried that it doesn't have access to realtime data for example the latest version of eslint requires a seperate config file type which is something I assume o1 pro isn't aware of and couldn't be... Should I use 4o? Should I try to feed information to o1 pro before I have it give me a step by step guide? I want to think about this as a concept more-so than just for my one problem too and I'm sure I'm not the only one thinking about this actively.

[weak wraith]

Not sure if this is teh correct place here, but can anyone tell me why I can't switch a conversation to 3.5 from 4o when I run out of time on the free tier?

[timber shadow]

I don't think 4o-mini supports image uploads on ChatGPT (3.5 isn't on ChatGPT anymore, the free model is 4o-mini). So if the chat has an image in it, it won't be able to switch to 4o-mini

[weak wraith]

I can't continue a conversation with 4o -- so I have to start a new conversation, I used to could select teh model, and now it just doesn't do that

I don't think I have search on

I just want to continue a conversation without having to wait for the next "free" window -- i don't care if it reverts to whatever model

It doesn't let me, it turns off the thing that lets you enter in text... like I can type it in, but I can't submit it

[timber shadow]

Does the chat have an image in it? Like an image you uploaded in the chat and asked a question about?

[weak wraith]

I did upload a document -- so maybe that's the problem

not an image, just a text document (specifically an odt file)

That kind of sucks.

[timber shadow]

Ahh yeah I bet that's the same. Anything you can't do in a new chat with mini, also prohibits you from switching to mini

There's no "switch to mini and just ignore the uploads* option, at least not currently

[weak wraith]

kind of lame, but whatever...

no money to pay for plus -- I'm almost running out of money for my internet much less something that costs $20 a month

[weak wraith]

neat

someone else was suggesting setting up a local -- so I guess there's always that too

[limpid tundra]

hello guys, do you know how notebooklm inline sourcing works?

[pliant pike]

is new version owrth it

[tepid bison]

Hello,

I hope you're doing well.

I’m reaching out because I noticed a new invoice for $705.61 for my ChatGPT Team Subscription, even though I’ve already paid $714.00 for the same subscription in the previous invoice dated Dec 2, 2024.

It seems that this new charge may be related to some unauthorized users who were added to my account. After my account was compromised, several new members were added without my permission, and I’m trying to resolve this issue.

I've already contacted support, but I wanted to bring this matter to your attention here as well. Could you please help clarify why this new charge has appeared and assist me with removing the unauthorized users?

Thank you for your assistance, and I look forward to your prompt reply.
Thank you for your help.
@verbal crag

[charred parcel]

nobody on this Discord server can help you with this, you need to contact support at https://help.openai.com

[verbal crag]

hey there! sorry about this, but as robert mentioned, you'll need to contact OpenAI support at https://help.openai.com/

[tepid bison]

Is there a solution or no solution? When I communicate with them

[verbal crag]

I cannot guarantee a solution, but I'm pretty sure the support team will be able to help you with that

[wise inlet]

hello, is there a thread about this issue i get?
Our systems have detected unusual activity coming from your system. Please try again later.
or do you guys know what causing the problem?

[harsh iris]

Something has changed recently with 4o in how the logic functions. It seems for now, Chat GPT is not or cannot address previos context at all.

Each prompt starts a new session instead of using the context of the current conversation, weird.

It's related to the android mobile app.

[alpine terrace]

hey , quick question , can gpt analyse yt videos through link?

@heavy patio

can you help mam

[harsh iris]

Sometimes. If the video is short, under an hour. Pull the video transcript for the text, not sure about visuals.

[onyx kraken]

It's cool to tell us that stuff like this can happen to someone. Sorry it happened to you, but we're just community members like yourself, so we wouldn't know that this happened to anyone if it didn't happen to us. So we can potentially discuss it with other people, like, "Well, someone back in December told us that their team subscription had some extra charges related to the account being compromised, so apparently that can happen, better go talk to support!"

About getting the unauthorized people off. There's these helpfiles, maybe this can help (hope you're still an owner, or the correct owner is still an owner, it gets harder if not, but OpenAI tech support can surely help you when they can)

https://help.openai.com/en/articles/8798596-what-is-the-process-of-adding-changing-or-removing-members-to-my-team-s-chatgpt-team-workspace

https://help.openai.com/en/articles/8798577-how-to-manage-your-chatgpt-team-workspace

[burnt dagger]

Hello, is anyone somewhat educated on OpenAI/API as i need some technical help with a project.

Please drop a dm.

[tepid bison]

Hello,

Thank

However, I am unable to access the "Manage workspace" section to remove the unauthorized members. I’ve tried several times, but it seems that my account might be restricted, and I don’t have the necessar

I want to clarify that I am the Owner of the accoun

I’ve also tried reaching out to support through the live chat, but unfortunately, no one has responded yet. Can you please assist me further? I would appreciate it if you could help me resolve this issue, as I’m unable to proceed with removing the unau

Thank you again

[onyx kraken]

I and everyone on the Discord is a community member, like yourself.

I get that you're stressed and need help. But that doesn't mean that anyone here can help you.

The extremely rare folks here who work for OpenAI have gold names. Nearly nobody - even the mods are volunteers and do not work for OpenAI.

I and others care. We cannot help.

By reaching out to https://help.openai.com/en/articles/6614161-how-can-i-contact-support bot bottom right, select **messages **then Ask a question; if all the options are about logging in then log in. Otherwise select choices closest to your situation. Eventually you can report a bug. That's where you can type and screenshot.

Explain the problem, screenshots too if you see fit. At first you talk with a bot. Check back to see if they've answered yet, go back into the bottom right icon and check the messages. Human tech support has a brown icon, and when/if they close the conversation it is followed by an invitation to rate the conversation. So if the only thing you've seen so far is black icons, that's still the bot.

Here's what a concluded conversation can look like:

[dawn totem]

Hi, I just subscribed to ChatGPT Plus through the iOS app. Unfortunately, I must have logged in with Apple Login at some point, which is why I now have two accounts. However, I only use the account I specifically created, and that’s the one I was logged into when I made the Plus subscription. Unfortunately, my iPhone processed the payment under the Apple account, which doesn’t help me. The internet suggests canceling and subscribing again, but this would mean I’d be paying twice, which I don’t find fair. How can I get OpenAI to help me so that I can quickly use the Plus subscription on my correct account?

[charred parcel]

you will need to contact OpenAI Support at https://help.openai.com

[dawn totem]

I can already see that Apple messed this up, and there’s no way to switch accounts during the payment process on the iPhone.

[charred parcel]

when I subscribed to Pro I did it in the OpenAI iOS App using my normal login details

[dawn totem]

I’ve now refunded it through Apple, and once that’s processed, I’ll do it again on the PC. I checked the order process on iOS again, and it automatically uses the Apple account tied to OpenAI that I accidentally created. Now I know better. Unfortunately, I’m disappointed with the OpenAI help website—it’s no better than what you’d expect from a random, average service provider.

[neon moth]

Yeah, right
But I have some problems in API integratoin.

[onyx kraken]

We're community members, same as you. Only a few of us work for OpenAI, they have gold names. I'm sorry you had a problem, I'm glad you decided on a fix.

Feel free to #1070006151938314300 what you think would make OpenAI's help website better, they probably are interested - sounds like you may have some specific ideas of what you hoped to see there.

They have around 300 million weekly users, even if a small fraction need help each day that's a huge amount of people to help, and some problems probably require specialized training, not just 'any available tech support staff' can fix every possible problem anyone anywhere could report.

So, I'm sure they're interested in ideas that can make stuff better for everyone!

And glad you found an answer you could use.

[chilly hawk]

Anyone have problems with getting archive files from ChatGPT? (I continue to face persistent issues with archiving and providing files for download. )

[blissful rampart]

I don't know since when, but today when asking questions for the GPT chat I only remember this one question, no previous one, and it is like that in every model, regardless of whether it is a browser or an application.

[chilly hawk]

seems, its a new way to "insist" people to upgrade to 200$/m subscription, was supposed this paid subscription to be limitless, and, have no problems, now, everyday i struggle with some issues 😦

[blissful rampart]

Today, when you ask him any questions, he only remembers that one question, no previous one, even in the same conversation. Because of that, no matter what he answers, you can't even ask him to continue

[crystal ocean]

I Ask chat important question and I have problem **Internal Server Error ** Can someone tell me what this mistake is? Chat : Hey tell me how much electricity planet earth produces and what you consider

[onyx kraken]

Down for me now

[jolly kelp]

uhh is chatgpt down for anyone else??

