#gpt-models
1 messages · Page 177 of 1
someone locket in to my acc and everytime i changed it he is still in there what can i do
How long before creating the application is feasible without programming skills thanks to ChatGTP? At least the first versions, for proofs of concept or local minimum viable products.
Frankly, I don't know how you can undertake it without feeling out of context 😂
I sent you a DM. I hope you dont mind
Until now, if I wanted to think about launching a proof of concept, I thought about hiring a flutter freelancer, but today, I don't know which profile would be right.
I wonder how effective it will be in detecting products to sell on Amazon? That is, for qualitative data analysis
Good vpn and some luck
q
Does Turnitin works accurately than gptzero?
Sadly to say there isn't any luck involved
Hey, guys, any suggestions how to get shortlisted in the waitlist?
Contribute on the evals repository
That seems like quantitative analysis, but you need current data, which means access to datasets. It can probably interpret a dataset reasonably well, but without it you won't get far
Hello, Can someone help me, I always get error after some replies, cleared cache, converstations, disabled extensions etc. Still recieve it after some replies.
Thank you.
I understand! But that's incredibly powerful per se, yes, it's an aid to making the right decisions once you have the data collected
what is the exact error message you get?
let me get it .
Definitely. Gpt is a really good intermediary tool because imo the best thing it can do is take nearly any unstructured data (request, text, whatever) and turn it into structure, even if you don't know the end structure. So that applies super widely
Hi folks! Does any have acess to the plugins yet? I signed up for them on day one and still nothing.
There Was an Error Generating a Response.....
many people are waiting still, patience is key
what browser are you using?
Chrome latest
can you please try Firefox?
Sure
Is there a good explanation anywhere of best uses for System and Assistant messages or has no one really figured that out yet
with gpt4 the system is kinda injection attack safe instructions that users cant override
*The system message helps set the behavior of the assistant. In the example above, the assistant was instructed with "You are a helpful assistant."
The assistant messages help store prior responses. They can also be written by a developer to help give examples of desired behavior.*
Yeah I was thinking more rigorous examples. These aren't super useful, it's hard for me to tell the difference also
Maybe on the instructions end, like enforcing output
The "you are a helpful x" is minimal, feels like there is much more you can do with it
yes that sets the behaviour eg. you can say "you are a sarcastic pirate, do this, dont do this" etc
Interesting. if it can accomplish that much better than just user message on the 3.5 API as well would be great, forcing output structure is not easy on chat
guys tell me the best browser that i should use
Much easier on 4 in user message anyway
My account has been blocked, and OpenAI still deducts my $30 plus fee per month. How can I cancel it? I can not log in to my account. I tried sending an email and contacting on the help.openai.com website, but it has been a month and I still haven't replied. What's the difference between this and robbery! Who can help me cancel this plus plan? I don't want to be deducted $30 per month anymore!
you probably need to call your bank instead
^ and you can provide documentation that you didn't have access for X months due to being blocked and you can get some monies back.
email support@openai.com
anyone know the character limit for a message in gpt 4?
Since last month, I have sent six emails without any response.
how to generate images?
I did not receive an effective solution from the bank. I plan to go again next week, but this issue was caused by the openai team
In api, 8000 tokens typically around 20-24k input+ output but depends on the text too. Unusual words or (from my experience) lots of whitespace produces more tokens
thanks 👍
@vale summit Was it because it was an Asian account?
are you able to use https://help.openai.com ?
and have you tried emailing refunds@openai.com ?
can chatgpt can be online again for Italy ?
@digital bridge No, it is Oceania's account. but I use Asian VPN. So it was banned
Italy has blocked ChatGPT, so until that is lifted, no
LLL my country for doing that
No, will I receive a response when contacting this email(refunds@openai.com)?
yes, you should be refunded and any further debits should be stopped
@vale summit You cannot use an Asian IP address, you need to change it to a North American one.
Thanks, I will try it
@vale summit Are you Chinese?
