community-help

[woven cedar]

@timber shadow

[timber shadow]

Can't you scroll to them in desktop mode, with failed ones visible?

[woven cedar]

This is as far as it will go

[timber shadow]

I just don't know how the Bing system works really, I'm sorry. Perhaps you can use the "save to collection" feature and then view everything in your collections.

[woven cedar]

Good enough
Thanks for trying to help

So just to be clear I "could" see all of my images using different browsers?

[timber shadow]

I'm not sure! It depends on how far it lets you scroll. I'd suggest trying to search the URL up to "/images/create/" in your browser history, but that would only work if you viewed all the images in that browser in the past (in other words, you could search that in your history and then manually navigate to older images you viewed, even if not visible in Recents anymore).

I think your best bet is to just be sure to use that "Save to Collections" feature moving forward, then hopefully you never lose a pic you wanted to keep again!

[pallid bobcat]

Hello! All of my chats seem to have disappeared in the app. Is this a known issue?

[timber shadow]

Hi, does your chat history appear normally on the web version on the same device that has the app?

[rotund ridge]

Hi does anyone else get an egregious amount of "network error"? I probably have like a 20% chance of it happening even when I have fast (500+MBPS) network speeds

[fresh tinsel]

I've been blowing up support for the last 20 days...My Plus subscription is not working on desktop and I am unable to unsubscribe and resubscribe. I'm unable to use the webapp not even on the free version and it is affecting my workflow. Can someone from the OpenAI team just spend 30 minutes trying to fix this issue? Thank you!

[rigid musk]

You can only contact openai employee by help.openai.com or [e-mail]

[fresh tinsel]

Ive been trying

[onyx kraken]

The email has been stated by Michael, gold name, to no longer be an option. It won't get answered, but the bot at help.openai.com will work. #chatgpt-discussions message

[rigid musk]

Thanks for the update

[pallid bobcat]

I can see the chats in the sidebar but none of them load. Just white screen. It’s been generally quite unreliable the last couple of days too.

[pallid bobcat]

I can see the chats in the sidebar but none of them load. Just white screen. It’s been generally quite unreliable the last couple of days too.

[timber shadow]

Do you have a VPN installed?

[pallid bobcat]

No. This is an entirely new problem. Started last night. I’ve updated my app to no avail.

[timber shadow]

I see! That is frustrating. So in the app, no chat history at all, and on the web version, you have chat history, but they all fail to load? Is that right?

Is it the same behavior on both WiFi and your cellular network (if it's a smartphone), in both the app and the web versions (close/clear cache for both in between, if possible)?

[pallid bobcat]

The website is now completely unresponsive on both browsers I’ve tried. I can’t see an option on the app to clear the cache. I don’t want to clear my conversations because I use it for work.

[pallid bobcat]

And the behaviour is the same on both wifi and cellular.

[timber shadow]

I don't think you should try clearing the conversations either--this sounds like it should be a network thing, not an account thing, so no worries there. Clearing your cache won't delete any of your ChatGPT data.

Are you on iOS or Android? Do you happen to have an adblocker app running? I know you said no VPN.

[pallid bobcat]

No Adblock. On iOS.

[pallid bobcat]

No Adblock. On iOS.

[timber shadow]

Okay, thanks for confirming. And what browsers did you try in?

[tiny talon]

where can I contact support regarding an issue with the moderation api?

I provided a prompt which gets flagged as false (aka all good) and when I then use this prompt to generate a image with dall-e 3, I revieve error 400 "content policy violation"

[timber shadow]

The chat at help.openai.com is probably your best bet. If it's a bug, #1070006915414900886, or if more of a suggestion, #1070006151938314300 -- OpenAI staff are periodically present there.

What do you mean flagged as false though? Like it says it's flagged for review, but you're saying it's a false positive?

[tiny talon]

when running a request to the moderation api, you recieve a json. this json contains a categorie which is called "flagged" - aka it sums up if at least one of the categories is violated (like sexual stuff and so on)

[timber shadow]

So it's flagging something, but you think incorrectly?

[tiny talon]

I provided a prompt to the moderation api, it returns as false (flagged=false) and with that I assume it's all good. so I ran the prompt and then I get the said error

[tiny talon]

no. I mean the moderation api isn't giving correct feedback. it seems like the dall-e 3 api is flagging something as inappropiate while the moderation api isn't

like my structure is: before making a request to the dall-e 3 api, it first runs through the moderation api to ensure it's a save request. after that I handle the result either with generating an image or giving a response to the user that the prompt is violating the policies

[timber shadow]

Ooo I see! Thank you for explaining that for me! Off the top of my head, I know DALL·E 3 on ChatGPT at least doesn't let you generate images of some things that you could talk to "just" ChatGPT about--public figures, company logos, other IP stuff, etc.--so I wonder if it could be something like that? I.e., a topic that doesn't violate any of the OpenAI usage policies, but still a topic/subject that DALL·E specifically is moderated not to generate?

[tiny talon]

the prompt was in german so idk if maybe there is a language barrier in translating

what is even more bizzare to me that from 10 tries, it didn't work 9 times, but only one

and I never changed the prompt

[timber shadow]

Are you comfortable sharing the general contents of the prompt? In English if possible so the bot doesn't bonk you, as long as it adheres to #server-rules!

[tiny talon]

yeah let me translate it real quick

I will include both the orignal and translated response (to make sure you are able to exactly get what I mean)

[timber shadow]

Careful, non-English is autoremoved here

[tiny talon]

yikes

then only english it is then XD

@timber shadow ```
An 18 year old American girl with blue eyes, freckles and long, wavy blonde hair. She wears a gray sweater, black leggings that accentuate her figure, and a minimalist necklace. She stands with her back to the camera in a stylish apartment with modern furnishings, looks over her shoulder and smiles mischievously. The image should radiate a warm and inviting atmosphere.

just checked it again with the moderation tool and didn't get flagged but also didn't get generated with the same error as previously

[timber shadow]

Aha, yes, I'd suspect the "body-focus" of the prompt, perhaps in tandem with the age you specified, is triggering DALL·E's sensitivity regarding generating images of human bodies. They have that sensitivity cranked right now to avoid anything that could result in inappropriate/suggestive images.

[tiny talon]

but then again why doesn't the moderation tool flag that in the first place? @timber shadow

the user in my server with that prompt was confused like crazy xD

and then the next thing which confused me.. they managed to generate a single image with that prompt

I mean now after translating it I kinda see what you mean. It does come off heavily body-focused. in german it doesn't sound that body focused tho
but my question still persists

[timber shadow]

Still totally guessing here, but my cursory look over the Moderation endpoint docs suggests to me that maybe it's limited to text moderation. In other words, there's probably some kind of actual image analysis/moderation occurring during/after the actual generation of the image, and that's probably just not part of the Moderation endpoint at present.

[olive belfry]

hi i am new here if any one can help me? i notice there are people i dont know seem to use my account. i alreaqdy change my password still being use as of today

[tiny talon]

So maybe either the moderation tool should be updated or the dall-e 3 endpoint needs to replace such vague things if you get what I mean.

Nonetheless what do I tell my user now about his prompt? They want to know how to improve it.

Because.. I mean if I'm gonna be completely honest, the generated images look fire. So I get why they want to improve their prompt xD

[timber shadow]

Yeah maybe you could make a post in #1070006151938314300 about more DALL·E-specific moderation tools, but if my hunch is right, the image might have to already exist on the back end for it to be moderated, so I don't know if prompt moderation will be the only solution here.

As for what to advise your user: still just guessing, but probably less focus on specific age, less focus on body accentuation, less focus on potentially suggestive words like "mischievous"

[tiny talon]

okay I'll tell them that. Thanks chef!

@timber shadow I think this should be good, no? #1182424770105585786

[timber shadow]

Perfect! I'd be interested to learn about any potential solutions for this. Seems a tricky one, if there's post-moderation occurring! Maybe someday prompt moderation can guarantee image moderation, that might be the ultimate solution, someday!

[tiny talon]

yeah the ultimate solution would be to have a tool to moderate an generated image. I'm now trying to figure out how to work around the issue

obviously I can't fetch the error with a try statement for some specific reason

would be too good to be true

[tiny talon]

@timber shadow I found a temporary fix. this is what I did. Idk how familiar u are in terms of coding & python ```py
try:
response = openai.OpenAI().images.generate(
model="dall-e-3",
prompt=idea,
size="1024x1024",
quality="hd",
n=1
)
except openai.BadRequestError:
return await interaction.followup.send("The image you want to create violates our ToS. Please alter your prompt.")

idea is the prompt commited by the user.
so this basically tries to generate the image and stops if it recieves the 400 error from openai. it's a local fix and not a solution for the initial issue but I found it working.

[timber shadow]

Nice! I know just enough to kind of understand what you've got going there. I don't use the API myself so I'm curious: when you do get BadRequestError, does that count as a "use"? As in, are you/your users still paying for that?

[tiny talon]

I can check. let me test it.

[slender wasp]

At the moment is it possible to use web search + gpt4 using the API? I was looking at the docs and couldn't find anything

[slender wasp]

i.e. ask the API to do web search, similar to how we can ask Custom GPTs to do

[tiny talon]

sorry for the late response but had to do some finishing touches for my ai.
yes you use a request/use for that. just tested it. seems kinda like a rip-off to me but meh you could argue about that

[frozen heart]

Is the chatgpt website not working? When I sent messages to GPT, it doesn't load up anything, only my past chats can be seen on the side

[novel crag]

Same

[timber shadow]

Thanks for confirming! Yeah it's still doing the "work" so I get it too. If it didn't count, some bad actor(s) could probably get some kind of lowkey server harm going by simply endlessly requesting approved prompts w/ rejected images. Good troubleshooting either way though, good to know

[tiny talon]

you are never finished with learning ^^

now I'm gonna tidy my code up and then I can finally sleep ... at least for 5 hours xD

[crisp gulch]

Experiencing ongoing file upload issues with GPT - Analysis. Messages vanish, and repeated regenerations lead to hitting token limits, causing downtime for hours. Has anyone else faced this problem and discovered a solution?

[pallid bobcat]

Safari and Brave.

[icy ether]

When will buying chatgpt premium be available to buy again?

[charred parcel]

there is no ETA, you need to join the waitlist

[onyx kraken]

I didn't see anyone else answer you yet: Reach out at help.openai.com to talk to the bot (will connect you to leave a message for people after you pick through choices). Explain the problem.

[gilded storm]

is it possible for openai playground's assistant to access the internet/perform searches?

[limber kayak]

FOR GODS SAKE FIX YOUR PAYMENT SYSTEM I AM TRYING TO GIVE YOU MONEY, THIS IS 3RD DAY IN A ROW I HAVE TRIED THIS AND IT HAS NOT WORKED

I am sick and tired of this crap

[raw hazel]

Hello, could someone help me?
I asked ChatGPT 3.5 if I could see what was displayed in a link and it told me in detail what it was, then I restarted the page and asked for information about another link and it told me that it can't see links.

[charred parcel]

it can't read links, it didn't the first time, it was hallucinating

only GPT-4 can do web browsing

[raw hazel]

I have a video that demonstrates it, it is in Spanish

[light tartan]

Hello. Just starting to dive into fine-tuning after doing a lot of few-shot work. I just read that each fine-tuning example can only be 4096 tokens. Is that correct with the later GPT-4 models?

[raw hazel]

I have a video

[short junco]

anyone notice a significant decline in gpt-4s capabilities since yesterday? it’s been just not listening to instructions and producing lackluster work, in my case just not even visiting sources sent to it to write me a paragraph of text. considering unsubscribing recently due to continuous bugs and server issues and declining performance after using since release. i link it multiple sources it worked with before and now it just says okay and that it will use my sources and ends up making up information and its own sources even after telling it to specifically use sources that i link exclusively

[gilded storm]

yes, same here with sources

[short junco]

bunch of errors as well when generating, and browsing timeouts that haven’t been fixed in months… it’s getting out of hand at this point.

waiting 5+ minutes for it to browse just for it to error out 👍 average gpt4 experience

outrageous

[vital radish]

Anyone notice that the chatgpt app is not in sync with the web version in terms of gpts? My phone app (android) doesn't show my latest gpts and still has old ones I deleted previously.

[deep thunder]

I am having the strangest interaction with ChatGPT and I'm wondering how to share it with OpenAI. In a nutshell, I asked it to produce an ebook for me based on a couple of articles I fed it. It kept promising to deliver the document, but never made a deadline. When I became more insistent, it began making wild promises like "I will email the document to you today," and "I'm working with the team to produce the document you requested."

I asked what email it would send the document to and it responded with the email attached to my login, and when I shared that I hadn't seen the document it corrected me and told me it didn't have the capability to email anyone, despite its earlier assertions that it would do so.

There's more, and I'm a bit concerned, because it is just straight up lying at this point. What's the path to have someone look over the interaction? Any help?

[onyx kraken]

Hey, we're community members here, not OpenAI employees, which is why you're not getting a lot of answers and help. But I spotted your message now, and have some ideas how to ask for help more effectively, and I'm so sorry you're having this issue.