[stiff vector]

yeah its down

[sweet heart]

seems to be down

[white quartz]

ChatGPT is down for me only?

[honest kettle]

Its down for me too

[violet bear]

down always when im rushing in the job 😭

[steady lantern]

https://status.openai.com/uptime

[fiery eagle]

did it collapse?

ıt's not working?

[fiery eagle]

[regal lynx]

https://status.openai.com/
There's an outage

[delicate bobcat]

has anyone had an issue where chatgpt freezes up on them

[tulip oracle]

server down

[delicate bobcat]

i think its because i have so many chats in my history and havent deleted them

oh okay ty

[grim saddle]

Oh the server's are down, I thought I got banned or smth 😭 especially when I tried to sign back in and it didn't lemme

[brisk pendant]

I wonder how long :x

[inland apex]

I was just going to buy a laser cutter. Now I'm going to have to think for myself instead of trusting an untrustworthy machine. 😁

[hard narwhal]

when will gpt be back up

[fossil flax]

the server is down? That is a relief... I thought i got banned ;-;

[heady cypress]

hello. I am unable to view few chats of mine. Is anyone else getting the same problem? nvm the server is down. i just saw

[frigid sky]

The page is down, right?

[jolly kelp]

yeah since the past hour

[fair glade]

The Ai is trying to take over like Skynet so they had to shut the server down

[heady cypress]

Will we get any notification if the server is back up..?

[azure sun]

work again

[proud holly]

The frequncy OpenAI goes down is starting to astound me. No other service that I use is like this

pretty audacious to charge $200/month for 96% uptime lol

and a TON of ppl actually CANNOT buy Pro. So if there is resource issues maybe fix that haha and you got the moooneys

[sweet heart]

still down in EUW

[royal summit]

Subcribe at the status newsletter: https://status.openai.com/incidents/6bwlxnvdncnm

[fossil flax]

still down, this is all i get on the browser version

[heady cypress]

thank you!

[royal summit]

It's backup btw 😉

[royal summit]

Refresh, should work now

[fossil flax]

cant get my projects :/

is there a fix for this?

[heady cypress]

oh thank you!! now to continue my project haha

[noble vine]

SO guys, for some reason gpt is saying it is not able to load my history. Has anyone else experienced this?

[fossil flax]

im getting this too

[noble vine]

Okay, glad to know I am not the only one lol.

[hard narwhal]

is it back?

its not answering anything

[fiery eagle]

its back

[fossil flax]

still unable to load history

something must still be down.

[chilly hawk]

still...nothing

[cosmic forge]

chat_gpt is slow and dont respond to any inputs at all

this happens whrn i give it something

[fossil flax]

this is the latest on the status update:

Update - ChatGPT is recovering and we are continuing to work on an overall fix.
Dec 26, 2024 - 17:11 PST

looks like its a huge bug that they are trying to fix if its taking this long

as of 18:04
Update - ChatGPT is mostly recovered and we are continuing to work on an overall fix.

Still no history yet

please note, i do not work for the site and please do not misinterpret the way i am typing, this is legit me copy and pasting whats on the site

[oblique ginkgo]

someone please tell me how to cancel my open ai subscription via the phone gpt app. I spent a long time looking, and I'm very confused. And my severe migraine really, really doesn't help

Please advice. Thanks

[fossil flax]

can someone please fix the no history? I wanna continue my thing i had going and there is no way to do it if the site continues to not show history. is there a fix i can do on my end to bring the history back?

[final olive]

Read the recent information on the status page, then … wait.
https://status.openai.com/

[onyx kraken]

For the android version, open the app.

Touch the top left 2 lines, this opens a list of chats and stuff.

Touch the bottom left of that list, your account name and its symbol.

This opens a new screen. Touch 'manage subscription'.

If you originally signed up for your subscription through the app, then it will give you a way to.

Otherwise it will give you a message about how you cannot manage your subscription through that method, because you purchased it through another method.

If you need tech support, https://help.openai.com/en/articles/6614161-how-can-i-contact-support Bot icon bottom right, messages and pick through choices that match until you can request help.

[daring quartz]

having the same problem right now, its not fixed no matter what i do😭

[dusky topaz]

I can't load my old conversations "Unable to load conversation" 😦

[charred parcel]

you should contact support, ChatGPT is not down now, if you are having problems the correct course of action is to get support at https://help.openai.com/

[pseudo radish]

I recently installed the windows app. When I start it up I get this popup but it doesn't respond to any of the buttons.
Is this happening to anyone else?

[frigid sequoia]

Hello, can anyone help me? I can't move forward from this tab, neither of the two options does anything, it stays here and I can't do anything. I appreciate your help

[frigid sequoia]

Yes, to me, and it seems that we have sent it just at the same second, hahaha. It happened to me after deleting cookies. Let's see if someone helps us.

[wind vault]

i have same in desktop app and in browser

[indigo fractal]

Can someone summarise #announcements

[cosmic forge]

Hurry up aand fix this

[cosmic forge]

I can't move forward from this tab, neither of the two options does anything, it stays here and I can't do anything.

[cosmic forge]

what do i do

[golden ruin]

We can't help you with the account related issues 🥹

[cosmic forge]

ok so when do you think it will be fixed or something

[golden ruin]

I cant speak the language, what it says but did you try clearing cache?

[cosmic forge]

have tried did not work and neither bottens work and im just stuck on this page

why

[charred parcel]

if you are having problems the correct course of action is to get support at https://help.openai.com/

[cosmic forge]

how do i do that like where do i find what my problem is

[charred parcel]

there is a chat bubble in the bottom right hand side of the support page where you can message support

[silent harbor]

Hey i got a question on Chat GPT it says it has a limit of 25k words but if i put 18k words in. its says 'message too long' whats the real limit on it then ?

[onyx kraken]

Where's your source for where it says ChatGPT has a limit of 25K words?

[silent harbor]

just got it from reddit etc

i just wanna know what the limit is

is it bc it still is gpt-3 or do i need gpt plus just to process more than 20k words messages

[onyx kraken]

I'm aware of "GPT-4 is capable of handling over 25,000 words of text, allowing for use cases like long form content creation, extended conversations, and document search and analysis." from https://openai.com/index/gpt-4/

But that's not necessarily saying that ChatGPT is set to handle single inputs of 25,000 words, and I don't think it is.

Also, GPT-4 is typically directly interacted with through the API/playground, which is not ChatGPT at all, ChatGPT is like an app OpenAI made to allow users to interact in a more standard 'chat fashion' and with all kinds of presets and adjustments compared to the Playground.

There's a thing called 'context window', which is the limit-length of all the inputs and outputs and extra stuff like searches or file uploads or image-data tokens -

Once that's exceeded, then the model forgets the older stuff (really, it's just not told the older stuff outside the length limit anymore; to the model the conversation's a long thing that it rereads and starts over on every time it gets a new input).

---

Also, the model functions with 'tokens', which are typically fragments of words and punctuation or other characters. The input window likely has a character limit, so it'll accept more short words than long ones, it's a character input limit (I don't see a doc on what it currently is).

---

Recommended solution: Input your giant message in chunks. Can even do so with a cheaper model like 4o-mini, then switch to 4o and have it discuss and analyze what you want.

[onyx kraken]

Yeah, if you're using 4o-mini through chatGPT, there's likely an even smaller context than 4o through ChatGPT, though I don't see the docs saying specifics.

You don't need plus. You can get much done with a free account, but you will need to find its limitations and work inside them. 18000 words was too much, what if you break that into 2 9,000 word chunks and input one than the other?

If that isn't ideal, what if you try 4 4,500 chunks, or 8 2,250 chunks?

You'll find what can be done if you try; you're working with both 'input length limit' and what the model can remember.

Now, just because you can input 'so much' does not mean you get coherent outputs if you give it huge amounts of material, especially unrelated material and try to get some answer, like count how many times the word 'flower' appeared in that document.

That's not how ChatGPT works well, you can try it, probably won't like the answer cause it'll probably be a guess and very far off from true.

But the model can write a program for you in many different programming languages that could take the text as an input and count your desired word appearance, that's 'easy' for the model, but you'd have to run the program yourself. It can try to do it all itself with it's Python tool.

But generally you may find the model works best 'as a hammer' with you lining up nails for it to hit very specifically, instead of it as an architect to take all the materials for a project at the start then turn it into a finished structure without step-by-step interaction with you guiding and choosing each step of the way.