Yes, I only found out after I was banned
Any update on when plugins will be widely available?
@digital bridge Yes
@vale summit wo ye shi a
@digital bridge en
I have sent an email and hope to receive a reply
Does anyone get "You've reached the current usage cap for GPT-4" even tho the only used GPT-3.5?
this is weird, are you really sure it was 3.5?
the chat changes to 3.5 when you run out of 4.0 usages
Yep, all the replies have the green logo and are fast.
As well I had GPT-3.5 select
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
Guys I'm subscribed to Chatgpt+, but I keep getting the
"Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
message. Is there a different link or something to overcome this issue? If it happens once or twice it's not a problem, but there are times when it happens 9 out of 10 times for hours, and it's really burdensome
disable vpn or try firefox
Try logging out and log in I do not know why but it seems to fix it for me
Yes weirdly it does sometimes fix it for me as well, but not always, it works maybe 30% of the time
I'm not using a vpn and I tried chrome, brave, opera and bing. I haven't tried firefox specifically, but trying out various browsers didn't fix the issue. I guess I'll download firefox and try with it
one of the issues is in chromium and all the browsers you listed are chromium based
Okay, I didn't know that, giving firefox a try now
man the filters are rough these days. i use chatgpt to create prompts for midjourney, and half the prompts are rejected because of "sensitive situations".
Firefox actually works, I'm positively surprised. Thank you a lot!
Is Auto-GPT cutoff by 2021?
Would it not be the same cut-off as regular gpt3/4?
It just prompts itself, it's rare for it to achieve desirable results
Altough I have a gpt plus acc
Look in #1019652163640762428 for a post on it.
That has absolutely nothing to do with the API though, which is what autogpt relies on.
Congratulations in that case, me too.
have u checked out auto-gpt guthub?
There's literally a post about it in #1037561385070112779 my dude.
majority ppl that mentioned auto-gpt here, say it is just overhype
It is.
actually have not seen any one that said positive stuff for it, but well 🤣
takes some time for ppl to find the max potential, just like some ppl said chatgpt was nothing special than a chat bot
well sorry "dude" im not here often
I told you earlier
Hello nerds!
And?
Im not a nerd 💀
j
And you're asking me if I saw something I already referred you to.
allahakbar
then this hello wasn't meant for you!
test
and? ...
Sup?
Okay!
doing alright
Keep it respectful, please
yes sir :)
Yeah course. Just kind of annoyed about this auto gpt hype
Isn't that better for #dev-chat anyways right?
Cannot find it.
Whats the name?
if it's about gpt-4 bneing the brain of autogpt.. here or #ai-discussions ¯_(ツ)_/¯
It is a good technical demonstration of viral marketing, without being able to do very much of anything
How can we use chatgpt 4
https://chat.openai.com and click on "Get Plus" below the history side panel
Thank you guys
Hi I am new here
interesting you guys say it sucks so much... it looked so dope on tiktok lmao
Has the model of ChatGPT been updated to GPT4?
GPT-4 is available from within ChatGPT for $20 USD per month
When i try to buy chatgpt, it declines my card every time, why?
i tried with 3 different cards
what country are you in?
that might be the issue, only support@openai.com will be able to help with that
Okay! thank you so much!
try vpn into italy
in italy chatgpt is blocked
If i wrote to openai chat, is the same or do I need to write to this email?
write to that email 🙂
ty
works in kenya
i wanna go to kenya
Hi all. I use chatGPT 4 (plus subscription)
After message cap limit reached, I wait related time. But I can not switch conversation from 3.5 to chatGPT 4. I am at the same chat conversation, downgraded to gpt3.5. I am searching a way to get back chatgpt4 again for the same conversation. Any idea ? 🙂
Once it switches you there is no way to go back, it's best to not consume your GPT-4 quota in a conversation you want to continue in GPT-4
Thank you very much @hallow owl
is there auto-GPT that I can use online available ? or support to build locally by gitlab autoGPT open source ?