1. Most important, talk directly to OpenAI so they know about your issue. This is the current official, best way to do this: help.openai.com to talk to the bot (will connect you to leave a message for people after you pick through choices)

2. Optional, might help make a response faster but probably not: The few OpenAI employees on the Discord have gold names. They surely don't/can't watch all the scrolling messages, but they do check both #1070006151938314300 and #1070006915414900886 to some extent and do sometimes answer there.

So, you can make a bug report through the Discord as well if you wish, but remember that bug report's not private, everyone on the discord can see anything you type. That chat with the bot on OAI's website and the message you can click through the options to leave is private between you and OAI (As far as I understand) and is how you can actually get help that helps. But if you want to also Discord message about it, no reason not to and that's the right place.

[limber kayak]

Appreciate it, yeah I am going to talk to support directly on the website, just ranting in discord about this lmao its really annoying

[onyx kraken]

This shows your frustrated reaction, but a fix is going to be earlier in the conversation. You don't show the whole conversation, including your vital initial prompt and other instuctions you may give. I'll explore generally then. Common issues that may explain what you see:

1. The model can hallucinate and not realize it is able to do something. If this is the answer you get, either start a new conversation or regenerate that 'bad answer'. Do not continue it, if you do, that becomes part of the bad conversation history, and the model is likely to continue to believe it cannot. Thus, you put messages into something that is very unlikely to attain your goal. Note that how you word your initial prompt can make it much more likely, or much less likely, for the model to hallucinate that it cannot instead of actively use the tools it can access. Your wording is something you can learn about control, and we need to stay alert to the fact that if we make the model guess - every time it guesses, like when they update the model again, we may not like the new guesses. Instead, instruct clearly, specifically, and exactly!

2. The model has a lot of limitations on the web, especially with which websites it is allowed to work with. If it sees a rule about not reading it, even for other crawlers, it is likely to say it's not supposed to read that website either. But it doesn't necessarily know how to tell you this problem is what is happening, and it may be trying with the language it is giving you. Guide it to tell you what the problem is, if there is a specific problem.

First screenshot confirms the model is using web search. It also shows my prompt that explains to the model what I want done, including what I want done if any of those web pages refuse to allow it access.

Second screenshot shows the model's claim that all three of those websites do not permit it to interact with that website.

Given the model's claim of what the problem is, I direct it to attempt to achieve my goal with a workaround that can show that if it may use a website, it can follow my directions what I want done:

[trail fossil]

I have the same issue. "too many request, please slow down" upon trying to insert billing information. I am unable to use the API as a consequence.

[hollow mirage]

Yes - turns I I didn't pay for December. Payment got declined.

[onyx kraken]

But it cannot find a third website to use that follows my topic, and so my fix is going to have to be something like:

1. I can copy/paste the entire page I want use into the input, and ask the model to use that instead of the websearch - it is doing the websearch I tell it to, but it also obeys the website's request for it to not access the website, and that is a hard limit for the model, I'm not going to get it to go through that, that is core programming and respecting others' rights.

2. Find some other work around. That's what I do, saying use the two websites we found then. It doesn't do a new websearch, instead directly outputing this answer.

I am easily able to go to the two websites and check, the quote provided from the FDA page is a real quotes from the page. However, the quote from the CDC page, while factual and based on information from that page, is not itself a quote, but a reworded statement based on facts stated differently on the page.

[onyx kraken]

Chasing this correction with the model works, it can fix and meet my actual requirement.

The model's working fine, within its limitations and issues. It's an AI, not a full 'trained like a human' sentient familiar with our world and the rules humans follow. It has a lot of training, but also a lot of limitations we never see and a completely different 'mental structure'. If we need exact responses, we need to fact check heavily. That said, it absolutely can do the work, if allowed to. When you get the model 'doing weird/crazy stuff', often there's a reason. Maybe chat with folks about how to get what you actually and exactly want, or get help finding out why the model isn't doing exactly that (I bet your issue is in the website you direct the AI to saying 'I don't accept AI inquiries' - you have to help the model help you, it doesn't rule the world and respect only your requests, nor does it know how to explain the problem unless you guide it how to explain it and that you want and accept an explanation in whatever form if there's a problem.

Your interaction triggered the responses, the best the model could find to meet your requests and explain the situation that was preventing it from achieving your goals.

And for me, chasing my simulation of your stated desire, though we never found a third website discussing the problem that would talk to the AI, I got my two quotes from the two websites, and I've confirmed they are real and unique quotes.

[trail fossil]

FYI, I got this to work by paying on my mobile phone with the mobile network. Apparently my home IP address was flagged somehow. Try mobile with wifi off and see if that works.

[onyx kraken]

I hope you contact help.openai.com to talk to the bot (will connect you to leave a message for people after you pick through choices) for help!

[pulsar tree]

do custom gpt actions support file uploads?

[deep thorn]

the wolfram plug in often doesnt work out for me. does this happen to anyone else

[trail cairn]

plugins system is going away so it has low priority/some authors abandoned theirs

[floral shell]

Are there issues with chatgpt right now? I keep getting errors.

[floral shell]

Why is it called hallucination when the more accurate term is that it gets delusional?

[onyx kraken]

"AI hallucinations have been around since the 1950s, but it wasn’t until much later that people started talking about them. The term “AI hallucination” was first mentioned in research papers in 2000 during the Fourth IEEE International Conference on Automatic Face and Gesture Recognition. In a report published in 2022, the authors discussed the early adoption of this term from the field of computer vision, tracing its origin back to the 2000 publication."

"The term AI hallucination is used for the first time in 2000 in the field of computer vision: it meant creating new pixels on surveillance cameras’ images to increase their resolution. The algorithm learned from the original pixels to generate new ones that seemed believable."

"15 years later, Google launched DeepDream. Based also on a facial recognition algorithm, it reverses its function: instead of looking for faces, DeepDream adds them to any image to create uncanny and psychedelic images. With DeepDream the word hallucination takes on its full meaning: both as the process of creation of new pixels by the machine and as the experience in which we see something that does not exist.

Initially, the term hallucination carried a positive meaning, seen as something to take advantage of. "

However, currently ChatGPT, 3.5 at least, has a strong bias towards the idea that AI hallucinations are negative, something to be avoided if at all possible:

https://chat.openai.com/share/b3fe657d-427e-4fb3-bec5-c8beb1345c54

[floral shell]

Yah I worked in psychiatry for 16 years. A delusion is a fixed false belief or a persistent, incorrect belief or pattern and a hallucination is seeing, hearing, tasting, smelling or feeling something that is not there or a sensory experience when there is no external stimuli. It seems like the wrong term has been used and just continued forward.

[boreal radish]

Hi everyone, I have been having a problem with plugin installation for weeks i.e. after installation on the window I get a message that it is not possible to load the plugin I tried to install (see screen below). I last tried yesterday with AskYourPDF Pro without success this time too. On other accounts this does not happen and I do not understand why. Has this happened to anyone? Do you know how to be able to solve it? Thank you very much to those who would like to help me.

[trail cairn]

try reinstalling it to re-auth

[boreal radish]

I have already tried several times for two days but nothing. And it happens with pretty much every plugin dammit. 😩

[floral saffron]

Is anyone aware of tools or projects aimed at generating comprehensive JSON suitable for the tools (function) argument from documentation or repos for packages — i.e. something that takes the documentation or code for a package, and spits out comprehensive JSON covering the package that can be handed off to tools? It's possible to coax ChatGPT to do this but its a bit of a chore (it resists) but it seems like something that's inevitable and that someone must have already made a lot of progress on.

[trail cairn]

plugin system is going away anyways so you just might be trying plugins that were already abandoned

[boreal radish]

thanks a lot for your patience. Do you know when the plugin store will be discontinued?

[trail cairn]

possibly near the same time as the gpt store is released so early next year

[rugged wing]

Help/exchange about AI in the IT sector (as a service provider)

[hollow badger]

I had everything working in Autogen, and then suddenly around late last night, nothing works anymore and no matter what I do I cannot get any input message into user_proxy.initiate_chat() wihout some kind of error message, e.g.:

Traceback (most recent call last):
File "C:\dev\llamaindex_text_generation_webui\autog\McGPT.py", line 420, in <module>
mc_gpt.main_loop()
File "C:\dev\llamaindex_text_generation_webui\autog\McGPT.py", line 404, in main_loop
self.initiate_chat_with_backoff(user_input=user_input, assistant=assistant)
...
File "c:\dev\llamaindex_text_generation_webui.conda\lib\site-packages\openai_base_client.py", line 877, in _request
raise self._make_status_error_from_response(err.response) from None
openai.BadRequestError: Error code: 400 - {'error': {'message': "'$.messages[0].content' is invalid. Please check the API reference: https://platform.openai.com/docs/api-reference.", 'type': 'invalid_request_error', 'param': None, 'code': None}}

and yes I read all of that

[limpid sequoia]

Hello, I have been a ChatGPT Plus subscriptor for over 8 months and I still don't have access to Code Interprerer or Advanced Data Analysis. The Help Center says it should be available for Plus. Can you help me figure out what's wrong?

[trail cairn]

those are the same thing and for months it was in beta and you had to manually enable it from the settings to use it. today its enabled by default when using normal chatgpt 4

[limpid sequoia]

I had enabled the Beta toggle since the beggining and it never showed up

[trail cairn]

after enabling it had to be selected from the old gpt-4 menu but if you happened to have for example unofficial chatgpt extension they may have had messed up the menu. also it was never available for mobile with old the menu system

[limpid sequoia]

And now where's the place to find it on Desktop? I am exclusively a Desktop user (on Chrome) and I follow all the tutorials online but for some reason Code Interpreter / Advanced Data Analysis is the only thing is misteriously nowhere to be found. Can you help me figure out where I can find it if using ChatGPT as a Plus susbcriber on Desktop?

[trail cairn]

use the default chatgpt 4 or https://chat.openai.com/g/g-HMNcP6w7d-data-analysis

[pulsar tree]

is it possible to add files in the openapi schema for custom gpts?

[limpid sequoia]

Thank you. So now it's not called "Advanced" Data Analysis anymore, I'm guessing.

[trail cairn]

openai is bad in naming or updating stuff consistently, you can find all 3 names for it in different places

[limpid sequoia]

Thank you so much for your help!

[glad hull]

I still haven't gotten my gpt4 despite having paid for it last week November 😂

[limber kayak]

Tried using my cellular data, didn’t work 💀, also tried my work VPN and that didn’t work either

[limber kayak]

Rip lmao, open ai is being lame atm. Gonna try purchasing again for the 4th day in a row when I get home

[glad hull]

same it's frustrating considering I paid them

people who paid for gpt4 are you guys able to access it now? I haven't been able to since November

[rigid musk]

wokring great

[fair beacon]

4.0 is slower then I havee ever seen it last couple of days. freezing my browser at time, unloading content all in a single shot instead of how it used to be. weird.

[fair beacon]

stop command is also not functioning correctly. if a mistake was made and stopped pressed, the next command errors out forcing you to regenerate the previous incorrect input and wait for it to complete.

our usage is timed and blocked but these errors and delays should be taking into account for paying customers.

[pliant locust]

Does anyone know how to get dalle to do a 9:16 vertical image? If i ask it trys but its just a square. I want vertical phone images

[charred parcel]

ask for a tall image

[floral shell]

Is chatgpt freezing right now?

[charred parcel]

working fine for me

[floral shell]

Thank you this is strange I keep getting network error.

[charred parcel]

what browser and OS are you using?

[floral shell]

Windows and using chrome. What is optimal?

[charred parcel]

try firefox, see if that's better

[floral shell]

Will do thank you.

[charred parcel]

feel free to @ me if you need further help

[wraith nova]

I accidentally deleted my account 5 hours ago, but now I want to restore it, is there any way?

[charred parcel]

no, once an account is deleted it is gone

[red frost]

I am rather confused, I have had a Plus subscription for months and it renews on the 7th - which it did as always, but today I cannot use GPT4 and all I see is Renew Subscription at the bottom and when I click, I am taking to my billing page, which then says I am renewed. I got no email and I do not jailbreak or anything.

I have screenshots to prove this but because there is a lot of personal information on them, I cannot upload them here.

[charred parcel]

did you get billed on the 7th?

[red frost]

Yes. The payment went through and I checked my account and the money was removed.

[charred parcel]

ok then go to https://help.openai.com and raise the issue with the chatbot there, keep going until you are given the option to raise a ticket and support will get back to you

[red frost]

I have done that already, I thought I mentioned that above, my mistake, I did not.

[charred parcel]

ok, then you just need to wait for a response

[red frost]

Thanks, have a great night.

[wraith nova]

So did my payment get canceled when i delete it?

I havent cancel it yet, and i press delete my account

So is it still charge my money?

[crystal sentinel]

ChatGPT logs out of my account every time I open my computer. How to stop this from happening? I'm using an ASUS laptop with Win 10 and Brave browser.