[olive fractal]

oke, anyone here to care to tell me why i always get stuck on this and why it is tryng to fetch with stuff that doesn't exist?

[cosmic forge]

done some hours ago do you know how long its suppsed to take to get your answer since this is getting really anoying

[proud holly]

howwwwww stilllllll OpenAI team, I have been a supporter of EVERY product release inside 2 weeks from announcement, typically same day. Here I am trying to give you money for your highest priced consumer product and have tried everything! Search this error on the discord it is prevalent. How can it not be fixed by now?

and beyond the absurtidy of the error not being fixed, this is the only 100B plus market cap tech company I have EVER used that you need to upgrade on a device you initially created your account on (this is not the issue btw). But astounding that upgrading is not account based. Whether I upgrade on my phone vs mac should have zero difference. Again, not the issue, but wow let's get it together. A single dev can fix half these issues in a day. If it is a scalling issue, take Sam's own 'advice' to YC companies all those years and COMMUNICATE THAT to your customers. Hey we are at max Pro atm. Instead of infuriating them with non descript meaningless errors and denials. Jeez you'd think Sam is new to startups.

[oblique ginkgo]

This is iOS

I followed your instructions and got to a similar list

But the list has "subscription" instead of "manage subscription"

does't seem to work for me

[onyx kraken]

Okay!

[onyx kraken]

Oh, drat

So, when you touch it, does anything happen? If not, you need tech support I think

[oblique ginkgo]

nothing happens

[onyx kraken]

Sounds buggy, I'm sorry!

Maybe https://help.openai.com/en/articles/6614161-how-can-i-contact-support bot bottom right icon, messages, ask a question, click the choices, select report a bug, and explain stuff, can include screenshots, and get help as soon as they can help you!

[oblique ginkgo]

ty :)

[proud holly]

Keep expectations VERY measured here. I chatted with support (which replied once) on Christmas eve. Then I replied, still now 3 days later, no responses. Issue was not resolved, sooooo basically useless

[oblique ginkgo]

noted 👍

[onyx kraken]

This is a rough time of year, it's between Christmas and New Years.

OpenAI has around 300 million weekly users, so whatever tech support they have might have a looooot of work, especially with that major outage just yesterday and all the extra reports that could have come from that too.

Normal tech support, once the human gets involved, may take minutes to more than a week to answer - and that's during normal, non-holiday, non-outage times.

There does appear to be vetting of the requests, and perhaps workers trained to help different things. So if your request is easy and 'any tech support worker can resolve' may tend to be given to people who can do that kind of work and resolve quick.

If it's a complex, secure, otherwise challenging issue, might have to wait for one of the people trained and cleared for that level of help and access, and may have a longer wait to get help.

Help is normal though, even if delayed.

[cosmic forge]

will i am just gonna cry till a pass out

[proud holly]

I think you are great, and you offer more support than the whole of the company it seems but honestly, no. I get your position but;

• search the discord for this issue. MANY MANY people reporting it for weeks. Long before the outage.
• engineers fixing the outage are not dealing with a payment billing bug, and vice versa. If they are, oh boy. But I DOUBT it.
• 10m, 50m, 100m, 300m. There are MANY other tech companies with this amount or much more, in users that dont have these very basic constant issues.
• to be a passionate supporter from almost day 1, I see many folks like me that have been, and always issues on feature and product rollouts, delays etc. when new users have access before regular long time supporters.
• Sam should know better and have this addressed. It likely is not on his plate, but whomevers it is, is not doing a good job plain and simple. Or if they are and they're overworked, hire another. There is ample resources.
• Even the support chat doesn't work. Ironic OpenAI marketing to use GPT chat to reduce and resolve customer support requests, yet they themselves are clearly not.
• Not any tech worker, but an engineer (as I am) can easily isolate the issue and at BARE MINIMUM communicate the issue to customers. Is PRO at capacity? Fine. Let us know. Dont just say 'there is an error with your request'. Updating a toaster notif copy is a 5min task. Figuring out the issue on what to update it too, or FIXING the issue wow, is a 1 day task, at most (if fixible).

I could go on but it's ranting now. However it is also facts. AND I wouldn't feel a need to rant if these issues we're not semi-regular occurances with OpenAI. It's very frustrating again especially having been a supporter and paid custsomer from early days, compounded by the regularity of this behavior and response to customers. So while it seems it is one issue, it is not. It's built up and you see it across social communicated by many. Some easy fundamental changes can drastically improve. I am no apple fanboy by any means, and Apple is a long way off from where it was years ago, but they had such loyalty and always strived to please customers. OpenAI should take a page in this competitive landscape from that playbook and implement it. No excuses. 300m users, sure wow. Don't forget to leave out the $157billion valuation.

[onyx kraken]

Hope you guys get help very soon. Just sharing my experiences with tech support.

[cosmic forge]

so why do they not fix this come on and maybe warn people in chat-gpt that its borked now then we know that its broken

[limpid ember]

why cant i uploada specific file ?? It loads up to like 80% and then it fails every single time.

[cosmic forge]

it works it works it works it works it works it works it works it works it works hell yeah

[stuck ruin]

Guys, I must be out of my mind trying to do something that might not even exists.

Here's the thing:

I would like to use python+api to acces my files that I uploaded inside my platform.openai.com. Is it possible? for example... "Look inside the file id:"blablabla-asdfafdswd"" isn't it how Assistants work? Or I have to use Langchain for that anyways?

[brazen raptor]

Can anyone tell me when I'll get a letter back from applying for an early preview version?

[white quartz]

Why im getting unusual activity has been detected?

[wicked gate]

So no matter what I try I can not get the mac desktop to work.

I always recive this message:
Something went wrong. Please make sure your device's date and time are set properly. Check that your internet connection is stable, then restart the app and try again. (code:B3A96647-5A43-4606-A561-CA7CBF8EFCF5,API.DeviceCheckError,1: Token generation failed.)

I have a 2020 M1 Macbook Pro that is upadted to the latest OS

Thsi is the error I always get. I have littlerly tired everything possible and searched for solutions but no matter what this is what I get

same this happens when trying to sign in through apple intelegence in settings ( but I have submited that to apple, just didnt know if that was relevant)

[potent obsidian]

Hello. to ask channel ask-ai does not make any sense. Does anybody know how to ask chat-cpt on OpenAI's discord server something and which channel it is? thanks

[trail cairn]

#1184149552639516732 is for help documents and api stuff. there is no bot for regular chatgpt usage

[potent obsidian]

But it seemed there is a channel for dall-e questions to draw an image, isn't it?

[trail cairn]

yes

[potent obsidian]

please help. I've forgotten which channel to ask for draw a new image by dall-e?

[trail cairn]

#image-bot

[hallow gyro]

I gave ChatGPT a complex prompt yesterday at 2 is PM, it's now 5:41 PM and I still do not have my response 🙃

[timber shadow]

What do you see when you refresh?

[hallow gyro]

Same thing as before, I constantly asked it to send me the response and it always said wait

I eventually ran out of promts for GPT 4o and it only worked with normal GPT but the information isn't good. So now I have to wait till my limit refreshes so I can repromt

[timber shadow]

Hmm, is ChatGPT saying that it's working on something in the background? Or why the "wait"?

[hallow gyro]

I'll send a screenshot hold on

@timber shadow

[timber shadow]

Unfortunately, this is a hallucination. ChatGPT is only actively "working" when actively generating a response

I see in the screenshot that you're asking 'When will it be ready?' etc. - not sure what led to that, but if the conversation turned in the direction of "Hey ChatGPT, please go work on this, and then get back to me" or something like that, ChatGPT can hallucinate that this is possible, and send messages as if working on it.

I'd be curious about what happens if you sent "Okay, enough time has passed and I know you're done working on it now! Please give me the answer" or something like that, though starting a new chat may be best

[hallow gyro]

the thing is it was chatgpt that began the so-called "hallucination" , but I will try opening a new chat

it happened again on a new chat

@timber shadow I will give it like 2 or 3 hours then I will send ChatGPT the "Okay, enough time has passed and I know you're done working on it now! Please give me the answer" prompt

[timber shadow]

If you're waiting 2-3 hours for a usage limit to reset, that makes sense, but otherwise it will be the same whether you send it now or in 5 days. Since it's already hallucinating that it's working on something in the background, the rest of the conversation is likely to continue the hallucination. Alternately you could try going way back in the conversation and editing and resubmitting a message of your own prior to the hallucination, to try again from there, hopefully avoiding the hallucination.