Hi new here! Do we have a chat where I can input images to do analysis?
You need to have gpt4 api
you needa apply gpt's api
Not a public feature yet, and there's no bots for using the AIs in here.
Ask in #1094133504402522142
We can't really help you.
I do have gpt4 and I’m on the waiting list of api alpha
Thx
Then wait in line slowly.
the api list for GPT-4 isn't necessarily first-come first-serve. I've heard multiple people signed up for it day one (including myself) and still don't have access to it
Same here, signed up day one, still no access. Same thing with plugins
I think it’s based on your field of expertise, if you are a developer or potential client you’ll get a faster approval
I haven't contributed anything to evals though, and I rarely used the GPT-3/GPT-3.5 api 😅
I think it depends on when you first subscribe to plus.
I have gpt4 api
and I also have copilot X
Because someone I know who’s registered as an entrepreneur/ developer get access on the second day
no no no
shouldn't have told them I'm a student! They looked at my application and were like "nahh, he won't pay enough" 
No Nono I subscribed on plus the first day
when
they're classists lmao
2.2?
Mid March I think
?
The earliest subscription plus refers to February 2nd.
I think I only paid twice for the monthly fee
Oh I guess I was talking about the public release
I have paid three times.
ChatGPT Plus went public Feb 2, 2023 👍
you just go through twitter finding all the names of people that work at OpenAI and then in your API application you compliment them on their choice of clothes and how great they look today... ez
yeah, even if you approach microsoft, they are only interested with use cases that has big enough traction to use their azure openai
I mean, getting api depends on when you first subscribe to the plus service.
not true
I don't think that's the only reason though 
you can get gpt-4 api access without plus
But I have gpt4 api.
previous api usage and contribution to evals are likely other factors
Yes the plus was on feb
I paid on February 2nd.
Just activity and and use case, I think. I signed up to OpenAI back in 🤔 july of 2021, so that helped, but I'm still waiting on Plugins...
I got the api of gpt4 on March 20th.
I started using chatGPT at the end of November.
I agree with this statement. I also think evals contributions play a role in it for some users
I really don't think it's based on when you signed up for ChatGPT-Plus
I dont think plus takes a role, the api is more for dev and companies
I know someone who has already used Plugins. It's amazing.
plugin waitlist maybe
I'm so jealous of the volunteers with plugins access 😅
But for me, I'm a graduate student of JHU.
I think it’s based on your contribution too the friend I mentioned only registered plus late March
no a developer
oh don't remind me 😔
I'm also in the waitlist of plugins. I tried my friend's plugins.It's amazing. I only have the api of gpt4.
I'm writing my thesis in autogpt now.
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
I’m on both api and plug-in waiting list
just have more patient and take things slow, when it get released, it will overwhelm us (self comforting)
Yes autogpt and agentgpt are my new fav
Also try gpt and markshow for generating BP and ppt
waiting to see if anyone would build something useful from autogpt/agentgpt etc
all of the indicators are there for Plugins to get a wider distribution soon
(keep telling yourself this)
Lol
literal manifestation
In all seriousness, the website has received a number of upgrades and lots of big companies have built Plugins
Ohh, did Y'all know you can now download your conversation history?
click on your email address at the bottom of the ChatGPT page and then go to settings, then to export and confirm.. done!
they email you a zip file with the chat in a bunch of formats
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
Are you still on waitiist?
what
?
One important improvement we need to see in OpenAI is the output of mathematical statements using true math symbols.
just wait for wolfram plugin
Hope for the best with wolfram.
I badly want to try out plugins 😭
whos' not 🤣 just need to be patient (self comforting mode)
What is the best url to access GPT4 to test it?
do you want ChatGPT GPT-4 or the GPT-4 API?
When is plug-in support coming?
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Read a good book
will gpt4 ever release for free users?
who is this
Nobody know
If I buy GPT-4, do I get the 32,000 token limit?
wdym "buy GPT-4"
do you mean chatgpt plus?