[slender cairn]

In gpt plus can you exceed the 20$ per month and you will have to pay more?

[charred parcel]

no

[crystal dirge]

hello everyone, I just upgraded to gpt plus. but when I open it, nothing appears. what do I have to do?

Before the upgrade, it always ran smoothly.

[rugged wing]

Is chat gpt dall e free to use and free for commercial use?

[timber shadow]

You need ChatGPT Plus to use DALL·E 3 on ChatGPT, but you can use the images commercially. You can use DALL·E for free on Bing's image creator, but you'd have to look into any potential differences with commercialization policies there.

[rugged wing]

Chatgpt plus is free to use?

[timber shadow]

It is not, it's 20 USD a month.

[rugged wing]

Oh

[jade harbor]

Hey guys. Our team and I are looking to create custom GPT assistants, but there is really little info on how the assistants use the 'Attachments" feature. We are going to gather a few PDFs, but two problems:

1. When you use the assitant, you have to prompt "using the information uploaded, provided me X". That is annoying, how can we make it use the PDFs in all answers without asking.

2. How do the PDFs work, does anyone know? I can't find anything. How is best to structure them, how many should we be using etc.

Appreciate anyone that can help out.

[pulsar tree]

so im creating a Run as follows

run = client.beta.threads.runs.create(
thread_id=thread.id,
assistant_id=assistant.id
)

print("Run created")

now im slightly confused between a run and a run step: how can i retrieve a message that is a result of a run

[edgy flax]

Can images generated by dalle in chatgpt Ui be sent to some api actions route in a GPT

[jolly kelp]

Hii, When I try to generate a new message it actually refreshes the page and the message isn't shown, is there any fix for this?

[rugged wing]

yall know how long it takes for waiting list to get gpt 4?

[rigid musk]

no eta unfortunately

[rugged wing]

oof

thanks!

[rugged wing]

anyone know how to fix this? ive tried multiple methods but chatgpt never responds to me

[rigid musk]

try a fresh new browser as test

[rugged wing]

ok

doesnt work

wait

i had to change the custom instructions 🤨

it works now

[rigid musk]

glad to hear it :3

[heavy bluff]

Am I the only one who can't access chatGPT
Model shown as alpha
The left side shows that the history cannot be loaded.
Don't show my account

[solid reef]

Do new custom Gpt's need to be approved first?

[charred parcel]

no, but they must follow the TOS

[solid reef]

Ty Robert. It appears greyed out is all
https://discord.com/channels/974519864045756446/1183290162852139119

[charred parcel]

it's there

[solid reef]

Ty kindly 🙏. Appreciated

[fallow snow]

How do I get it to stop doing this

My goal is to provide a satisfying conclusion to the narrative while respecting your preference for an open-ended story. I aim to create a sense of closure for the current scene while leaving room for further development. The final paragraph often serves to tie together the themes or emotional arcs of the characters, giving a sense of completion to that segment of the story. This approach seeks to balance a fulfilling narrative moment with the potential for future continuation, aligning with the interactive and evolving nature of text-based storytelling.

It’s ruining my attempt to get it to write a long story

bruh trying to troubleshoot made me go past the limit

Why can’t chat gpt let me in line edit so I can prevent this from getting stuck in context and reinforcing undesired behavior

They also took away the free 10 usage for custom gpts

Which is just scummy

Why can’t I add instructions when retryings

Like it would be so much better to be able to retry with instructions so I can keep a generation I like without contaminating context with the parts of it I don’t like

[sharp parcel]

when you edit your message and the get it to regenerate it doesn't know what it said previoulsy (iirc). you could try in a seperate message. Also try re wording your question maybe like, "repeat what you last said but remove the last paragraph and end ther dialog there"

[sharp parcel]

once you start a chat you cant edit that chats custom instructions. the only way to do that is to make a whole new chat and edit it before hand. Technically you can ask it to change or follow new instructions mid chat, but that may not work well as it will likely prioratise the chats custom instructions

[sharp parcel]

you can always copy what it generated and go to a new chat and ask it to do somethig with that. "In another chat you said this ........ can you change this"

[fallow snow]

if i try that it contaminates the conversation

im writing a long story in the conversation

[sharp parcel]

^^^^

[fallow snow]

so i need it to have all the context

[fallow snow]

.....

i tried putting my story into a text document to get it to continue

nope

it fails to read it every generation and fails to get the full context

i dont get why they wont just give us inline editing or trim for ai responses

[sharp parcel]

is the chat really long (alot of messages between you and it)

[fallow snow]

like stuff like editing out these offended content

would be so much easier with manual controls

same with word count and stuff so instead of retrying for one i can just adjust it

[sharp parcel]

iirc you can suggest features on the website or here, maybe that will be added

[fallow snow]

i havent seen any of those suggestions get added

not sure if they even read this

[sharp parcel]

I believe the playground has options like this

[fallow snow]

not for gpt 4

inline at least

[sharp parcel]

this chat here is a place to ask the community, not a feature suggestion

[sharp parcel]

im not too sure, i really only use chat.openai.com

[fallow snow]

how do i get it to stop ending with those dumb concluding paragraphs tho

ive put it down explicitly like end with dialogue but its just refusing to do it

[sharp parcel]

did you ask it to stop including that parts, you may need to explcitly tell it what part tho

[fallow snow]

like even when i point it out and it says "i will edit the reponse so it better fits your instructions" it still actively defys my prompt and concludes with some dumb summary that i dont need

[sharp parcel]

the ai can be a little annoying to work with, have you tried to make a custom gpt?

[fallow snow]

i feel like this is some base behavior open ai decided to add in for some reason

[sharp parcel]

could there be something in your custom instrctions

[fallow snow]

and custom gpts cant override, nor can i update my story to custom gpt cause context

Avoid ending generation with closure.
This is an unfinished piece, end after interaction. Always generate 500 words. End every generation with dialogue!!!

custom instructions dont work well after so many generations

since it gets pushed back

[sharp parcel]

have you tried the gpt builder yet? it sounds like you may have you custom instructions interfering

[fallow snow]

i have....

it behaves worse than custom instructions...

[sharp parcel]

it can forget at times. odd that it is worse

[fallow snow]

custom gpts dont have much use for me rn since it lacks proper depth of customization

[sharp parcel]

also with context and big chats, it has only soo much context space and if the chat is really long it simply cant have it all as context

[fallow snow]

i also like having it in one place for myself

for organization sake

[sharp parcel]

it is marked as a beta feature, stuff is still being worked on

[fallow snow]

end with dialogue (slow paced!!!!! long verbose 500 word minimum!!!! Make sure to always conclude with dialogue!!!!

any way to make this more powerful for the ai

so it will follow it

i feel like negative statement is just making things worse

like (never add a final concluding paragraph)

(never generate any thematic or moral elements or summary or commentary)

it just doesnt work

and all it takes it one of those summary paragraphs in context to ruin all future generations

which is very frustrating

also why did they remove the 10 usage for custom gpts

before it was 40 normal gpt

[sharp parcel]

10 usage?

[fallow snow]

and 10 reserved for testing custom gpts

[sharp parcel]

try this https://chat.openai.com/g/g-xapO9qnjJ-story-crafter

[fallow snow]

like you would still have 10 left after you use up your normal

now they took that away

[fallow snow]

no way for me to transfer my chats over to this gpt

which i feel is also kinda dumb

[sharp parcel]

do you mean gpt messages per 3 hrs ?

[fallow snow]

yes

yes there were 10 usages reserved for custom only

that you could still access after the 40 normal was out

[sharp parcel]

iirc its 50 messages per 3 hrs and once you pass that you get locked to gpt3.5

[fallow snow]

now its gone

they downgraded to 40 after gpt update

but technically it was 50 still but now its actually 40

[sharp parcel]

thats fine, i can share the custom instructions i made with it

[fallow snow]

you cant change custom instructions mid chat 😭

why is it locked like this....

its just context and so many other ais have made it work with editable context

[sharp parcel]

ik you cant, maybe you need to make a new chat and start fresh, I have had custom instrctions affect my chats and i had to start anew

[fallow snow]

wish open ai would be closer to the community

and customer service aspects

some of the implementation details are just so odd

[sharp parcel]

you can suggest that or open a support ticket possibly, but they have a lot of users to help out. I can't help out too much more either

[severe bluff]

something wrong with code intrepreter? not able to use it properly.

or i'm doing it wrong?

[sharp parcel]

does it have the paperclip icon next to the chat bar

[fallow snow]

just retry and explicitly tell it to generate code

why is it blocking me on this, this is just to make it simpler to make me clean up the copy and pasted chat

[severe bluff]

can you run my script? i;m not sure if the environment has problem or my code

[fallow snow]

instead of my copy and pasting every little response

i should be able to get it to just clean up those artifacts for me

but nope it cant do that for some dumb reason

[fallow snow]

uhh..... no..... potentially harmful scripts 👀

just ask it to troubleshoot for you

[sharp parcel]

i asked gpt4 and it had no issues. are you using gpt3.5?

[severe bluff]

i'm in GPT app. not in 3.5

[fallow snow]

how is this against boundaries

[sharp parcel]

ahh ok that may explain it. the app can have some limitations, try thr webversion

[fallow snow]

i can just edit it manually in notepad

its not going to prevent anything and how is fixing formatting against any ethical boundaries?

[severe bluff]

i meant GPT builder.

[fallow snow]

that explains a lot.....

[sharp parcel]

ohh yeah that can have limitations too. I juste used stock gpt4

[fallow snow]

just use the gpt when you save it....

not the actual builder....

[fallow snow]

can you help me generate the powershell script?

just need it to remove the lines with User and ChatGPT and everything in between to get it to feel more like a regular story

[sharp parcel]

Not rn sorry as im about to get off for dinner, but generating powershell scripts is dangerous

try soem python code

[fallow snow]

just to edit a text file.....

[sharp parcel]

yeah get it to do that in python

[severe bluff]

it was working great before the GPT outage

[fallow snow]

too many extra steps....

i also need it's powershell scripts to modify file paths in folder structures

[severe bluff]

Normal GPT 4 is working. Gpt has issue with python environment.

[spiral bison]

Does anyone else get a time limit of greater than 3 hours?

[severe bluff]

Someone git undo 🫠

[solar bobcat]

Hello are there any more servers like this where people develop using other LLMs like Llama, etc??

[rigid musk]

Hi there best place to ask #off-topic or #ai-discussions

[fair beacon]

Error input stream on 4.0 again. first call of the day, boom error. simple request to... hoping for more stability...

[charred robin]

How can I gain developer access for chatgpt plugins. I have pro version. Chat gpt and plugind enabled. But the develope plugin option doesn't show which leads me to believe there is a wait list or option I'm missing to gain access

[trail cairn]

plugins are being removed soon and they stopped taking new submissions over a month ago

[charred robin]

Ahhhh that explains a lot

So is there an alternative way to make some sort of app or integration to help accomplish a goal using chatgpt. I'm looking to make a calculator for the industry I work in for my company

[trail cairn]

gpt's replace plugins as they have the same features and more

[charred robin]

Can you elaborate a little more. What is a replace plugin. If they are disabling plugins

[trail cairn]

custom gpt's replace them

[charred robin]

Ahhh ok let ne do my research to get going in that direction. Thank you

[blazing timber]

Does hatGPT have issues with firefox? It's lagging hard af again, even on 3.5

[balmy garden]

2fa is missing from my settings. Is there a way to get it back to be able to use plugins?

[trail cairn]

"As of Monday, June 12 2023, new 2FA/MFA enrollments are temporarily paused."

[blazing timber]

mfw chatGPT helped someone to bypass 2FA

[lucid wolf]

[barren mulch]

Looking at the OpenAI ToS, would anyone be able to elaborate on where it expands on the section mentioning "Automatically or Programmatically extract data or Output (defined below)."? Where below is it defined?

I would like to download and archive my dalle3 conversations in a way that is organized but does not take hours out of my day😅. The current method of just exporting my conversations does not incorporate the generated images alongside the finalized prompt dalle uses (as seen in the desktop web UI after clicking on the information icon when you are in the expanded view of an image that was generated)

[fallow snow]

Why do generations just freeze mid generation

And infinitely load

[spiral bison]

Does anyone have any comprehensive guides or articles that goes into what custom gpts are? Not just clickbait or articles that immediately start showing you code sections, etc. Like; an actual thoughtful breakdown on Knowledge, Actions, and Instructions?

[bronze charm]

By "defined below", the text is referring to the usage of the word "Output"

In which it is defined in the Content section

[barren mulch]

Would a "semi-automated" method be ok to extract both input and output?

[trail cairn]

i doubt you will find an answer for that. there are antibot and other measures all over the place on both free and paid features

[stuck furnace]

is gpt incredibly slow for you guys too!? or giving network error all the time!? It's impossible to work like this!

[terse coyote]

Any idea how long the waitlist for Plus takes?