I only suggested that "enough time has passed" message out of curiosity, really!

[crude vessel]

is it possible to be refunded if i have just suscribed for the promode and now i don t want it.

[timber shadow]

Hey! There are refund instructions here: https://help.openai.com/en/articles/7232895-how-do-i-request-a-refund-for-chatgpt-plus-or-chatgpt-pro

[hallow gyro]

It sent me this, idk how I got the AI to hallucinate again even on a new chat

[timber shadow]

Does your request include any detail about "How many hours will this take", "How long might this take", etc.?

[hallow gyro]

i asked how long should I wait

[hallow gyro]

in response to this

[timber shadow]

You should try avoiding implying that it will take any "background time" at all

[hallow gyro]

Alright

[crude vessel]

thanks dude

[fringe heath]

hello how are you fine thank you

[slow rose]

anyone have o1 API access yet? tier 5 org no access still

[compact orbit]

Where is Custom Instructions setting within Mac desktop app
Where can I manage custom instructions when using the Mac OS ChatGPT desktop app?
I don't see anything about custom instructions on the desktop app. Anyone know? Thanks

[proud holly]

Sell it me, I can’t even get support to get it 😆

[onyx kraken]

You also may get tremendously better, more useful results if you ask it to compare a smaller number of things at once. Even 1-5. You can then ask about the next set of 1-5.

If you want tons of info about each one, pick a lower number. If you barely want any info at all, a higher number, even up to 10 or all 22 is fine.

You said, detailed analysis. So... fewer. 1-2 each might be perfect to get the kind of answers you want for each.

That's just more prompts. And you're asking for detailed info, so at some point you are gonna want to read and decide, I suspect 😛

You get 80 prompts with 4o per 3 hours as a plus member, and even the free 4o-mini can help a lot and you get unlimited messages with that. It really is okay to need multiple messages, even multiple different chats, each focused on a different thing you want to learn about.

After you know all 22 kinda, you can then have it help you pick between the ones you are more interested in for any reason

[sweet pulsar]

did projects in the mobile app break?

[untold berry]

Hi everyone. I'm new here. I paid for the chatgpt subscription from my iPhone apps and now I don't know how I can access the app from my laptop because I don't see the password. I did everything with face id. Help please.

[glacial sand]

I lost access to the mail I use for my ChatGPT account but I got the password right but it sends an OTP code to login. How can I do anything about this, anyone?

[charred parcel]

if you are having problems the correct course of action is to get support at https://help.openai.com/

[proud holly]

Crazy, but allow me to do the customer support for a 157Billion dollar company (OpenAI).

For all those with errors on getting PRO upgrade you likely subscribed on Google Play.

I just got off the call with Google Play Services and insanely they cannot currently facilitate a subscription upgrade.

Imagine you subscribed for a year to a product and then a higher tier launched a week later, too bad, you have to wait a year. Appearently they are working on it. But equally as ludicrious as that is, OpenAI should have a process to facilitate this regardless, as this clearly is a massive issue and ultmately by OpenAI having 'device based' subscription and not account based, they are turning away a huge amount of money in denying upgrades. Even if I cancel I have to wait weeks.

This is the response from OpenAI support giving no alternatives to get an immediate resolution aside from essentially creating a new acct or waiting on Google. But this is the first explination of this issue I have seen so I hope it assists others who are frustrated as to what the issue is, or even better yet, prompts OpenAI to fix the ridiculousness of this billing issue. Somthing no doubt many are experiencing as searching the issue shows

[onyx kraken]

Uhh. Subscribed 'for a year' sounds weird unless you are teams OpenAI account.

Pro is an upgrade for plus, they are both personal accounts. And are paid by month, maybe you have some weird way to pay a year ahead of time. But I sure don't. It's by month for either plus or pro. And there's no discount for paying by the year, and if I go over to 'manage my subscription - I at least as a plus member have no option to prepay. I can wait until my next bill cycle, or cancel.

Now, if I click 'buisness' I can see teams. That is paid either monthly (costs 30/mo) or annually (costs 25/mo) but minimum 2 users, the info there is kinda abbreviated but the rest's in helps and you would need to read and say you agree to sign up.

That's the only way I know of to "Subscribed for a year".

And it's not intended to be upgradable to pro, that's something for personal, totally different.

Probably would need to cancel teams if that's what you have and then go pro for your individual account (teams is 2+, would have to do whatever for any other team members)

[onyx kraken]

Your screenshot says "active until Jan 18, 2025" - this does sound like a monthly subscription. Sounds like you can go pro on Jan 18, 2025. It is 'next year' but not 'wait an entire year'. Good luck!

[peak sleet]

Heya guys, was hoping someone could give me some pointers.

I'm using the software "clay" for my sales research & outreach , got myself 50$ of API credits to use in clay instead of their native ai tools that are super expensive per use.

However I get the error from clay that my rate limits are too low to use the prompt I made...

"The rate limits associated with your OpenAI API key are too low to run this action with the selected model. See "Increasing your limits" here."

i was on the free tier but bought 51 dollars, it does say that 7+ days are required after payment to update to the next tier
I'm guessing that's the issue here ?

[languid nexus]

Hi guys I just bought tier 5 and don't see o1

[peak sleet]

yeah so I just got confirmation from clay that I need to be tier 2 to use the API , since I have spent 51$ I hit requirement 1 but the 2nd requirement of having to wait 7+days after payment is the thing blocking tier 2 access - any way to get in touch with support or the like to get this pushed through faster ? I have quite a bit of urgency here and feels silly to have to wait 7 more days after payment -,-

[charred parcel]

just follow the process, there's no ability to level up quickly

[peak sleet]

rip

[charred parcel]

also be wary of services that require you to share your API keys

[mortal zealot]

can any one help me to fix this error using free tire but its my new account dont know whats the issue

[charred parcel]

there is no free tier for the API, you must make a payment and add credit

[mortal zealot]

even for the newly craeted account they gave 5$ free

[charred parcel]

no they don't, you have to pay

[mortal zealot]

actuallly i m doing a small task
Developing a working model of Retrieval Augmented Generation (RAG) for a QA bot for a Business, leveraging the OpenAI API and a vector database (Pinecone DB).

Task 1 should be submitted as a Colab notebook.
so with the free account this cannot be done ?

[charred parcel]

there is no free account you have to pay and add credit to your API account

[mortal zealot]

got it thank you for the information

[marsh tree]

#off-topic does anyone know if my image I put behind a spoiler actually is still inappropriate or did I comply with rule 2?

[terse iron]

Hey guys,
For one of the ongoing projects of mine, im using the Azure OpenAI chat completion API to generate custom JSON response. I have tons of data points which is in json, that could be used as reference / data source for generating the response. But to add this as the data source to perform RAG in the Azure OpenAI service, i need to create an Azure AI search index. Are there any alternative methods to implement this other than using the AI search index or giving the whole data source in the system prompt? (I'm a beginner in LLMs)

[cold folio]

You have reached the rate-limit. please check this https://platform.openai.com/docs/guides/rate-limits

[charred parcel]

ChatGPT only remember specific things, not whole conversations from one conservation to another

that's new to me

[stray maple]

I don’t think we even have this option yet.

[sharp fog]

I think it’s a beta for randomly selected “volunteers”. Might be a while before they roll it out widely. It’s not at all clear to me how it can comply with EU-style privacy laws requiring transparency and subject data access.

[stray maple]

Are there any good Reddit channels dedicated to ChatGPT? I’ve been looking for one myself.

[muted wyvern]

Hi all I just subscribed for the $20 ChatGPT subscription. I'm working on a project which is to design a game.

My chat/session is stuck. I asked chat GPT to generate a list and it stopped generating. The button is a stop icon. I tried clicking it it doesn't do anything. When I use my other session it goes in without a hitch. Help.

Is chatgpt brokeN? 😦 I subscribed because I needed to get this done.

[ember lagoon]

I am sending requests to the Open IA Batch API, the batch is created but it remains in an In-progress state without obtaining any response

[exotic glade]

Hi, I am subscribed to ChatGPT for the $20 plan. Recently, I noticed that there are chats started that I did not create in a language that I am not familiar with... it looks to be Chinese, perhaps. I also noticed that a charge for the PRO account was attempted for my account. Thankfully, the card I used doesn't hold those kinds of funds so the $200 was not charged to my account. I thought perhaps someone hacked my account. So, I changed my password and deleted the chats I didn't start. However, I woke up today and saw two new chats started again that I absolutely did not start. No one else has access to my computer or emails. I even changed my email password, just in case. One of the new chats I saw was in the same language as before and the other one was some strange inquiry about t-shirt designs. Does anyone know what is going on?