The limit is still 4096, while 32k is exclusive to the API.
I think 32k is not yet available for ppl who have api, so only 8k for api
someone mentioned this previously
@peak surge @mystic nest
Gotcha. Thank you both!
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
can I somehow delete some of old questions in the same chat so that the bot doesn't get the wrong idea?
Does openai stopped accepting new users for their api? It's been 1+ month since I first signed in the waitlist and 2 more times since then, how am i still not in ? :/
Same...
4 weeks is not a long time
hi how does this work i thought they said it was free ..lol
Have you got your access @wind urchin?
Yes, but I originally signed up for the OpenAI beta program back in Jan of 2021, it took 6 months to get accepted
Today, GPT-4 as acting real funky. Feels like it is losing conversation memory
I'm jealous
I'm still waiting for Plugins
But yeah it took long enough for you too
If you can think of a really interesting use case, sign up again with it
im also waiting for gpt4 api e plugins, 2 plus accounts and nothing haha
Hmm ... Like what? Any ideas?
¯_(ツ)_/¯
https://openai.com/careers/ <= fast forward access with this 🤣
so i have to buys to use dall-e?
Did you get it?
I love how sometimes it's like GPT-4 has infinite memory in regards to what i'm talking about with it and then at other times it doesn't remember past 2k words..
Good idea
you might ask in #images-discussions I'm not 100% sure what the situation is right now
I understand this is related to the servers workload, usually mine works better at certain times than others
but to whoever has the api v4, its better to simply build an application that will save your previous info in a local vector database
"Here's the discussion section of an article, remember this when I ask you questions in regards to it in future okay?"
"Yes, understood. Ask your questions"
"In regard to the discussion section I just gave you, what are its weaknesses and strengths?"
"Sorry, you have not provided me with any discussion text"
discussion is roughly 2k words long- so it should definitely remember with the 8k token limit
does any one knows how to reach out to support? im having a different issue. I'm using an amazon vps to bound my lte connections for better stability in a far country area, but for some reason chatgpt blocks it
you can also get fast tracked if have an existing system in which LLM's could be a component https://github.com/openai/evals
thats forgetful haha
i have no clue what this means 😢
Have the bot help you
ChatGPT GPT- 4
i doubt I can get gpt-4 to help me make evals lol
Why not?
To understand them i meant
like... say you have a product you built that you think might benefit from GPT-4 being used with it, ... say you run a time table system for trains....
Oh I see.
Yeah definitely couldn't be me
I actually wanted to see if it could help you play poker but i'm not even a novice when it comes to software development so
as in, in realtime as opposed to asking it constantly
If anyone has it, they could share it with others tho?
bro if you have the goddamn holy grail would you really risk losing it?
they give some starter credit
I got deez funds
I'll pay for it if I could just get in
Well, I would not, having access is like a free ticket to exclusivity, I might use it in conjunction with business partners, but I'm not about to give access to to people I don't have legal agreements with
Sounds like I need to become influencer
It's hard work, hard work gets rewarded, I agree that sometimes... it's unfair, but there is absolutely nothing stopping anyone from just putting the time and effort in and coming up with a product
Agreed. I'll just work hard like ant 🤣
actually working hard alone doesn't work so well, need to be innovative/creative too nowadays 
you can entertain people, or you can save people money, or you can save them time, if you're super lucky, you make something that does all 3, or 2 out of 3, but any one of those will get you rewards
though that's the basic
unable to get to human, the bot didnt help, since my problem is quite specific.
what's the problem?
I'm a loser... No creativity
That mentality will not help, you must put a positive spin on events
Anywhere I can give some feedback regarding the chat app? It's great overall but I have a minor complaint I'd like to bring up.
It will probably be lost in the vast sea of silly, pointless complaints, but I may as well try 🙂
they all get read
Okay, cool, all I meant by that is the app is probably popular so you probably have a lot of questions
sure, but I see people come in every day with unique ideas and suggestions, something may stick out and get noticed
Anyone managing to sign-up for GPT? Cant signup, it says signup is currently unavailable...