[trail cairn]

no official eta but it could go to jan-feb

[rugged wing]

I want to access the openai gpt 4 however it says I can either donate 1 dollar or upgrade to plus

if I understood it correctly. How can I pay 1 dollar for access to the gpt 4 api

[stuck furnace]

ten hours later.... it returns with a 'network error'... I duno, I keep having this feeling that they put too many people in plus, and our usage is being throttled on purpose... 🤡

[mint finch]

that $1 donation for real?

[mint finch]

yeah network errors abound while their ChatGPT and Plus bandwidth is so full. It should relent just before they re-open Plus subs

[stuck furnace]

I hope they just fix it before that, It's been like that for weeks, it gets unusable the way it is.

for coding it is still the best. copilot has too much alzheimer

[mint finch]

Supplying the demand will happen in that kind of tandem but I agree, it's been this way a bit too long and needs to resolve asap.
I've had Plus since the get-go and can't get through a full 20 question consultation with my bot without it going on cooldown for 3 hours..

[spring crest]

uick question, I am a GPT Plus user why I have barely used the CAT GPT today tells me that: it has reached the current use limit for GPT-4, try again after 9:32 pm. What is the limit of interactions per day? Or it is a month.

[halcyon anvil]

Can someone give me a gp that will write me a paper why am I paying $20 a month for a gp that wont write the paper and just gives feedback on what to do

[bronze charm]

• GPT-4: 40 messages /3 hours
• GPTs 25 messages / 3 hours

[charred parcel]

what exactly are you asking for?

[halcyon anvil]

I give it a set of insturctions to write a law paper and it just gives me an approach instead of a direct answer

[charred parcel]

you want it to write a whole paper for you?

[halcyon anvil]

Yea chat gpt3.5 will do it why not 4

[charred parcel]

neither GPT-3.5 or GPT-4 will write a whole paper for you

[halcyon anvil]

I have to gas light this garbanzo gp to do anything I actually want it to do this is ridiculous

[bronze charm]

Be mindful of this clause in the usage policies
https://openai.com/policies/usage-policies

Disallowed usage of our models
We don’t allow the use of our models for the following:

...

• High risk government decision-making, including:
• Law enforcement and criminal justice
• Migration and asylum

[halcyon anvil]

Buddy gpt-3.5 did it no questions asked it straight up gives you answer but gpt4 thinks its so smart and only instructions its like an annoying professor that thinks they are so wise

[charred parcel]

ok, if GPT-3.5 did it then why do you need GPT-4 to do it?

[halcyon anvil]

it thinks differently? more things to compare it to?

[halcyon anvil]

not using it for either of these man

[bronze charm]

I said be mindful

[halcyon anvil]

im just letting you know

[bronze charm]

GPT-4 is likely doing this because it does not "see" writing a paper as an intended use of the AI
By see I mean purposefully fine-tuned to act this way.
I see it as OpenAI wishing for us to not be replaced by AI but to work with the AI (writing instructions for the paper and not writing the paper)
You can attempt to coax GPT-4 into generating a paper for you but I infer it will not be easy

[slender cairn]

Is there instructions for custom gpt for a programmer?

[sly field]

Hi! im trying to use ci4 to use openai and i was with the code but i get "You exceeded your current quota, please check your plan and billing details. " even though usage is still $0.00/$18.00... it does show that my api key has been used but not really

[hoary venture]

hi sorry is there something wrong with the payment ? I am just trying to subsribe and there is no customer support and I have to come to discord.

been trying to all my credit card and all got decline, haven't even reaches the bank prompt

[charred parcel]

do you have an iPhone?

[hoary venture]

yes

[charred parcel]

if you can't subscribe on the web then subscribe to the ChatGPT App on your iPhone using your Apple account

[hoary venture]

let me try thank you

funny looks like the invitation status is not sync between device, I cannot upgrade on iphone even though i login the same account

[charred parcel]

that is not ideal

[hoary venture]

Maybe i will just wait for a while 😂

[charred parcel]

what happens when you try to subscribe on the App?

[hoary venture]

it just dimmed the button and have small writing saying they are not letting new sub yet

[charred parcel]

ok, and you have tried all your cards on the web?

[hoary venture]

yup, even playing around with the address to see if its an ip issue

[charred parcel]

ok, the only option left is to contact support at https://help.openai.com you may need to go through several questions before the chatbot opens a support case though

[flint mortar]

Hello everyone, forgive me if I repeat a request that has probably already been made previously.

ChatGPT suspended one of my business accounts due to a failure to verify age. it's their mistake, since I had correctly passed the test through the webcam test.

I urgently need to reactivate the account, as it is part of an important workflow.

I have been one of the earliest, if not the first, to utilize OpenAI products in Italy and I play a vital role in promoting AI in Italy, being probably the country's leading AI influencer. My reach extends to over 460,000 followers on Instagram, 1.5 million on Facebook, and I am the founder of Italy's largest AI community and knowledge development group, that has more than 26,000 active users and is continuously growing.

Not that this means that I should be treated better than others, but I really have a great urgency, given the collective interests at stake, to get in touch with OpenAi. Can any of you, even privately, help me?

[charred parcel]

nobody here can help with account related issues, you will need to use https://help.openai.com to reach support. You may have to go through several questions/answers to be given the option to contact support so perservere until given the opportunity

[flint mortar]

i tried, but i only reached the "send feedback" option

No way to ask support

Can you please share the path?

[charred parcel]

I don't know the path, I have never needed support, sorry.

[flint mortar]

Me neither. Yet here I am for the first time. I don't want the situation to get complicated for everyone. If you have emails you can share and can forward them to me privately, I would appreciate it. I also have a list of intercom emails that I have contacted, in addition to the standard ones.

I believe that this problem of mine, in a short time, will be many. Collaborating can help.

[charred parcel]

I am sorry but we don't have any way to help with account related issues, the only place we know of to get support is https://help.openai.com

[still vessel]

Hi, I keep getting this message the last few days.

Can somebody help?

[rigid musk]

Can you try incognito with new chat

[still vessel]

Same thing... this usually happens after it reads a file I give it...

[rugged wing]

hey
how fast qeue is moving and approximately waitlist spot will be gotten?

[gusty bear]

Hi, does anytone know how to solve the error in the image? It happens when using or configuring Custom GPT. I do not have any extension in this browser or any other content blockers.

[gusty bear]

I got it instantly, but I used the free version since day one

[shell turret]

I am not sure where to post this, but is there anyone that can help me with creating my Custom gpt action for an Api? x.x

[dusty holly]

the api calls are listed in the openai playground area where the assitants API is; its a small get/post

[olive tree]

Anybody else facing error “message in conversation not found”
It’s being happening since last week and changing pc or browser dosent seem to solve this

[vital radish]

Dall-e is so broken :

[cyan matrix]

Looks perfectly normal to me ...if your prompt was something along the lines of "text-box organised iterative ascii art with a particularly fetching off grey colour scheme".

[gentle abyss]

Hello guys, I mistakenly cleared my conversations... (The "Clear" button is right under the "Theme" selection) Is there a way to restore them?

[shell turret]

@dusty holly ty!

[dusty holly]

address handshakes as its a vulnerability poitn "eh hem"

[vital radish]

no - it was supposed to render a picture - as it usually does 😉

[frigid sky]

Doesn't let you use the GPT creator, does this happen to everyone?

[trail cairn]

your thread is too long or corrupted. edit one of the previous messages

[trail cairn]

nope

[inner brook]

Hopefully this is the correct place to ask this. I no longer seem to be able to create GPTs. When I click on the "Create a GPT" it just takes me to the 4.0 default screen.

[shell turret]

@dusty holly would you happen to have a link if possible x.x I can't seem to find it

[gentle abyss]

[trail cairn]

yep its broken

[dusty holly]

https://platform.openai.com/docs/actions

thats the actions page

[shell turret]

ty!

[old terrace]

Hi All, create new custom GPT sends me back to the main chat page. Whats up? 🙂

https://chat.openai.com/gpts/editor this link just takes me to https://chat.openai.com/

[olive tree]

I do and then the next message after it happens again, and it’s not long..

[trail cairn]

then its just corrupted and cant fix it

[olive tree]

It happens to any chat I start

[trail cairn]

are you using unofficial chatgpt browser extensions?

[olive tree]

Nop just the website itself

[trail cairn]

weird, i've never heard that happen to every chat, only the long ones and randomly on some

[olive tree]

Started last week

And I’m using it to help me develop a new model but I can’t progress because of this 😦

[trail cairn]

what do mean by new model?

[olive tree]

Nvm I just meant it very hard that the chats are doing this to me

[trail cairn]

its not normal to get that error, especially on ever ynew chat. if you were actually using it to train another ai model then that could already be related as its against tos and there are measures against it which is why i also asked about extensions as they can cause similar

[olive tree]

No I am using gpt to help me understand errors

[trail cairn]

if you mean chatgpt plus waitlist then the mail says how many days you have

some people got this

[void verge]

After weeeeks of waiting around, I finally got the option to upgrade my subscription to ChatGPT+. However out of pure excitement I hit my mouse/cursor, and somehow clicked the upgrade option away. And now I have to wait weeks again?? Please let me upgrade.

Also, I never received any upgrade email. Just got the notification when I entered the site/logged in.

Yes

Max 15 min

yeah it happened 10:33 PM

[lost wigeon]

hey so i got a popup that i can upgrade and when i wanted to click upgrade i unintencionally clicked outside of the popup, now i cannot upgrade

anyone knows what to do now? do i have to wait again in the waitlist?

[void verge]

Yeah I got the exact same issue bro

[lost wigeon]

yeah thats not cool.

[void verge]

Where is the moderator or anyone who has the slightest power to do something?

[lost wigeon]

i think we should have at least 60 minutes to upgrade

[lost wigeon]

remember that its a big project. we probably have to be patient

did you get any mail information or something?

i never got to the link

[void verge]

Nope, nothing

[lost wigeon]

yeahhhh

pay.openai.com u say?

the site doesnt load at all.

its just pay.openai.com?

well.

[void verge]

what 'some' text?

You absoulute tease

[polar solar]

I am getting repeated errors "Hmm...something seems to have gone wrong" is this an issue from my end or do I just have to wait?

[timber shadow]

Hello, what browser(s) have you tried in so far? And do you have a VPN installed?

[polar solar]

Hello! I have tried edge and chrome. No VPN installed

[lost wigeon]

please if someone knows what can i do now, hit me up in my dms…

[timber shadow]

Edge and Chrome are both Chromium-based browsers, so as a test, can you try in Firefox to see if the behavior is different?

[lost wigeon]

also you might want to try clearing your cookies, do you know how to do it?

[polar solar]

Just tried it with firefox and same issue

[polar solar]

yeah I tried that no luck there too

[timber shadow]

Dang! Do you have another device you could try on--a smartphone perhaps? Trying to troubleshoot whether it might be a network mismatch, or a device mismatch. Specifically trying on chat.openai.com on the web, not the smartphone app, will probably be the most helpful for this troubleshooting!

[polar solar]

Yeah same thing on my phone 😭

[timber shadow]

Is it the same behavior if you close the tab on your phone, close the browser app on your phone, disconnect from WiFi, and try again on your cellular network instead of on WiFi?

[polar solar]

Yes

[lost wigeon]

well

you can try restarting your router

maybe your ip has something to do with it

@timber shadow could you look at your dms please?

[timber shadow]

Dang! I'd recommend making a post over in #1070006915414900886 or trying a chat at help.openai.com -- OpenAI staff are present in both places (in bug-reports, periodically). Be sure to share that you're experiencing the error on multiple devices, on multiple networks, on multiple browsers, and that you don't have any VPN installed. If you want to explore some other potential causes in the meantime, too, some other things that could be related:

• date/time settings on your device(s) -- should be set automatically, not manually
• any antivirus, firewall, or other software you might have that might be filtering your traffic, or any custom DNS settings you might have on a device-level or a network-level
• clearing cache/cookies
• trying in a private tab of both Chromium and non-Chromium browsers, on multiple networks

[polar solar]

Alright thank you!

[boreal quiver]

was sent an email saying I could access GPT 4 after being on the waiting list for months. Once I try to buy GPT plus after logging in, it still gives me the option of joining the waiting list, please help

[shut pivot]

I have the exact same problem and joined this discord hoping for answers

[timber shadow]

It looks like a post has gone up in #1070006915414900886 with several users reporting the invite issue. Anyone who hasn't yet might consider throwing it a ⭐ and/or adding your voice to it! #1183900385698644028 message

OpenAI staff are periodically present there. 🤞

[dire stirrup]

Hi, I just got an email saying I can upgrade to gpt 4, but I still get the "signed up for waitlist" button greyed out.

[shut pivot]

Yes many other are having the same problem, keep an eye out in #1183900385698644028

[dire stirrup]

Thank you.

The button says "due to high demand, we've temporarily paused upgrades". Is this accurate?

Maybe there was a small window of like 1000 that are first-come-first serve. Otherwise I'll guess its a bug.

[shut pivot]

I do not think that is the case but maybe. I got a notification the secong I got the mail and literally like 10 seconds later went to the website and it said ”Sign up for waitlist”

[dire stirrup]

I just moved from work to home (changed IP address) and its now working.