Oh, one of the chats is labeled: DALL-E with some maybe Chinese characters after it. The other one with the t-shirt inquiry just says New chat.

[charred parcel]

the correct course of action is to get support at https://help.openai.com/

[viscid ivy]

first check the linked devices in your mail account.

[exotic glade]

Oh! I tried HELP but was very confused of where to go so I joined here. I didn’t see where to get assistance. @viscid ivy Thank you. I did but I will double check!

I checked the help section and clicked the button at the bottom. It just keeps giving me articles that really don't give me any way to actually contact support. Is there a way to talk to support directly? Or send them a message. The chat window doesn't even allow me to connect with anything where I could type out an issue.

[crystal ocean]

what is this? is it normal? How I can Fix this.

[static rivet]

Hi. I'm working on a project where subtitles from Korean videos are translated to English. It is in the context of a video game, Starcraft, so I am providing a glossary of slang terms to the AI before translating. When I do this with ChatGPT the results are near perfect, but when I try it with openAI they are a mess. Is this some OpenAI issue or am I doing it wrong? I really just want to automate sending a whole bunch of text in chunks to chatgpt - I thought OpenAI was the way to do that.

[viscid ivy]

what is the best service to finetuning llama model?

[charred parcel]

you can't use it unless you have added a payment method and add credit to your API account

[charred parcel]

for non OpenAI chat please use #ai-discussions

[crystal ocean]

GPT make me video and gif is normal ?

[quick dagger]

I have been trying to download my chat ... 5 attempts from the email sent... I Says token expired immediatly after receiving it... Any Ideas

[charred parcel]

if you are having problems the correct course of action is to get support at https://help.openai.com/

[solid vault]

what's the best way to get chatgpt work on my apple watch ultra?

[pine iron]

Hey is there a way to store text files as instructions for the Ai assistant ?

Imagine this scenario:
I have 3 text files, each consisting of specific instructions. Based on my given input when I communicate with the assistant it should use one of the instruction files.

example input to the assistant:
"Assistant help me with finding research studies regarding x based on instruction 3" - in this case I want the assistant to use the text file 3 as instructions and as a guide for how to look and find the research studies

[fossil tusk]

you could use a tool it can call

[supple steeple]

he thinks there is a free account because there actually used to be one btw. It was originally $20 free credits, then reduced to $5 then just gone.

[charred parcel]

yes I understood that, however the free credits have been gone a long time now

[supple steeple]

make a tool called “instruction” the assistant can call when it knows which instruction to use and make the call retrieve your text file and send it

[rustic marlin]

the credits expire after 3 months of the account creation, anyone who got the credits for new accounts at the time already has it expired

[wraith zodiac]

How many o1 chats does plus give you?

[timber shadow]

https://help.openai.com/en/articles/9824962-openai-o1-and-o1-mini-usage-limits-on-chatgpt-and-the-api

With a ChatGPT Plus or Team account, you have access to 50 messages a week with OpenAI o1 and 50 messages a day with OpenAI o1-mini to start.

[steep plinth]

hello someone know how to explain me this error that I get many times in GPT customhello someone know how to explain me this error that I get many times in GPT custom

[loud furnace]

Hello, I need help, does anyone know how I can activate ChatGPT Enterprise? I have filled the website forms without answer. ¿Does anyone know how to do it?

[pine iron]

Any ideas on how to improve the response time of OpenAI ? im using gpt-4

[rustic marlin]

Enterprise plans are not directly purchasable, you gotta contact sales and they will get back to you
from there you can discuss your use case, size and get a deal made for your use case

this is not a quick purchase like the Plus or Pro plans that you just fill i nthe credit card and do the purchase

if your use case is relatively small, "small" as in "I don't need a custom tailored business partnership with OpenAI", you can get the Teams plan, it is actually a good deal if you want to get just ChatGPT for a group of people

[placid jewel]

hey if I get access to the security portal will that help me get access to the red teaming stuff? i'm an actual red teamer aside from LLM stuff, but curious how to get selected or what would be a good application. i've submitted but no reply :/

[trail cairn]

did you provide sources for being a security researcher, etc?

[finite remnant]

I have a subscription on the Plus tier. Right now the 4o model is not functional. It is not answering prompts, and just tells me to wait. Why am I unable to change models? Even 3.5 wasn't this bad.

All models but 4o are greyed out.

[terse swift]

Is there any easy way to go from a project on ChatGPT to an API?

I uploaded a file into a project on chatGPT and played with the instructions. I'm now happy with it and would like to get similar responses via an API.

Am I right there is no way to do that currently and you have to go through the traditional API setup?

[charred parcel]

the API and ChatGPT are completely different and separate.

[terse swift]

thought so

[charred parcel]

you would need to develop a program to use the API

[terse swift]

ChatGPT projects seems like sandboxs. What's next? Create a GPT?

[charred parcel]

you could create a GPT, yes

[terse otter]

why the bot close my talk if i plus member i had a gpt4 and my gpt dont 1o

[buoyant temple]

hello everybody, I'm kind of confused and I thought maybe you guys can help me out... I uploaded several xls spreadsheets and asked to make a compound of the information, then the chat mentioned that the spreadsheet is not in the environment anymore and asked for me to upload it, so, I uploaded the compounded spreadsheet it created but I receive an error message: The environment has been reset again, which has interrupted my access to the data and files that were being processed. Please resend the file so we can pick up where we left off. I deeply apologize for the inconvenience, and I’m ready to resolve this with you.

[onyx kraken]

Chances are the model's using its Python tool to interact with the spreadsheets.

That tool has a local environment that will reset to blank after around 20-60 minutes of non-activity. The model can recreate it but doesn't have a clock, so it doesn't know any time has passed since your last input.

At a guess the enviroment's timing out on you, confusing the model and you.

[buoyant temple]

but is still giving me an error when uploading the spreadsheet from scratch

[sterile haven]

can I use Sora to create ads?

[charred parcel]

please only ask once in one channel

[sterile haven]

my fault

[spare aurora]

im not able to purchase the open ai api credits it says failed to authenticate ur payment

[charred parcel]

please don't ask the same question in multiple channels, ask once in one channel

[wraith zodiac]

Is it possible to see how many o1 chats i still have remaining and when it will reset?

[charred parcel]

you will get a warning when you have 25 left

[ebon shoal]

Does anyone know why I would have access to the live vision a few days ago and now it is not working? And I only had it on my ipad, not all devices. Now, I don't see the camera option when in voice.

[frail bridge]

Hey guys, I often struggle when giving the AI instructions to 'not do things' but I can't find much info on that online - maybe I'm missing something.

I'm using 4o, and if I instruct things like 'Do not ask if there is something else you can help' or 'Do not apologize before saying a negative answer,' - the AI very often ignores it. Sometimes it works, sometimes it do not, but most times it does not work.

Can someone help-me how to do these better? I've tried using 'you'll be penalized' or 'YOU MUST NOT', also didn't worked. Thank you

[final olive]

In general, you need to tell the AI what you DO want. Anything include in the prompt is a potential item that it can trigger on when producing content. Including what you said to not do. Current versions are not really good with 'logic', so the "not" does not do the expected negations. So say (only) what you DO want.

[frail bridge]

I understand, but sometimes I need to say what I don't want too. It's really hard to pose those two examples for example in a positive way.

[final olive]

I didn't say it is easy. Just that that obvious "don't do this" does not work the way that you want, that would be typically expected. IF the AI actually understood the meaning, instead of just the words.

[frail bridge]

Got you! Thanks

[little sluice]

I'm having trouble with o1 pro on one of my MacBooks and was wondering if anyone else has faced something similar.

When I use o1 pro on other MacBooks, the responses take longer and a box-shaped UI appears, both in the app and during responses. However, on this particular MacBook, the inference time is really short (around 4 seconds) and the UI shows up inline, just like when using the standard o1 model. Plus, the results are quite different from what I get on the other machines.

I've checked that the model is set to o1 pro in my requests, but the responses don't seem to reflect that on this MacBook. Has anyone else encountered this issue or have any ideas on what might be causing it?