Oh, there may have been a solution to my problem
Glad I checked before posting
"The previous model used in this conversation is unavailable. We've switched you to the latest default model." This was frustrating me because I thought there was no way to get the chat back onto 4, but it seems if you just wait a bit and Regenerate Response it switches back to 4?
Is there a keyboard shortcut to get to the bottom of the text in ChatGPT?
You can use ctrl + end like normal
Ctrl + End doesn't seem to work?
I use a 4g/lte connection bounding to have decent internet in rural area, the process involves a vpn usage, even tho I set my own VPN in an amazon virtual machine, I get blocked
Ok, if you 100% need a vpn to do channel bonding.. (I don't think you do, but lets go with it) then you could route the VPN through a static IP on a 1$ a month VPS with a non VPN address
thats exactly what im doing, I have static IP already..
as for the necessity to need a vpn to do the bounding (not a simple network aggregate, its based on openmptcprouter, although its theorericaly possible, i couldnt get it to work without a virtual machine with a different internet connection
right, but the IP you have is one of the VPN's range which is going to run afoul of the VPN blackholing that gets done in realtime by Cloudflare, so, what I would do is get a single vcore VPS and install a proxy service or openvpn on that, that way, your public facing IP is one from a reputable hosting provider and not one likely to get blocked every couple of hours
.
but im using AWS with a dedicated IP, not sure how to improve from there, im not using "shady vpn" or anything like that
🤔
do you know anything that i can try? hehe
lel, try azure.... I mean... hmmm
im about to test aws acelerator, which gets you within a different ip range, theoretically.
i tested vulture, went to aws, no luck on both
i previously used speedify service, no luck as well (that one is understandeable, since sites often block "public" vpn ips
yea, my hunch is it's IP range based, buuuuut, it could be some blocked port.... some NAT thing.. clearly some part of the auth is getting blocked or not returned in a state that CF likes
hey everyone, i was using gpt 4 model on a convo but ran outta msgs so it changed to gpt 3 now idk how to change it back lol is there a way?
The only way at the moment, is to copy the last few questions and answers from the old session into a new GPT-4 chat
It really does need addressing, and I'm sure it will be
awww f i had that convo for so long, i changed to gpt 3 and its so noticeably worse
amen
right, but only the last ~3000 words get used for context, so anything prior to that is ignored anyway
3000 words it spits out or 3000 words in prompts?
gpt-4 is the gpt-4 of gpt-4
the full context (memory) size of GPT-4 is 8000 tokens, 0.75 words per token, so... about 6000, but.... i "think" GPT-4 on ChatGPT has been reduced to 4k to allow more people to try it out, so... that being the case, only your last 3000 words will be used for analysis
there's so many people with "Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
on different internet connections, browsers and VPNs. why is this not being addressed?
right now, it's down to an idle timeout issue, hopefully it will get addressed soon
ahh thanks interesting info hopefully we get the full thing soon
F5 will fix it
it doesn't... that's the problem 😭
have you tried with the Firefox browser?
I tried on chrome and edge
not for some time, it's very expensive to run
edge is chrome with sunglasses and a hat on
they both use chromium, Firefox uses gecko so there is no common failure point
ohhh yeah 😎
alright, let me try firefox then... I've been having this issue all week and it is frustrating already
I've been having this issue all weak and I'm exhausted
openai it’s just thousands of indians & romanians peoples trying to earn some money. be respectful
it's actually 1 monkey 1 typewriter and a time machine
Yup the monkey sits in the time machine and never grows old, just keeps onnn typin
YO NO nononono, not appropriate
Hey, lets keep it respectful, please
jk relax
Nope not cool to jk either, not cool to joke about now let's consider how gpt-4 gpt-4 is instead
funny how people helped to train AI
Ur on the defensive now but we can move past it just tell me ur best gpt-4
and have no idea about it either
I just moved to firefox and it did 2 prompts in a row 🔥
thanks @wind urchin
chatgpt API is able only for py?
you can use JSON and any language you like, the boilerplate API stuff is there for Py and node js
ah, web client?