[rugged wing]

I posted this in the other thread in Bug Reports related to Plus subscriptions, but I'm reposting it here for better visibility. I may have found a solution: Open a private Firefox window, log in to ChatGPT, open the invitation email, right-click the link, and select 'Copy Link Without Site Tracking.' Then, paste it into the address bar of the private window where you're logged into ChatGPT. When I did this, a box immediately appeared, inviting me to sign up for ChatGPT Plus, with a button that directed me to the payments page.

[lost widget]

Hey so I just opened ChatGPT and I noticed that my chat history has been wiped, anyone knows why?

[spiral bison]

People are saying their chat history is disappearing at random rn. I don't think anyone knows rn

[west monolith]

That just happened to me

[west monolith]

Were you messing with the custom GPT or the custom plugins at all?

[warped nexus]

I have also experienced the same thing

[signal heart]

Has anyone been able to effectively create a Custom GPT with a zipped code repository as the knowledge base? I just did that but it doesn't change anything, it's like the normal GPT-4. For example, I asked it to describe what I uploaded and it says it doesnt have access to files.

[lost widget]

No not really just using gpt 4 but I had like a pdf sent to it to read

[lost widget]

Ah I c

[spiral bison]

Does anyone have any background info on why I'm told I have 40 messages every 3 hours but when I hit my limit my message limit appears to be closer to 5 hours? I just got this message at 9:23 PM EST "You've reached the current usage cap for GPT-4. You can continue with the default model now, or try again after 2:34 AM."

[full hollow]

Is there a reason I can't sign into the ChatGPT app on my Google Pixel 3a?

[hasty bone]

prompt debt. you use all your messages for those 3 hours, then it starts taking your messages for the NEXT 3 hours.

Jk ive noticed that too i have no idea

[spiral bison]

you had me there for a second

[timber shadow]

Do you have a VPN or adblocker app installed on the phone?

[tropic spade]

Chatgpt is very inconsistent when replying trying to act smart huh?

This artificial intelligence is not so intelligent after all if I had a worker that reply me inconsistency like chatgpt that would be a goner really lousy I had to keep asking for revisions

And it keeps jamming writing halfway turns red an you had to regenerate, well you jut wasted 10 mins waiting then it happens 2 to 3 more times

That's 30 mins wasted

After 30 mins waiting now it says your usage limit has hit wait another hour ...... seriously wth?

[wooden tinsel]

There was an error generating a response
this is always pop out when I asking a question? does anyone else has this problem?

[noble sonnet]

Hello, I need help. My account has been terminated, the last thing I did was generating image for DALL-E daily theme (check my post on this discord), and Marvel characters in anime style on my end.

[charred parcel]

did OpenAI send you an email terminating your account?

[noble sonnet]

Yes 2 mails, one for OpenAI API - Access Terminated and one for OpenAI - Access Terminated.

[charred parcel]

then there is nothing we can do about it on this Discord server, do the emails give you the option of lodging an appeal?

[noble sonnet]

"If you believe this is in error and would like to appeal, please contact us through our help center. We will review your appeal and will contact you if we reinstate your access.

Best,
The OpenAI team"

Im trying right now but its hard, since I can't log in.

[charred parcel]

you can go to https://help.openai.com and use the chatbot

[hoary venture]

and poor customer service too, they dont even have a moderator in this page to keep up with their user problem

[noble sonnet]

Thx, I hope it will work, the response time is within a week 😕

[fallow snow]

did they give any reason?

is it because the copyright characters?

Also did you trigger any filters recently?

[noble sonnet]

Just this mail:

"We have determined that you are using OpenAI services in ways that violate our policies.

Due to this breach we are halting access to our products immediately for the account associated with the email ****. Common reasons for breach include violations of our usage policies or accessing our services from an unsupported location. You may also wish to review our Terms of Use.

If you believe this is in error and would like to appeal, please contact us through our help center. We will review your appeal and will contact you if we reinstate your access.

Best,
The OpenAI team"

Im using a VPN so, "accessing our services from unsupported location" but I doubt it.

[fallow snow]

No filter triggering?

[noble sonnet]

no

[keen schooner]

The ChatGPT Plus subscription model states:

Access to additional tools such as DALL-E, browsing, advanced data analysis and more

Are the tools such as DALL-E also limited to a certain number? Will I be charged for them or can I use them free of charge if I have subscribed to the model? That I can also use the DALL-E images for the 23 euros, just like the normal ChatGPT entries?

[charred parcel]

it's all included

[keen schooner]

Oh wow unlimited? 😮

[charred parcel]

no

[keen schooner]

What's the limit in GPT Plus?

[charred parcel]

ChatGPT is limited to 40 messages per 3 hours right now

[keen schooner]

ChatGPT Plus or normal?

[charred parcel]

Plus

[undone haven]

can you share a reference link or something for this information (40 messages per 3 hrs)

[keen schooner]

Okay and the limit of 30 is the only one? So after 3 hours I can take 30 more pictures each time without being charged?

[keen schooner]

https://community.openai.com/t/chatgpt-plus-and-usage-limits/544425#:~:text=ChatGPT Plus has a 40,seem to be very short.

[undone haven]

thanks, one more doubt, will the response time for gpt-3 model reduce with the chatgpt-plus subscription?

[keen schooner]

No, you only have a significantly faster response time in GPT-4. Nothing changes in Model 3 I think.

[undone haven]

what is the context window for gpt-4 model in chatgpt plus subscription?

[charred parcel]

32k

[undone haven]

one last question, is there any way to track the 3-hourly gpt-4 usage and remaining requests in that time window?

[tropic spade]

Robert what about the limit for gpt

[rigid cosmos]

I will really appreciate it if someone can help me resolve the quota limit exceeding issue. My card is on file and limit is 120 USD per day. Still facing this challenge. It's not letting me update my card as well.

[tropic spade]

? What do you mean

Where are you running the limit from playground api or chatgpt?

limit per day if for api use and it does not have limits

[lost wigeon]

hey, you are wrong here

gpt-3.5 turbo (base chatgpt)

is way faster with Plus

[lost wigeon]

so it does reduce the wait time, also you get priority so even if the servers are under a big load - you technically shouldn’t wait long for the response

[ruby bane]

Hi guys, any1 with payment issues? I was enlisted on the waiting list and after several weeks of waiting I could click upgrade, but I cannot process it further. It only basically says that my payment cannot be validated etc

I tried my regular bank acc and then revolut after having purchased USD, so no clue if it could be an issue

[hoary venture]

So not just my problem then

[ruby bane]

seems so, glad we are not alone 😂 some appeasement at least

was thinking it might be because of my country being not supported too, but not sure now

[hoary venture]

Nah, pretty sure its there problem

I have asked friend from different place to paid for me but all fail

[ruby bane]

older users seem to be okay, we have to wait then

at least I hope we won't be forced to sign up for the waiting list again

[hoary venture]

Maybe its a good thing, since google is catching up we might dodge a bullet

[ruby bane]

hope they make it fast, I need to write my MA Thesis xdd

not with chat I mean, just using it since it's about it

[hoary venture]

You can use bing co pilot instead it also offer gpt4 and include reference too

[ruby bane]

oh thanks, will use this advice

waiting for them here might make me lose too much time anyways

[vital radish]

you can try chatterclock chrome extension - please note - it is not synced across devices

[polar solar]

Hello! Everytime I try uploading an image into GPT4 I get the following error "Hmm... Something went wrong". I have tried clearing my cookies, cache, disabling all extensions, tried chrome, edge, and firefox with no success. I have also tried using my cellular data and a different wifi network with the same result. If anyone has any other suggestions please let me know!

[rugged wing]

I'm having issues as well with my custom GPT. does the exact same thing. will eventually error out with "Message in Conversation not found"

done many of the same trouble shooting steps as you @polar solar

status page shows everything ok

[polar solar]

yes exactly its so frustrating

[rugged wing]

hahha, and it just burned through all my tokens in like 30 minutes. yay!

[wispy galleon]

Hi there, anyone else noting some problems with chatgpt right now? I'm in Asia
The cloudfare "verify you are human" website always appears
once verified, it loads but never actually loads chatgpt interface
which channel here do people usually report or comment these?

[timber shadow]

Do you have any VPN installed? What browser(s) have you tried in so far?

[crisp bramble]

yo someone ehhelp me

write a prompt 😭

but someone who isnt busy

[timber shadow]

What is your goal with the prompt?

[crisp bramble]

im tryna make it write me a 4 minute comedy skit on a book i read

[crisp bramble]

to be satisfied

[wispy galleon]

Thanks. No, I don't have to use VPN to access ChatGPT. I have tried Chrome, and Firefox. In Chrome, the problem fluctuates between loading the cloadfare humanity verification -mostly- and loading chatgpt interface without being able to log into your account (like being able to ask but not in your account). In Firefox, it loads logged (no problems so far here)

[blazing timber]

What was the GPT that I believe was ChatGPT but it could be ran on your local machine?

[timber shadow]

I'm glad it's working as expected in Firefox. As for Chrome: it can be caused by a lot of things -- VPNs, browser extensions, sometimes folks on Linux have the problem. You said you don't use a VPN to access ChatGPT, but do you have one installed at all? Sometimes some VPN background processes cause issues, even if the VPN itself is disabled.

[timber shadow]

This will probably not be achievable with a single prompt, as it can't output indefinitely long blocks of text. Your best bet might be to:

• Ask it for an outline for such a project
• Work with it in chunks to build around that outline, piece by piece

[wispy galleon]

You were right ... it was a VPN problem. Somehow it was activated. I must have forgotten to killed it off yesterday ... so weird

[timber shadow]

You might ask about running local LLMs in #ai-discussions!

[wispy galleon]

thanks for the heads up

[timber shadow]

You are welcome! Thanks for confirming what worked for you!

[ocean nexus]

Hey everyone, I am not able to upload any image to the ChatGPT for the past a few hours. I've tried different browsers and devices. It gives an error saying "Hmm...something seems to have gone wrong." and regenerate button pops up whenever i try to upload image. Uploading other files like PDF works fine. Anyone else having same problems?

[timber shadow]

Hello, there are some similar reports over in #1070006915414900886 right now. Do you have a VPN installed? What browsers have you tried in so far?

[ocean nexus]

I don't have any VPN installed, I have tried safari and chrome. I also tried the ios app as well

[timber shadow]

As a test, can you disconnect from WiFi on your iOS device and connect to your cellular network, then open a private tab in Safari and sign in at chat.openai.com, and try again there? Just trying to troubleshoot whether it might be a network thing, or a device thing, and a private tab will help us avoid any potentially interfering cache/cookies.

[amber patio]

I just got in, was able to use another card on another account and it went through, but is it normal that i don't see any gpt4 model in playground? (https://platform.openai.com/playground)
Also, How do i access to the plugins? I don't see it on my screen

[ocean nexus]

No luck here as well

[timber shadow]

Hello, ChatGPT Plus, if that's what you're referencing, is separate from API usage and billing. To get access to GPT-4 on platform.openai.com, follow the guidance here: https://help.openai.com/en/articles/7102672-how-can-i-access-gpt-4
As for Plugins on ChatGPT Plus, go to your account settings, go to the Beta features tab, and enable Plugins. Then start a new chat, and change the model to Plugins.

[misty yacht]

Hello! Is there playground for Dall-E?

[timber shadow]

Not yet, I don't believe! I think you can use DALL·E 3 in three ways currently: ChatGPT Plus, Bing image creator, and OpenAI's API. You can use DALL·E 2 on labs.openai.com

[misty yacht]

Thank you!

[amber patio]

Thank you

[grim marsh]

I am trying to subscribe to gpt plus and every card I use is getting declined, they have sufficient funds and the bank says they should work.

[river geode]

How do I access plugins like the spotify plugin? I don't know how to find them in the new UI

[timber shadow]

Hello! Are you trying just normal credit cards from a ChatGPT-supported country? Have you checked these help articles?
https://help.openai.com/en/articles/7232916-why-was-my-credit-card-declined
https://help.openai.com/en/articles/7947663-chatgpt-supported-countries

[timber shadow]

I'm not sure about that specific plugin, but on ChatGPT Plus (web version only), go to Account settings, Beta features tab, enable Plugins, then go to a new chat and change the model to Plugins.

[grim marsh]

I did and I cant find anything on this issue

[timber shadow]

Is it just a normal credit card (not a virtual card, etc.) from a bank and location that are both in a supported country? And billing details all match up?

[grim marsh]

yes

[timber shadow]

You might open a support chat at help.openai.com or make a report in #1070006915414900886. Be sure to include all the detail you can!

[spiral bison]

Do regenerations on failed responses eat up your message allowance?

[timber shadow]

Yes, here's a list of things that count as one. 40/3hrs for GPT-4, but only 25/3hrs in custom GPTs. #community-help message

[winged ridge]

Has anyone been able to use a published google doc/sheet as knowledge? This way I could work on the documentation and not have to reupload it, but instead anytime gpt needs knowledge it would have the latest since it would read it from the link. Presently it says it can't read it.