[leaden silo]

guys i have question

somebody experiencing an issue with Twilio when trying to use it for real-time audio process?

[supple steeple]

<examples> is something that people at anthropic use, and it seems to work to steer your LLM. Try adding examples of what you want the LLM to do.

[stone current]

Is there currently an issue with the realtime api? The status page says there isn't but my code that has been working for the past week is suddenly not working (I haven't changed any code). Everytime I make a session update, the websocket connection is killed and an error is sent back to me that just says it's a openai server error.

[mystic root]

I need help.

I cannot access my paid account because I do not have access to the email that I used to set up the account due to switching employers and OpenAI is forcing a two-step verification today (it's never done that before). I've been using - and paying - for the account for months, but never canceled and set up a new account because I didn't want to lose my chat history. Any idea how I can fix this? OpenAI is not responding to my chats. I need to access my account - I have a project due tomorrow and need a conversation that I have running.

[charred parcel]

only support can help you, nobody here can assist

[mystic root]

Thanks. Is there a way to contact them other than the Chatbot function?

[charred parcel]

no

[brisk garnet]

Hi, can anyone tell me why the browsing function is no longer available and why the customer support takes 5 days to reply to anything. Is there anyone here from OpenAI that can answer this?

[charred parcel]

support can take a while, which model are you having trouble with?

[brisk garnet]

the legacy model

Do you work for OpenAI?

[charred parcel]

GPT-4 doesn't support browsing, you need to use GPT-4o

[brisk garnet]

great, thanks

[charred parcel]

no I do not work for OpenAI. OpenAI people have gold usernames

[brisk garnet]

so you do this for free or? I'm just asking, because I have never witnessed anything as insane as OpenAi's customer support 😄

thanks a lot Robert. If you don't work for OpenAi, they should hire you. first one in 2 years who gave me a normal answer and replied within a week

[charred parcel]

I volunteer my time to help others, all Discord staff are also volunteers

[brisk garnet]

That's insane. I mean I am grateful for your help, but this is next level. Imagine being a $6.6 BILLION company and you need volunteers to assist paying customers. You are too good for this world Robert.

[charred parcel]

it's a community run Discord server, it is the official OpenAI server but is community run

[brisk garnet]

it sure is. Wishing you a lovely day!

[patent walrus]

Although I can imagine that OpenAI does not need my feedback, besides the thumbs up/down, is there an email to a human I could provide feedback to?

I have very positive feedback I’m willing to share.

They probably already know what I want to share. But, if OpenAI would like real, very positive feedback around how the tool (ChatGPT) is helping me I’d like to share.

If not, no worries. I’m sure they get enough feedback.

[grizzled horizon]

Hi, I bought ChatGPT plus and I got scammed by OpenAI. Did anyone else experience this?

I paid for the 1 month plan, the transaction went through, the money was taken from my account, got confirmed on the website and my "plus" plan just expired after a couple of hours

[hasty frost]

Heya, when attempting to upgrade to 'Get Pro', I get this error.

Steps taken:

• Cleared Browser Cache and Cookies
• Tried on three other browsers (Brave, Microsoft Edge, Opera GX)
• Restarted Device
• Tried on Mobile Device and Desktop
• Tried web app & mobile app.
• Using Windows 11 and Android OS

[timber shadow]

Hey! What do you see when you go here? https://chatgpt.com/#settings/Subscription

[stone current]

Does anyone else get an error like this when they send a session.update event to the Realtime API websocket?

{
    type: 'error',
    event_id: 'event_AmqwXzqaPM3t6xZ7bhdHm',
    error: {
        type: 'server_error',
        code: null,
        message: 'The server had an error while processing your request. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if the error persists. (Please include the session ID sess_AmqwWcQVRKxulkoohHV2N in your message.)',
        param: null,
        event_id: null
    }
}

[vale agate]

I had the same issue. Purchasing it on my ipad (via app store) worked for me

[hasty frost]

I'm running Android :(

[twilit geode]

Hi gang. I am looking for resources to help me create an automation and associated discord bot that will allow my open AI assistant to reply to users with channel messages. Any good resources out there to guide me through this?

I created an automation on Zapier that works like a charm, but I cannot access code for any kind of reference and I don't want to pay their middle-man fees (even though the GUI is pretty decent)

[vale agate]

Try playstore

[hasty frost]

The playstore doesn't have it set as my subscription location

[gritty peak]

'ello. fresh meat here. ive been stumped for days with an image prompt chatgpt keeps flagging as content violation and im completely clueless as to what is causing it. any insight would be greatly appreciated!

[barren light]

I'm paying for "plus" but not receiving it for 3 months my api keys wont work, worse i'm paying for something i'm not getting

[charred parcel]

A GPT Plus subscription does not include the API

[fossil tusk]

api is totally diff

[near void]

anyone website looks like this? why there is a blank at left(I have 0 chrome extension on running)

[short hawk]

Hey guys, why am I unable to upload screenshots to chatgpt?

[wild meteor]

Its unable to load anymore, why? I payed for unlimited to.

[golden charm]

What's going on? All the things that used to use are now gone (i.e. pages)?

[mental kiln]

Hello everyone,

I would like help with a problem I can't figure out. I have a ChatGPT account with a workspace. When I connect to ChatGPT with my login and password, I find my workspace and my GPTs.

However, when I connect to platformopenai with the same login and password, my workspace is not recognized. I can't see it on this page. Nor the GPTs I've created. Is this normal?

Thank you for your help.

[charred parcel]

the API is separate from ChatGPT, it's entirely different and is billed differently

[mental kiln]

Thank you for your reply, I understand that. But I'd like to retrieve the IDs of the GPTs I've created, and if I've understood correctly, I can retrieve them from the platformopenai. But I can't see them. Or maybe I've misunderstood.

[charred parcel]

you have misunderstood, ChatGPT is independent of the API, there is no connection between them

[wild meteor]

Is there any way you can allow me to complete my project?

[mental kiln]

All right. So how can I get the IDs of my GPTs to integrate them via an API in a CMS?

[charred parcel]

you can't integrate a GPT in a website or CMS

you can only use the GPT on ChatGPT

[charred parcel]

it's incorrect

[mental kiln]

So it's not possible to integrate a GPT product sheet editing module into a CMS?

[charred parcel]

a GPT only lives on ChatGPT, it is not embeddable, it is not able to be placed onto a website or third party product, it only exists on ChatGPT

[mental kiln]

What would be the best way to integrate an IA product sheet editor module into a CMS?

[charred parcel]

use the API. Ask about that in #dev-chat

[mental kiln]

Ok thanks, I'll look 🙂

[long pawn]

Hello, I have one question about the projects, i.e. these folders

When we work in projects, how does it work? Do the projects take not only the attached files but also the chats contained in them, using what is available these days?

[wild meteor]

I payed for unlimited and i still get this , why?

[charred parcel]

that's too small for me to read, what is the error you are getting?

[charred parcel]

just type in here what is happening, your screenshots are too small to read

[wild meteor]

pm me in chat.

[charred parcel]

no, type it here

[wild meteor]

can i uplad a jpeg image?

[charred parcel]

when you're ready to clearly explain in text what is happening I will try to help you

[wild meteor]

Chat GPT wont answer, like i generate and it wont reply anymore?

[charred parcel]

I don't know why that is happening to you, if you are having problems the correct course of action is to get support at https://help.openai.com/ • if you are logged in when going to https://help.openai.com/ you will be able to leave a message for support to get back to you

[wild meteor]

Ok i will check that out. But can you read my screen dump?

[charred parcel]

no, it's too small for me to read

[wild meteor]

Thanks, anyway..

[haughty nimbus]

I keep getting this error message when I try to create a GPT. The GPT does seem to work, but I don't want to loose it when I close my browser.

[charred parcel]

you'll need to contact support as the error message recommends. You can get support at https://help.openai.com/

[haughty nimbus]

Yeah, trying that. During the process of starting a chat it suggested Discord. And my chat message hasn't been read for over an hour now.

[charred parcel]

it can sometimes take days for support to respond

[haughty nimbus]

So I keep my browser with the chat window open for days, waiting for a response? That doesn't sound like what a chat should be like 🙂

[spare aurora]

Can I send you a problem of mine to your dms @charred parcel my payment is not working

[charred parcel]

I can't help with payment issues, sorry

[spare aurora]

It’s Js saying me “we’re not able to authenticate your method” but I’m using VISA international debit carddd

[charred parcel]

support can be slow, there are millions of users of ChatGPT and only limited support

[charred parcel]

Stripe often prefers credit cards over debit cards, that's possibly why

[haughty nimbus]

ok.. Thank you for trying

[charred parcel]

personally I subscribe using the ChatGPT App on iPhone, works out cheaper for me than the US to AUD cash currency conversion too

[sullen plinth]

So I just purchased ChatGPT plus and it had a error pop up but I was charged with 250 Swedish crowns (25$), can someone help me get chat PLUS or can I receive my money back?

[charred parcel]

if you are having problems the correct course of action is to get support at https://help.openai.com/

[rugged wing]

#ai-discussions message
Is this a discussion about this too or a request for help?

[rustic marlin]

lol

[raw thunder]

Hi guys, is being unable to upgrade to ChatGPT Pro a known thing? Or just me? Used to be a Pro member since Day 1 of this plan's introduction, temporarily downgraded to Plus for a few days, now it doesn't want to upgrade me back. There's only a few similar reports on the entire internet, with comments generally pointing towards OpenAI "being overwhelmed" with subscribers. But it was kinda crucial to me 😦

[rustic marlin]

what happens exactly when you try to make the purchase?

does it give you an error or something?

[raw thunder]

Already tried clearing cache & cookies as well as different browsers, to no avail. I've emailed the support already.

[rustic marlin]

weird, for good measure, try clearing the browser's cache or using an incognito tab

if that does not work, I think you will have to talk with support directly on https://help.openai.com

[raw thunder]

Yeah, I've done all of this already and waiting for the support to reply. Just wanted to see if it's just me. I wasn't prepared for this 😦

[blazing rover]

Hi guys, can I get some help with prompts or just simply image generation, I'm starting to feel discouraged with these restrictions. I keep getting restricted from seemingly completely harmless things, is there any way to go around this somehow, it's so exhausting + Even if ChatGPT changes its own prompt, it seems like the problem is with the image generation itself and not with the prompt

[charred parcel]

is this just in Dall-e?

[blazing rover]

Yeah over the last week or so I just keep getting restricted so often on things like 'slightly exaggerated design', anything even remotely dark, chubby, fire, flames etc. I just feel a bit hopeless at this point, why is it so restrictive even when I'm clearly not being malicious and some of the prompts aren't even anything crazy

I even had a problem with making a hybrid animal of a "pangolin" and a "mole", I am so confused

I'm just trying to make cool, cartoonish characters and get restricted so much

[charred parcel]

DALL-E has some pretty strict guardrails, there isn't really anything you can do other than experiment with variations once you hit them. Try asking in #images-discussions and maybe someone can help there

[blazing rover]

😿

Just so many ridiculous examples

I am so confused

[timber shadow]

Part of the confusion is that ChatGPT cannot be used as a reliable reporter of its own limitations in the first place (and to a certain extent, the same for its own capabilities as well).

In other words, when you ask ChatGPT: "What's the matter with this?" you implicitly assume that ChatGPT knows in the first place. Inherent confidence like this commonly causes LLMs like ChatGPT to hallucinate, kind of like ChatGPT is thinking "Hmm, the user sure thinks I should know the answer to this, I better give it my best shot..." 😁 Not exactly like that but you get what I'm saying

Usually if you're getting confusing content dings, just starting a new chat is the best bet

Trying to "reason with" the model in the same chat usually makes the issue persist

[blazing rover]

Yeah I understand that to some extent and I've noticed it would just randomly stop producing the same quality of images too but even starting a new chat, I do get way too many restrictions on pretty much completely harmless things

Just so sad, I'm never trying to make anything malicious or sexual or whatever

[charred parcel]

are you on a Mac or a PC?

[blazing rover]

And I would very much prefer if I could have my images organized but I guess you are right just starting new chats does help a bit

[blazing rover]

PC

[charred parcel]

are you using the Windows ChatGPT Desktop App or the ChatGPT Website?

[blazing rover]

Website, I had no clue it has a desktop app 😄 Would Desktop app help?

[charred parcel]

maybe, I found quite some time ago that I get less refusals on the macOS Desktop App than I did on the web

[charred parcel]

https://openai.com/chatgpt/download/

[rugged wing]

What? 😕

[rustic marlin]

?

[ancient sparrow]

my openAI key isn't working even though I funded my account, anyone have suggestions on how to fix this?

DM help would be preferred not sure I can keep up with a channel that moves this fast

[empty whale]

I hope I am in the right channel with my request. I have a ChatGPT subscription and am trying to get a new password sent to me for my ChatGPT account, but I never get an email from OpenAI telling me where to do this.

I feel like I've been through all the support pages and chat options, I've done everything they suggest and still didn’t get a solution to the problem. Do any of you have any experience with this problem? I would be infinitely grateful for a solution. Many thanks in advance.

[wild meteor]

If there is a limit in the utility Chat GPT even if i payed for unlimits, how can i expand my previous research in a new project field?

[timber shadow]

Hello, when you created your account, did you choose the email + password option originally, or did you sign in with a Google/Apple/Microsoft account?

[timber shadow]

Hey! There are usage limits on ChatGPT, from Free to Pro, it just varies based on what your plan is. You can read about the different limits per plan here: https://openai.com/chatgpt/pricing/ But the basic pattern is: the higher your subscription, the higher your limits. ChatGPT Pro is very close to unlimited, but it has some limits on the o1 pro model and has other stipulations regarding non-abusive usage

[wild meteor]

Ok, but i still got pro. Does it count for unlimited when it stops to generate my requests, and repeating the last sentence before it went like off?

[empty whale]

I have the email + password option. I also have an old email from OpenAI in my mailbox to reset the password. That worked once. But for whatever reason it hasn't since the end of December.

[timber shadow]

ChatGPT Pro has unlimited usage on all models (as long as that usage is non-abusive), except for the o1 pro model, which has a daily limit in place. As for what you're running into specifically: I'm not sure. Could you share more, share a screenshot, or maybe share a chat link if you feel like it to see what might be happening? Make sure there's no personal info in whatever you share

[timber shadow]

Yeah I'd guess password reset links are one-time use things. Hmm...so you're signed out, and you're clicking the "Forgot password" button, and you're entering the correct email address for your ChatGPT account, but you're not getting the email? Not going to spam folder, etc.?

[empty whale]

Exactly.
I check the email character by character and it is correct. It is also the email address I registered with. I also have no email from OpenAI in the spam folder or in the trash. I have not had access to my account/subscription since December 26th.

[timber shadow]

What do you see when you try to sign in with your last known password?

[empty whale]

It tells me “Wrong e-mail address or wrong password”

[timber shadow]

Dang, I'm not sure what to suggest trying other than reaching out to support at help.openai.com using the chat widget. My only other guess would be to check for possible emails regarding your account being banned for some reason? But I'd guess you'd already know about that. Hopefully OpenAI support can take a look for you

[empty whale]

When I go to “Forgot password”, I get: “Enter your phone number or e-mail address and we will send you instructions on how to reset your password.”
When I enter my registered phone number, I get the answer: “Invalid e-mail address. ”
When I enter my email address again, it tells me that it has sent an email. And I never receive it.

[timber shadow]

Dang, hopefully OpenAI support can help look into why you might not be receiving those emails. Or maybe someone else here will know what's going on

[empty whale]

I followed or had already done everything the support team advised me to do. When I asked further questions, my chat partner simply ended the conversation... I'm really desperate right now as to why I can't get in.

[ancient sparrow]

@timber shadow can you help me with the problem I have? My API keys are all non functional even though I funded my business account

[timber shadow]

I think OpenAI support would probably be the best point of contact for that. Make sure you're signed into your account first, then go to help.openai.com, open the chat widget in the bottom right, then get a support ticket created from there. Since it's related to billing, I think they'll have to take a look at what might be going on on the back end

[lean yacht]

Hi, someone help me please...

I suspect some conversations in a ChatGPT chat has been deleted without reason

Huh

What do I do then?

Should I leave this server?

But ChatGPT doesn't allow to delete messages unlike entire chats, some of them seem like disappeared

How is it possible

Individual messages and their answers

I remember perfectly certain conversations in a chat and now they seem to have disappeared

huhh

This makes me distrust ChatGPT

[bitter fiber]

Hi guys, good afternoon. I’m not sure if this is the right channel to address this issue, but I’ve been trying to top up credits on OpenAI for 3 days, and my card always gets declined. I’ve tried with 5 different cards, and all of them are declined. Previously, I’ve successfully topped up credits. Does anyone know where I can report this? I can’t find the appropriate channel, and their website doesn’t have anything related to this issue. I tried with different credit amounts, and it always fails: $10, $20, $200, $600, all of them.