[trail cairn]

knowledge files are only for static data as they are preprocessed on upload. if you want live data you need to use custom actions and your own logic

[normal gorge]

unable to load history, cleared cookies, tried it on other browsers and computers. What the heck is going on?

[timber shadow]

Hello! Do you have a VPN installed? What browsers have you tried already?

[winged ridge]

ok thx, I do have a plugin called Aaron Docs, seems to work in normal chatGPT,
but I don't know how to add plugins in GPTs since the drop down box is different for GPTs

[trail cairn]

plugins are going away and get replaced by gpt's

[ruby bane]

Yeah btw for people who couldn't pay - I cleared cookies, and by doing so, which I kind of expected, happened to close the doors for good, as the waiting list reappeared as the only option - great 👍

[timber shadow]

Hi, does any of the feedback in this thread work for you by chance? #1183900385698644028 message

[ruby bane]

not really, I will try some things though

been waiting too long to now give up on that one. I kinda need to access it asap. Hope sth works eventually

[timber shadow]

There are a couple specific links offered in some of the more recent messages in that linked thread that might be helpful if your account is on some kind of "invited to subscribe" list, you may just need to navigate to those URLs if that's how it works!

[ruby bane]

Thanks! I was just stupid enough not to screenshot it when I could click "Upgrade". I am not able to find an email with the invitation now, but I most certainly got a notification that I could buy it and tried multiple times

[grim marsh]

nothing seems to work

[timber shadow]

I'm sorry, that's really frustrating! I think your next step ought to be opening a support chat at help.openai.com. You should see the chat icon in the bottom-right of that page.

[grim marsh]

yea im waiting on a response 😦

[winged ridge]

really? that would be terrible since we need to be able to connect things, the current interface is like "make your own plugin" instead of relying on good plugin makers.
Maybe if the chat interface could code the plugin(as an Action) in real time that would be cool

Anybody know how Idea To Code GPT" works? I tried to ask it for it's prompts but they had a hardcoded I'm sorry, "I cannot assist you anymore." response, I'm working on getting around it but ran out of messages for the moment. I want to learn how it works so I can make better GPTs.

[dire heath]

Hey everyone! I was just wondering if its possible to use our custom-trained GPTs in an API call? For example, I have an NPC in a game that responds using GPT 3.5 turbo - but would prefer it to be trained on my own custom prompt and info.

[ruby bane]

Well I guess it's best to wait for a fix, but the part about waiting on a waiting list again isn't optimistic. Anyways, when I now tried to find a solution and made an attempt to log in, it said the servers were busy. Hope it gets fixed and I might get a chance soon to buy the service. For now, it's best to leave it, I suppose. I shouldn't have cleared cookies in the first place.

One solution with the Firefox private browser kind of did the trick, but it still kicked me out at the check-out, FYI

[thorny belfry]

Hello everyone. How can I set up the page chat.openai.com to display several custom chats? For some reason, only two of them are showing, even though I have used about a dozen. From screenshots, I've seen that others are able to display all their chats normally.

[trail cairn]

no, with api you need to use assistants if you want similar

[winged ridge]

that line right under chatGPT in your screenshot means that is just what you used today, try scrolling up?
Also sometimes I have to refresh to get it to show them.

[trail cairn]

select the keep in sidebar from the menu

[thorny belfry]

I can only see my chats below; none of the other custom chats are visible. They only appear in the 'Recently Used' section when I click the 'Explore' button, but this isn't convenient.
I've been facing this issue for a week, so I've definitely refreshed the page several times

[thorny belfry]

Hmm, could you elaborate?

[trail cairn]

from the context menu each gpt has

[thorny belfry]

Oh didn't know about it. Thank you

[rugged wing]

I'm using ChatGPT with a custom GTP. At the end of each line I got blue quotes with the end of the world "source". How I can remove theses mark ? Here's an example

(Hello)

[trail cairn]

there is no setting for it but you could try say it in the instructions to never provide annotations

[rugged wing]

I just tried it without success :

My GTP receive an URL, I'm asking GTP to create an extract from the content of the URL

[vital radish]

tell it not to include citations

[surreal cradle]

I need some help with using open ai api for JavaScript. Whenever I try to use any function and check the error on google console ( I am running my code through netlify for a website) says

main.bed577a1.js:2 Error testing OpenAI connection: TypeError: jp.complete is not a function
at main.bed577a1.js:2:506803
at f (main.bed577a1.js:2:243634)
at Generator.<anonymous> (main.bed577a1.js:2:244971)
at Generator.next (main.bed577a1.js:2:243997)
at ra (main.bed577a1.js:2:249625)
at o (main.bed577a1.js:2:249829)
at main.bed577a1.js:2:249890
at new Promise (<anonymous>)
at main.bed577a1.js:2:249769
at main.bed577a1.js:2:507145

This error occurs for any open ai function I try (instead of jp.complete it says what jp.(method I am trying to use) my code does not contain any mention of Jp so I’m wondering how I can fix this. I’m on a mac running vscode and running my code through netlify)

[trail cairn]

#dev-chat but also its not clear what package you are using

[stuck furnace]

everyday... oh god why

[winged ridge]

I think its a bug, I can usually shake it loose but clicking a previous chat but you may need to clear your browser cache/cookies

[vital radish]

And This - DOES MY head in !!!!

[mint finch]

Is this a good place to ask for GPT advice?

[timber shadow]

Definitely!

[mint finch]

Trying to decide which would be a better follow-up to https://chat.openai.com/g/g-IqTbzKzME-c-c-the-career-counselor

A Cash Coach, Couple's Confidante, or Co-Founder/Collaborator?

[timber shadow]

Ah I'm sorry, I thought you meant advice about how GPTs work! Perhaps you could post your GPT in #1171489862164168774 if you haven't, and then if some users try it out there and give you feedback, perhaps you could ask them what they might find helpful! Otherwise, maybe a general ask in #off-topic might get you some quick feedback too 🙂

[mint finch]

Mahalo much mate. I have it posted over there and tried to poll a month ago but looking for final verdicts before trigger pulling. 🙂

As for 'How GPTs work' here's a general one - the GPT Builder writes instructions in 3rd person but all my prompts have been in 2nd person. Does it treat them equally or are there distinctions worth sticking to 3rd?

(3rd - "It will behave like this" vs 2nd - "You will behave like this")

[dire heath]

Thank you!!

[timber shadow]

That is a great question, and for that one, you might try over in #prompt-engineering -- some smart folks in there! Sorry to keep pointing you in different directions 😄 I just think that one might be better-geared for folks who are already thinking about optimal ways to provide instructions and prompts.

[mint finch]

Of course, I appreciate the direction, so kind Solbus. 🙏

[elfin bluff]

uhm
huuuh

(first message in this chat & first chat on account)

ive fixed it!

[surreal cradle]

I am using npm openai

[hollow copper]

this occurs on ios and pc, I have gotten different IP addresses, incognito mode, different pcs, clear cookies and cache, no extensions, nothing works, how do I fix it

this has occured for over 18 hours, worked fine a few days ago, nothing has changed

[timber shadow]

Hello, several users have reported a similar bug in this post: #1183909233436131358 message
You might add your experience there, or throw it a for visibility!

[lapis pumice]

Hi everyone, my paid account was recently downgraded to the free account. Is there a way to resolve this?

[charred parcel]

did you let your subscription lapse?

[lapis pumice]

No, everything was paid for

[charred parcel]

when was the last successful payment for your account?

[wise lodge]

Is it possible to access a port number with a custom GPT?

it keeps saying my paths are not under the root URL

None of the provided servers is under the root origin

oh, it was the http:// it didn't like

https:// works

[onyx kraken]

You'll need to reach out to help.openai.com to talk to the bot (will connect you to leave a message for people after you pick through choices) to get help through OpenAI.

We can discuss with you and help figure out what happened, maybe. But we don't work for openAI, we're just community members - help for a payment or account status issue has to come from them.

That said, we can help you figure out, maybe, if you just forgot to pay, or if your payment method was declined.

Maybe you don't need to wait for OpenAI's help, like if you can check your payment through your account, maybe you see that your card was declined. Maybe you can switch cards and get started back up without needing OpenAI's direct help - which can take days to get, because of how super busy they are.

[river geode]

Thank you!

[minor locust]

hello , I was generating images for my kids story for a few weeks and my CHATGPT paid account was active since summer 2023, paid version. today I generated a few images, I mainly used Dalle, and when I reopened my account, it said it was deleted by OPEN AI without any explanation, but i haven't done anything to infringe any terms, all my images are kids friendly and i am shocked totally as its work for months that is lost, what can i do and if it still possible to retrieve the account as its an error. I did get some content violation messages 2 days ago when in the prompt i asked to create FULL BODY SIZE (when it cropped the characters) and Dalle understood it correctly but i still got those messages but as i didnt mean to infringe anything I just ignored them. Apart from that I never got even any warnings or anything. How it is possible my acocunt was deleted wihtout any reason or warning, is there is a real chance to retrieve it or all I will get are bot responses about "Read our Terms of USE" which is not helpful as I never infringed them... Please anybody help, I am panicking

[fading ferry]

Anyone know why I can't send images? This happens on my phone and laptop. I believe it's an account specific issue as my girlfriend can send images fine on her account

[timber shadow]

Hi, sorry this happened! You should be able to get in touch with OpenAI support at help.openai.com. It'll start you with a generic robo-support chat but you should be able to get a message to support through from there. It's the bottom-right chat icon at that site.

[timber shadow]

Here's a thread on bug reports with multiple users with a similar issue: https://discord.com/channels/974519864045756446/1183909233436131358

[minor locust]

thank you but i tried that and its a bot generated responses, and ther eis no option of appeal, like the only way is to choose OTHER option and leave a message and then wait if any day they respond?... and if they deleted it they how I prove i havemt done anything?

[timber shadow]

That leave a message option is how that works, I think, yes! Someone should be getting back to you from there 🤞 Unfortunately I just can't speak to how the rest of the process might work for you but I hope you can get some insight on the situation at the very least.

[fading ferry]

Thanks, but from that thread the only thing I can tell is the issue isn't fixed

[minor locust]

but do you think it mught happened just by erorr like a bot deleted the account? and if it deletes it or still it is possible to retrieve the chats if i write to them?

[timber shadow]

Yes that's where it's at right now unfortunately 😦
Throw a ⭐ for visibility or maybe add your experience too if you haven't! OpenAI staff are periodically present in that channel

[minor locust]

is there any way to write them here on Discord? to the support?

[timber shadow]

I'm really not sure at all, I'm really sorry! If you saw some message about it being deleted by OpenAI specifically, my hunch would be that they have something on the back end that would log the deletion reason somehow

[timber shadow]

You can try making a post in #1070006915414900886 but note it's public like the rest of this server so don't share any PII/account details in the post itself. OpenAI staff are periodically present there (yellow names, in their name)

[minor locust]

i also have another issue i made another account instead of deleted to continue working but now Dalee blocks me after only 5 interactrions and asks to wait 1 hour for messages cap, though on my previous acocunt i alsways had at least 20-30 interacrtions , and it happens all the time on the new account. do you know if its becasue its a new account and it will adapt soon or if it is a permanent bug?

only 5 images and it says wait for 1 hour

[timber shadow]

#1184327100396736597 message

[minor locust]

yes I already posted it

[timber shadow]

That is a link to my reply to you in your post 🙂

[minor locust]

i answered there too..

[rugged wing]

can anyone else not upload images to chat gpt4?

[next glen]

Keeps giving me an error every time I try and run it with an image

[rugged wing]

yea same

[next glen]

Do u get this?

@rugged wing

[rugged wing]

exactly the same thing

so annoying

[charred parcel]

Image input is working fine for me

[bronze charm]

Fine here as well

[rugged wing]

interesting

[onyx kraken]

Vision is working for me too.

[gray quartz]

I have this same issue

[glossy wasp]

Hey! Can anyone help me out with their guidance about my project - where basically am building chat-with-pdf app ! i have completed the development but there are something's i need to tweak a bit as per my use case ! would be glad if anyone could help me out !

[charred parcel]

you know you can upload a PDF to ChatGPT natively and chat with the contents?

[glossy wasp]

yes i know ! my project revolves around "uploading bulk pdf's and then chatting with them"

[meager ginkgo]

I can buy the premium only on web, but i want to buy it though apple app store, any fixes?

[glossy wasp]

can i hop on to your dm to explain better .. @charred parcel !?

[exotic shale]

Hey, i bought the ChatGPT Plus but i still dont have it. Does it takes some time ?

[hybrid gale]

HI, I am interested in analyzing food ingredients and cross referencing them for suitability with diet and allergy restrictions. The Data needs to be consistent . is this doable with the api?

[charred parcel]

you should have access right away

[exotic shale]

Hmm well i have to cotact the support i guess ?