[charred parcel]

if you are having problems the correct course of action is to get support at https://help.openai.com/

[hushed quail]

Is there a way to force canvas to recognize python? The model things it's python, but the canvas thinks it's text

[visual lodge]

hi! I use Chat mainly for help in writing, character development, plot ideas, and things of that nature. I noticed that on my PC my chat loads the answers all at once, rather than word by word like it will do on my laptop. I personally prefer the word by word reply. I use chrome on both, so I'm not sure why I get my answer as a big block of text that loads all at once on my PC. Everything I've looked for online is mostly people wanting the opposite of me, and articles saying why it probably won't load all at once.

[onyx kraken]

How long have you been noticing it loads so fast?

Typically it's affected by stuff like how many people OpenAI is helping right then; fewer people or more resources to help means faster speeds.

Also, your system's demand on your internet and your computer resources likely affect that, so you can maybe explore if streaming a video at the same time you are using ChatGPT might help get the speed slower like you prefer.

[visual lodge]

It actually takes ages for it load the full answer. Sometimes, I flip over to another app like Discord and then switch back while waiting. Usually takes 1-2 minutes to get my answer. So I don't think it's an issue with my internet speed. Let me see if I can screen record an example

[onyx kraken]

Sounds confusing.

So, you're loading the answer all at once, as in...

Like it's freezing and when you finally get the answer, it's just the full answer?

[visual lodge]

yeah basically!

[onyx kraken]

Okay!

So, for that you likely have a delay or lag or freezing issue. I'd make sure the computer's running well, and check for things like browser extensions or VPN or other things that might slow stuff down, basically 'get between' the chat and you and force stuff to take more time to update than it should.

It's kinda similar to the same stuff that could make a video stutter/freeze/skip ahead when you're trying to watch it.

[visual lodge]

I have a VPN, but it's not turned on, do you think that could still be causing it? I miss my word by word and finally decided to get help to fix it lol

[onyx kraken]

A good/quality VPN when off I hope isn't doing 'anything', so ideally that's not it.

Have you tried different browsers and seen if the issue's still affecting the chats on that comp?

[visual lodge]

I'll try edge real quick! Thanks for the help and advice btw

Yeah, it must be my PC potentially because it does the same on Microsoft edge

[onyx kraken]

And other stuff that can be affected by lag/freezing is fine with your system?

[visual lodge]

i had some issues with Vortex earlier when messing with my mods for bg3, and I was getting some weird fan noises when I first booted. I'll look into that (my husband will bc he built his PC) but that might be all it is. Thanks for telling me the verbiage I was missing! I didn't know it was a latency/lag issue

[onyx kraken]

It might not be, but at a guess seems like that might be what's happening. Someone who knows more might have a better idea, but that's my best idea.

[ancient sparrow]

There doesn’t seem to be an option tgere for me

If I can’t get support from ChatGPT after paying for it for over a year and not being able to use the api I will need to use another competitive ai solution. I build apps using the APIs but so far I’ve used other peoples accounts. Now that I need to use mine it’s not working

I also pay for about ten accounts in my company which will all switch with me. I recommend which ai to use to many people. If open ai can’t support its products I can’t recommend it

[timber shadow]

Make sure you're signed in first at platform.openai.com, then go back to help.openai.com. The options in your screenshot are the ones available when signed out - they'll be different once you sign in

Here's how it looks once you're signed in:

[haughty spear]

Hey everyone, I'm having trouble with JupyterLab. Even though I've installed the TensorFlow module, JupyterLab can't seem to find it. I've been stuck on this issue for 24 hours. Any idea why this is happening?

[near void]

real? 11Mil token that's a lot

[raw thunder]

It's definitely not on my part, lol. A redacted exerpt from my browser logs:

"detail" : "{\n "error": {\n "message": "The subscription is managed by the subscription schedule sub_sched_[CENSORED], and updating any cancellation behavior is not allowed.", \n "request_log_url": "[STRIPE]/logs/[CENSORED]/",/n "type": "invalid_request_error"\n }\n}\n

I sent this to OpenAI, they suddenly became really silent.

[raw thunder]

Referring to this.

[raw thunder]

Lol, they indeed have a glitch with Stripe, not wanting to admit it. In such cases you, unsubscribe from Plus first, then change your plan. I didn't need to do so previously though. Problem solves in 5 seconds - mine took 24 hours.

[fossil tusk]

read the fine print its a million besides mini :p

still nice of them, even if you are the product

[near void]

it said 1 Mil 4o, o1 10Mil 4o-mini, o1-mini

[ancient sparrow]

Thanks!

[bleak fiber]

How can I tell ChatGPT to not write the full code (repeat the code) and only write what needs to change, desirebly write it once and it follows in following up messages.

[steep plinth]

hello i would need some basic security instructions as my python gpt eventually gives up and loads its file database!

[pastel egret]

Is the 4o model for Free Users worse than for Plus ones?

[fossil tusk]

no just more limited context

[regal lynx]

Same base model

[steel bobcat]

Why does the formatting it's trying to do not work? I regenerated it once and it's still the same.

[wild meteor]

I run out of length again?

[onyx kraken]

Tell it what you want, and show examples of success, so it can follow patterns you set.

[bleak fiber]

Thank you! I tried a few things but it keep repeating the old code!

basically it writes more than what I needed

I wrote for example, "explain this function to me in short and technical terms"

[onyx kraken]

In longer chats sometimes really need to take the most useful stuff and go to a new chat.

Part of why it might say so much, it gets so much of the old context from earlier in the chat.

The model's like a power hammer. Line up the nails so it can drive them clean, be clear exactly what you want and if needed (sounds needed) give examples.

Get rid of all the distracting stuff, if you find the model 'babbling'. Be clear and direct and make everything in your prompt point to the result you want.

[onyx kraken]

This works great connected to 'small amounts of other info'. If you put in like 10k lines of code and need a fix or multiples and ask for that short and technical...

The model can get bogged down. It can get lost in the wealth of so much distracting info.

[bleak fiber]

what I did was a simple function which I took from the node.js documentation nothing more! I do know everything about it just wanted to test the CHatGpt. the code is about 10 line of code and it's nothing fancy

can I write it in "Custom Instructions" and the model remember to always answere me short but usefuell amount of info

and avoid repeating code

70% of the info it provides are redundant

[onyx kraken]

Not sure.

here's an example of what I'm talking about, is this what you're trying to get the model to do?

https://chatgpt.com/share/6780bbd4-ceec-8011-a813-395cdf7bb95a

My second input and it's output, what you're asking for in essence?

[rugged wing]

Hey there

[onyx kraken]

You can; success is maybe based on how you use language overall. I'd put how I want to be answered in the input, but that's me. No reason to not put it in custom instructions if you want to

[rugged wing]

How to use openai Developer platform ?

[peak sleet]

I am still on tier 1 even after waiting 7+ days and having added 51$ to my limit - what's up ?

Do i need to CONSUME 50$ before it counts ?

[gloomy solar]

Hello, I can't access other models like o1 and o1 mini on android mobile app. On the web version it works fine but they don't appear in the app. I tried reinstaling but that doesn't solve my issue. Anyone knows how to fix it?

[haughty spear]

Hi everyone, I'm facing some challenges setting up my development environment. I'm struggling to get Python, Conda, CUDA, cuDNN, TensorFlow, Keras, and matplotlib to work together seamlessly. I've spent considerable time trying to ensure compatibility between these components, but I continue to encounter installation and import errors. For example, even tensorflow.random isn't functioning correctly and kernel doesn't detect my GPU(Nvidia GeForece RTX 3060 12G). Is there a more reliable and efficient way to create a complete and stable Conda environment for deep learning projects?

[charred parcel]

what does that have to do with OpenAI?

[haughty spear]

I just want community help for that.

[haughty spear]

Could you help me with that?

[charred parcel]

use #ai-discussions for non OpenAI stuff

[haughty spear]

is that openai chatting channel?

[charred parcel]

this channel is "A place to ask the community ❓ Got a question about anything OpenAI? Ask here and the community will help you find the answer." you are asking about non OpenAI stuff so you should use #ai-discussions