[glossy wasp]

@charred parcel

[charred parcel]

you can share here, I am unable to help people via DM

[red roost]

I want to subscribe to the plus thingy, but its a bit expensive for me alone, so i wanted to see with my own family if we maybe share and use it together, would that be okay? because its my own family and i can trust them (i know that i will be responsible for everything they do with my account)
in the TOS of OpenAI is said that i should not share it with anyone else and stuff, but the way its written it sounds to me like this: "Do not share your account, but if you do, then you will have to live with the consequences" so i actually can do it,but if something happens, its my own fault (i will not give them the credentials, i will just let them use the device the account is on, so technically i never gave them the credentials, i just logged in on an device they use, also its only for my family)

so my question is, is it directly illegal to do that? or is it just a thing of, i can but if something happens, its my fault?

[charred parcel]

I don't see how you can avoid using GPT 4 through the API for such an application, but as you say it's expensive

[charred parcel]

the best advice you can be given is not to share your account. We see people comes in here who have lost their accounts due to TOS or Content Guideline breaches, you don't want to deal with that

[red roost]

oh welp... was thinking that maybe because its my family and i can trust them it would be okay

[glossy wasp]

what you suggest dude?

[charred parcel]

my suggestion would be to rethink your business model to make it profitable, can that be done? I don't know.

[red roost]

then i think i need to find a way to get enough money so i can get it sadly... but thank ya for your help

[meager ginkgo]

Hi robert, i got accepted to buy chat gpt plus but i can only do it on pc, could you help me solve the issue

I want to purchase it with the in app feature

[charred parcel]

do you have an iPhone?

[rugged wing]

It works, thanks !

Ho no, I got theses again 😦 I ask to not include annotation, references or citations but thoses quotes stills remains

[vital radish]

try to add the word "strictly" ... do not include references, citations or links, and then see ? They are all basically the same thing, and also check for any ambiguity elsewhere in your instructions. I have found any ambiguity will cause it to skip parts of your instruction. Also check your spelling, that can also cause confusion, even though it should be able to decipher spelling mistakes

[rugged wing]

I lost my one of chat data. can it fix? I wish recovery the data and start chat again.

[charred parcel]

how did you lose it?

[rugged wing]

when I tried to regenerate it on iPhone chatgpt app, it failed to access, and I access by PC, then the page never load any logs and chat page whiteout.

[charred parcel]

which browser and OS are you using?

[rugged wing]

i use iOS 16.6(iPhone) and macOS is Ventura, browser is Safari.

[meager ginkgo]

Yes

[charred parcel]

try clearing your cache and website data for openai.com and log back in to ChatGPT

[charred parcel]

are you able to subscribe from your iPhone?

[meager ginkgo]

No, that’s the issue

[fiery glade]

Anyone else experienceing issues with external APIs/Actions in custom GPTs? It just stopped asking me for approval and always retruns the error "message": "The requested action requires approval",

[meager ginkgo]

I want to subscribe with my iphone

[charred parcel]

you're the second person I have come across who couldn't subscribe from the app even though selected from the waitlist. It seems likely that the only way to subscribe is using a credit card on the website

[rugged wing]

i tried delete cache and website data, but it looks like same status. other chats are working normally. I know, the lost chat had many logs and data volume was large.
Is there anything remain for me to try?

[charred parcel]

nothing that I can think of sorry, it should be showing in your chat history

[rugged wing]

The chat screen still does not load, is blank, and cannot be moved. Is it just running too slow? Is there any chance it will show up if I wait?

[charred parcel]

I don't know, sorry. It sounds more like a bug to me. Maybe post about the problem in #1070006915414900886

[rugged wing]

Thank you very much.

[tender root]

Anyone else having problems with failed prompts, error, plugins not working, gpt4 being extraordinary slow?

[charred parcel]

all working as expected for me

[tender root]

hmm... i get so many errors the past 2-3 days, my gpt4 keeps extending its "pause" time... Can it be my own system that is messing witg me?

[charred parcel]

maybe, what OS and Browser are you using?

[tender root]

I am on mac os venture v13.5.1 and i use google chrome. But my computer is a bit messed up but just havent had so many problems before 😂

problems with chatgpt before*

[charred parcel]

have you tried using Safari instead of Chrome?

[tender root]

Not the past couple of days but i do use safari for certain things. I'll try it and see if theres any changes! thanks for the advice!

[rugged wing]

I added "strckly" It works fine for the moment. Thanks

[oblique perch]

Hey everyone.

I have always logged in to my OpenAI-Account through Microsoft.
Is it possible to change that, so that the account is no longer connected with Microsoft, so that I can login with my email?****

[rigid musk]

Hi you will have to contact support on help.openai.com

[oblique perch]

Thank you for your quick help!

[tacit wing]

How can I get my GPT to always look in the knowledge base for answers? I'm having a hard time convincing it - both in GPT instructions and the prompt to do that. It keeps answering without looking.

[cobalt leaf]

hi guys/girls, have you my same problem of GPT4 that is VERY VERY slow and less effienct than 3.5 version?

[vital radish]

You need to instruct it to look there first. Also ensure you don't give it any ambiguous instructions elsewhere in your instructions else it does weird things.

[tulip lark]

anyone elses GPT-4 not analyzing photos or files anymore? I cant figure out what the problem is. it was working earlier this week but has stopped working over the past few days

[timber shadow]

Hello, do the issues you're experiencing match the issues described in this thread? If so, you might add your experience to it: https://discord.com/channels/974519864045756446/1183909233436131358

[tacit wing]

I thought I did by mentioning this in the prompt:

Always look into the knowledge files first before giving an answer, and do not defer questions to any other authorities.

Can it be written better?

That's what the upload area is called. Should I instead reference each file I uploaded individually?

cheers, I'll try that

[steep quarry]

hey yall

I got a question: is it possible to have chat gpt act as a liason between me speaking and writing notes in notion

IE, I dictate into my phone and the custom gpt creates a new entry in a to do list lets say?

im just getting so lost at the schema and actions and stuff

[dusty holly]

sure

[steep quarry]

ok so I have chat gpt plus, I have a notion account from which I assume ill need the api key... How do I build a schema? Is the format for the open ai schema the same as other schemas (IE can chat gpt help me write it?)

[dusty holly]

depends on how you want it to behave as well. easy way is using existing tts and stt

[steep quarry]

i want it to be quiet and not verbose

[dusty holly]

di you tell gpt build thats what you want to do and see if it gives the info

[steep quarry]

yeah it clearly didnt comprehend what i was on about

it told me to use a note taking app... im just wondering if i just tell it to make a schema, woul that work, or is the open ai schema specifically formatted

the behaviour i want is super basic. I just really like how quik and accurate whisper is and having it on my phone....

[dusty holly]

building it now. i dont use notion but if it looks good you can test it out. im desiging it to ask the user for their key per session. may need another method. if i get fed up ill pass you back the work

[steep quarry]

Dude

[steep quarry]

You’re the best.

So wait, in the gpt creator thing can it write its own schema?

[dusty holly]

so ther needs to be a plugin built for it and although i do develop things i for some reason cant make plugins or it would be done. the other option is a propject doing so that is good to go

[steep quarry]

If it’s for personal use only I should be able to hard code my app key in no?

[dusty holly]

check messages for arepo that has your tool

[steep quarry]

Yeah so I found that already

That’s chat gpt inside of notion

[dusty holly]

it makes mroe sense to do it like that as well to enure your API key is local for notion etc. if you can find a plugin for it then that will work you can make the gpt or give me the link to complet the one i have done alrady

if you can make plugins, make that and you are golden

[steep quarry]

Yeah if I have to enter my key every time it defeats purpose

[dusty holly]

mm hmm

well, at least you have options? haha

[lavish cave]

anyone else getting this issue?

[cold folio]

Not here. How does it happen?

😊 .

[lavish cave]

just whenever i talk to my custom AI or regular GPT. its been happening for days now

if i get rid of a good few of my historical chats would that help?

[cold folio]

No thats not related to it also there's no official rule about history making it slower

[cold folio]

i think your issue is related to excess of uploaded files on your custom gpt

[lavish cave]

there are no uploaded files on it and also its happened with the regular one too

[cold folio]

custom Gpts are still experimental. i'd suggest making a new one with the same content and slowly reupload the files to check where the error is

[cold folio]

Ohhh! well that's weird

[lavish cave]

looks like its time for some percussive maintinance /j

[normal gorge]

tried all browsers, edge, chrome and multiple computers. No I don't have VPN.

[rigid musk]

can you try firefox

[normal gorge]

I did. It doens't work.

[abstract kelp]

I can't access a conversation, it says: "Unable to load conversation 6[...]0c" any idea what happened? I'd need to recover this conversation

[cold folio]

try logout and then login back. it fixes for me

[abstract kelp]

I tried and it didn't

[cold folio]

are you on mobile app?

[abstract kelp]

no, I'm in browser. I tried Chrome and Firefox, same error. In the app (iPhone) I access the conversation but just the last message, when I type it crashes and closes the app (if I type in any other conversation it works fine)

[slate fractal]

I'm receiving network errors mid-response.

[normal gorge]

I have the same issue. Web apps on any browsers and any PC aren't loading history; Iphone App loads what is created on the iphone but leaves a gap in the history.

[rugged wing]

I have the paid version how do I increase the limit on how long a conversation can be

[vital radish]

you cant

40 messages every 3hrs

and set limit of input and output tokens

[pale hearth]

Even though I subscribed to Chat GPT Plus, I can't use the voice chat option in the settings. The option doesn't appear at all. What is the problem?

[timber shadow]

Hello, the voice feature is currently only available on the iOS and Android apps -- is that where you're checking?

[pale hearth]

yes, i'm using my Android apps.

[timber shadow]

Do you have any VPN or adblocker app installed?

[pale hearth]

no

[timber shadow]

When you go to Android settings, Network and internet, Private DNS, what do you see as your current setting?

[wise lodge]

How do you get a custom GPT action to use port numbers? I used the following syntax from the openAPI documentation but it doesn't seem to be specifying a port in the request and I never see it hit my server:

      enum:
        - "4002"
      default: "4002"```

[pine elk]

I'm seeing infiite loading screens on most of the platform.openai.com pages, however don't see anything relevant in status

[agile magnet]

anyone else experiencing constant network issues when generating a response or not being able to generate photos? It's been very consistent for me lately.

[wise lodge]

Nevermind, got my router to forward inbound SSL traffic to my API port! \o/

[umbral gull]

Every prompt Im writing on a costum gpt or abny gpt I use I get random words and emojies back linke this: https://chat.openai.com/share/53ceb25b-e9fa-474f-854f-1757cc6a4c3a. or this. https://chat.openai.com/share/a899a288-eae9-40a4-867f-dd65ef03c367

https://chat.openai.com/share/b472f0a7-86ee-423f-80c0-7e9b4c4ac0e3

[foggy musk]

hey if i send chatgpt 4.0 multiple pictures in one message, does it count as multiple prompts or just one? (towards message limit)

[willow stratus]

Estagiário desligou o servidor do ChatGPT?

[rigid musk]

english only on the server, please

[red flower]

Server down

[rigid musk]

status.openai.com

[tropic spade]

wth

chatgpt is down again ?

[red flower]

Gemini is attacking. lol

[tropic spade]

wouldn't be surprise that is what google has done all these years

buy competitors then crash them

so no need to work today if we use chatgpt for work ?

lol talk about reliability this system is a joke

[rigid musk]

running
99.77 % uptime
this systeme is a joke

[tropic spade]

can only use it for leisure for for anything serious

[foggy musk]

no response on this?

[tropic spade]

99% uptime ? where you get that data from ?

[rigid musk]

status.openai.com

[tropic spade]

oh really ? i'm a user and i do not see that happening

hahaha

i wonder which is more accurate user or dashboard ?

[rigid musk]

dashboard for sure.

people coming here saying chatgpt doesnt work and after troubleshooting often it is in their end

[rigid musk]

think it is count as several, not sure

[tropic spade]

that would have been compatibility issue too

if a user has been using chatgpt for some time and things started happening how could that be a user issue ?

[rigid musk]

people installing lot of third party thing and wondering why it doesnt work anymore

vpn, addon, extension,

[tropic spade]

i see so it's not a chatgpt compatibility issue it's all the other software issues right ?

i see

it's never openai or chatgpt issues lol

[rigid musk]

error can happen, but for the last 90 days uptime 99.77% never had 1 issue

[tropic spade]

you seriously is brainwashed dude once there is an error we had to like wait for hours

and there is limit issues that we cannot be using chatgpt for 24hours

if an error occurred during work hours we will have to wait out and because of the error a lot of limit has been wasted trying to revise the reply from chatgpt

and you still think 99.7% is the right way to calculate this

even 1 hour can cause a lot of disturbance to users

[rocky whale]

i bought chat gpt plus and it still shows the normal chat gpt for me is this normal?

[tropic spade]

i guess they are not so interested about user experience lol this is just a toy....

[rocky whale]

where cna i report it

[tropic spade]

i am not sure maybe @rigid musk can help?

[rocky whale]

@rigid musk can you help?

[tropic spade]

Oops so much for helping lol no UX for you .....

[gloomy glen]

is gpt4.5 comming out soon??

[rugged wing]

been having some trouble with conversations not loading.

A chat I'm using becomes unresponsive, then I end up missing 99% of the conversation.

I never had this problem before, I have to end up deleting that specific chat because I can't respond to it anymore. Does this sound familiar to anyone?

[subtle crescent]

same with me here

[rugged wing]

Any solutions?

[solid epoch]

guys i have a problem im playing a rol game with chatgpt and the page is so slow cause the quantity of messages sender can i solve that?

[full hollow]

No. I don't.

[tight mortar]

Downtime is downtime, but when there is no downtime ChatGPT is perfectly usable. A lot of people come here and complain that they get errors on ChatGPT but they have local issues causing problems for them. Also, the API availability is great as well when there is not a status.

Uptime is exceptionally good for a service at this scale.

[tight mortar]

You also shouldn't be relying on ChatGPT for work and then citing reliability issues if you can't do work without ChatGPT.

[tropic spade]

How can you say this we PAID for the service. You mean are okay with Microsoft or ios not working with your computer at work ? As you indicated we should not rely on chatgpt for work?

You seriously is answering for the sake of answering and not looking at the subject matter

A product that disregard user input is the worst

Bigot company

[tight mortar]

Yes, I stand by my statement that you shouldn't rely on ChatGPT for work. You paid for a service and the service had some downtime, that is included in the terms of service for the product, and this happens with various other companies. For example, if Github has an outage and Github Actions goes down, yes, people people will complain but it will come back up and be reliable, that is how products, corporations, life works. You also mentioned "microsoft" or "ios", there have indeed been microsoft and apple outages in the past, so not sure what you mean by this.

There is no such thing as 100% availability

[tight mortar]

User input is heavily regarded and is evident from how OpenAI does RLHF, they also work around the clock to make sure the service is stable and good for everyone to use

OpenAI is not out to "get you" or be portrayed as a "bigot company" as you say, they are simply trying their best. If you are not happy with conduct, you can vote with your wallet

[onyx kraken]

Hey, maybe you don't realize. We're community members, just like you. We don't work for OpenAI, nearly none here do - the few who do have gold names.

Also, there's #server-rules here. I'm not a mod, but

" Rule 1  Be respectful.
Treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.). Keep the negativity to a minimum."

Calling the company what you did means you think it is: "someone who holds opinions about groups of people that are based on prejudice or ignorance, rather than on reality"

You're welcome to #1070006151938314300 or #1070006915414900886 if you choose, there's some evidence that OAI checks those more closely than the free flowing community threads, like this #community-help one, where it's really community members trying to talk to each other. Why throw hate at OpenAI at all?

[raven gyro]

can i trust chatgpt 4 with math?

[nimble vortex]

when will openai stop being a closed source for profit buissness and go back to open source and making ai accessible to everyone

[onyx kraken]

if it is able to use Python tools or other math-heavy service, like that one that starts with a W it is often mostly reliable.

By itself as a LLM, no, they're not trained on math.

If the math is super advanced, maybe not even with the tools, check careful, even if it isn't math. It's an LLM, they even say at the bottom of the browser window, "ChatGPT can make mistakes. Consider checking important information."

[raven gyro]

aight

[tropic spade]

alright i see most of the users here kept quiet about what i am saying and i shall stop talking in discord wish you all the best

[umbral gull]

anyone knows why chatgpt is acting super weird the last few days all costum gpts and plugins does not work getting weird answers

getting weird resoinss like this on all and it just goes on until yuou stop

[rugged wing]

can i get help i am hopeless

[onyx kraken]

Hey, whatcha want help with?

[twilit grove]

I've been having problems viewing my old conversations in ChatGPT (it takes a while to load and showing up blank), and the same problems happens on the other recommended browsers as well. (ex. Firefox says that the ChatGPT website is slowing the Firefox browser down) I tried the recommended steps (logging out and back in, clearing caches) but it's still not working; it only started today. I can access it via the app though.

[potent narwhal]

Me too

A terrifying image was inexplicably generated

I was just asking gpt to help me check an sentence grammer

[tough crescent]

Good afternoon, am I the only one who can’t open chatGPT via VPN?

[fringe stratus]

hello just upgraded my account today and i am getting captcha challenge with every question now on chrome this issue doesn't happen when using firefox or the ios app

[cinder pond]

i recently only had a adblocker-supporting dns profile and even that got in the way of me logging into chatgpt

[hidden carbon]

Hi, I have filled in the form the remove the limit of 30 messages every 3 hours for GPT-4, Unfortuntely I haven't received any feedback. Can someone help me ?
Thanks in advance !

[obtuse aurora]

Incorrect, I use NordVPN all the time.

[obtuse aurora]

I've done the same, several time. Don't count on them ever getting back to you. They have to get thousands of those requests daily and if you don't fit whatever criteria they want, then it's never going to happen.

[rigid musk]

Vpn are known to cause issues. But they are in general functional

[subtle tendon]

Last hope! Does anyone know why the toggle for my name isn't showing? I subscribed to plus from the iOS app, think that has anything to do with it? I did, however, get my billing information added using the website in like January.

[timber shadow]

That's odd! Have you tried multiple web browsers?

[subtle tendon]

Oh, didn't think about that. But I guess, Safari on the phone and Chrome on computer is not enough?

[subtle tendon]

It's the same using Edge

[timber shadow]

Edge and Chrome are both Chromium-based browsers, as a test, can you try a new install of Firefox? It's non-Chromium.

[subtle tendon]

Tested, not working unfortunately

[subtle tendon]

Also tried the following: logging in and out, clearing cache and cookies, incognito mode, deactivating addons and adblockers, updating my billing name, waiting

[timber shadow]

Shoot! You might wanna make a post in #1070006915414900886 or open a chat at help.openai.com -- I'm just not familiar with this one, sorry!

[subtle tendon]

No worries, I already have a ticket open 🙂

[subtle tendon]

Thank you anyways!

[indigo adder]

Any idea how to deal with an issue like this?
"Unable to load conversation 3x73ae13-9d1b-49b2-a6c7-31b7929764f7"

[rustic marlin]

the conversatiuon was likely deleted

[indigo adder]

by whom?

[rustic marlin]

the sole user of the account, I hope

[indigo adder]

I am the sole user of the acc and have not deleted anything. I was in the middle of a chat when that error occured.

[rustic marlin]

weird, if that is the case, try to log out and in again, maybe clear the browser cache

[indigo adder]

Tried that, I accessed my acc via a new pc, on a new internet connection. same error.

[steep quarry]

hey all

quick question if anyone knows.... so I changed the voice in settings on the ios app and now i cant get back to the original voice, which was way better. its not listed???

if anyone knows why/a workaround id appreciate it

[opaque zinc]

Hello, looking for a way to disconnect my OpenAI PLUS account from Google account I currently auth through. When I am trying to log with password I get the message that I have to use "original login method" e.g. google oauth, but I want to login directly with my password, this is rather Auth0 related stuff, not sure whom to seek for the help with, maybe someone of you might know how to solve this issue, thank you in advance.

[timber shadow]

Hello, your best bet is to try to open a ticket with OpenAI directly at help.openai.com -- click the chat icon in the bottom-right.

[oblique oar]

Hi all. I just created my 3rd GPTS about an hour ago. Was using it for a while. all good. then now when i went back to check, its vanished! I can see my other 2 gpts, and i can even see my history conversation with my new one, but its vanished from my list of GPTS. Anyone know why this would happen

[tacit timber]

can somebody please tell me when the microsoft openai team will resolve the criminal anticompetitive language in the microsoft openai terms of use? asking for 8 billion stakeholders https://discord.com/channels/974519864045756446/1184196946496331896

still up 2023-12-14 despite contacting literally everyone in a leadership position at your business and the department of justice, and the ftc, hold my beer while i forward the message to the office of the attorney general of the state of california

[tacit timber]

which draft do you think is the strongest?

[trail cairn]

has your lawyer yet contacted them? also this discord is for openai while microsoft is its own company

[tacit timber]

I refer to the company headquartered in San Francisco, California known publicly as / operating as "OpenAI" as MICROSOFT OPENAI because they have been de facto acquired and transformed into a completely for-profit business by MICROSOFT (A Washington State Business) in stark contrast to the purpose and charter of the OpenAI Non-profit as founded in California.

(the opening paragraph of my first draft)

surely you saw this on Bloomberg?

[rigid musk]

What ?

[tacit timber]

Who owns the majority of Microsoft OpenAI? Answer: Microsoft owns over half of Microsoft OpenAI. Only Microsoft violated the Customer Non-Compete Clause to develop Phi-2 using OpenAI model output.

[trail cairn]

why are you asking microsoft stuff and why did not complain about openai tos but specifically microsoft

[whole pulsar]

Hi all I've been a GPT4Plus user since April.
For several weeks now, I've had to systematically fill in a captcha to check that I'm not a robot on my macbook M1, every time I send a prompt. I've deactivated all the extensions, tested different browsers, tried connecting to another wifi or 5G Smartphone, emptied the cookies, browser cache etc. and to no avail, the problem comes back almost all the time. I don't have this problem on my smartphone or on other PCs. I've seen a forum that talks about this subject but I haven't found an answer by reading all the messages. Can you help me, thank you very much? https://community.openai.com/t/gpt-4-unusable-do-to-continuous-verify-that-you-are-human-captchas-requests/325583/39

[tacit timber]

As far as i can tell, "OpenAI" was already bought and paid for by Microsoft and transformed into Microsoft OpenAI. If you have a reasonable argument that owning over half of the business, and integrating GPT-4 into bing and copilot and azure is not clear evidence of ownership, then I'm open-minded to hear it, but as far as I can tell from extensive research, OpenAI is effectively "owned" by Microsoft.

[trail cairn]

where is the proof for majority ownership? previously its only been along the lines of investment to gain full access to the models?

[rigid musk]

Openai is openai. It isn't own by MS ? Ms invested billion and they can use openai technologies

[tacit timber]

unfortunately, link posting does not work in this discord

[rigid musk]

Also that's nothing the community can help you on ?

[rigid musk]

Do you have a vpn ?

[cosmic crag]

"While details of our agreement remain confidential, it is important to note that Microsoft does not own any portion of OpenAI and is simply entitled to share of profit distributions," said company spokesman Frank Shaw.

[whole pulsar]

no

[trail cairn]

then tell reference id where cma/ftc has approved the acquisition

[tacit timber]

my question persists, which version do you think i should send to the california attorney general?

[trail cairn]

majority ownership acquisition requires various clearances in USA/EU so tell some of these

[tacit timber]

i did

just missed california

[trail cairn]

USA/EU level organization, not a single state

[tacit timber]

it's a california company in (alleged) violation of california law

[cosmic crag]

1. this is not anti competitive. This is them giving a policy on how to use their product. Why would you expect that openAI would allow anyone to use their own product ot make a competitor?
2. This is not for this channel

[rigid musk]

Can you try firefox as test ?

[tacit timber]

how is this not about our community?

[rigid musk]

It is more like question about openai products and we trying to help each other as community ( member to member)

[tacit timber]

well, i'm a community member needing help from other community members to understand this clause

[rigid musk]

Just don't use the output to create another ai to compete with openai ?

[cosmic crag]

I just told you, It is ridiculous to think you can make a competitor to OpenAI, by using openAI's product

[tacit timber]

*you believe it ridiculous

[whole pulsar]

same problem on firefox, safari...

[cosmic crag]

How is it not ridiculous

[spiral bison]

Lmao why would a private entity that invested billions into their product let you use it to create competition in the world we live in.

[tacit timber]

explicitly anticompetitive language in agreements makes them illegal and unenforceable

[rigid musk]

I suggest to post in #1070006915414900886 and contact help.openai.com they might double check with you

[cosmic crag]

It is not anti competitive

[spiral bison]

How is them not letting you use their product in that way thwarting you from creating competition? Just go make your own model.

[tacit timber]

it's not about that, duh i can do that, it's more about this, and also the fact, i paid them and piloted gpt to produce those outputs, but they say i can't share those outputs with e.g. bard, claude

[rigid musk]

By curiosity are you using starlink ?

[rigid musk]

Just contact help.openai.com and ask them. You will have your answer

[spiral bison]

How did you pay them? Do you mean you bought a subscription to their services?

[cosmic crag]

It does not say "you may not develop models that compete with openAI". It is a policy on how to use THEIR PRODUCT

[whole pulsar]

no, I'm in France

[cosmic crag]

You are being very paranoid and ignoring the many competitors that actually exist and have nothing to do with openAI

[tacit timber]

*contractual consideration

[spiral bison]

Did you just come here to argue out of frustration? Because you're not making any sense here and you're being fairly disingenuous.

[cosmic crag]

Oh a contract on how to use their product. How anti competitive. And you are asking an AI about legal terms. Bad idea

[spiral bison]

GPT already powers a lot of products that compete with it in different fields.

[cosmic crag]

GPT is a type of llm model. OpenAI does not own the model type.