Why dont... Why don't they tell us these things? Haha
not really a good idea
I just joined, but is it true it can scan images
computer vision, yea
API keys and such, unless you mean web into a 3rd party server.. then sure
Okay thank you
i mean own implementation of API using web client written using target language
sure, as long as you don't fall foul of CORS restrictions in browsers
not API by self ofc, only request/response side
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
sure, as long as you're not trying to access local resources from the browser.. you're golden
Incredible
OpenAI has clients for a variety of langs I think! Also, if ur into that langchain thing which is a whoole lotta work, they got it there too
I haven't started w/ langchain yet, but it just seems real, too much work I'd rather build my own loop
yea, it does, but last time I used the API creator it... borked a little
Which lang? I got it working in js nicely
js for sure, I tried C# and C++
Is there a general REST API?
both went... suboptimaly
Oyyy — uhhhh REST API tho? Sounds like work since OpenAI built wrappers around it and you'd have to dig into that code. But if they got a REST API then any lang should be good
Azure OpenAI Service REST API reference, there's that
to be fair, I just sked GPT-4 for an API to itself and it aced it
Sweeet!
i did feel dirty after tho
loLLL naw it's ok! gpt-4 doesn't mind!
basically, it is possible to use API to generate randomly obfuscated x86 code? 🤔
I mean, an API you build yourself could do that
Why do u need that? (Just curious)
ye, it’s just response machine
well, it’s takes incredible long time to write an IR for do really good binary analysis project. AI may solve my task’s in seconds, obfuscation just idea from the list.
I built an API in node that reaches out to OpenAI. This way you can modify any responses before it reaches the user
it’s a diff task. right question can be: “what chances of false-positive results & how to avoid them”.
Here's a secret idea that I'm sure other ppl came up with. Write another bot to double-check the initial bot's conclusions
hm, i may think bout that, cus rn i have no idea of how to get rid of the ai hallucinations
Which model r u using? gpt-4 is better at avoiding hallucinations. I believe after some testing I've found it's more likely to hallucinate after a lot of text is in the conversation window
If ur passing up code which could be a huge amount of tokens, it could be interesting to try to figure out how to split it up and tell the bot about that
both. they’ll give wrong output about x86_64 asm & absolutely cannot work with special libraries API’es (e.g gives non-exist method name). there much more but i’m unsure of what i can tell.
Have they stated what the token limit is for the CharGPT implementation of GPT-4
About how many characters are u sending up?
Yes. From what I can see, ChatGPT 3 has a 4096-token limit, GPT4 (8K) has an 8000-token limit and GPT4 (32K) has a 32000-token limit.
I've been able to manage a pretty lucid conversation window of 20k characters on gpt-4 (8k version)
Not much hallucination from what I can tell
Sending up code tho, that's gonna eat up so many tokens. We'll need to build a framework to help gpt parse and respond in chunks
The langchain community I'm sure has strategies around that. I'd refer to them for strategy but I'm not sold on using it at all, since I know how to build APIs and don't see it as necessary at all. Not at all!
embeddings
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
Is it cost-wise reasonable to programmatically create embeddings for a large upload? That could help with Q/A mode on the code, but not necessarily transcribing it into obfuscated code
I think the task of rewriting a codebase or even a file of code to obfuscate it using an LLM may be above the abilities of the LLM, just because of context space
Or perhaps the quality of output may be related to the percentage of upload you give against it's max token limit
Managing the context window of a convo is a big thing, I've managed to do that on a project, like I mentioned, 20k characters seems to be a sweetspot for the base gpt-4 model
Can someone plz help me get a job??
I just wanna work at OpenAI and tell them they're doing everything wrong
Probably not the best attitude to have. I doubt they're doing much wrong. They just have information which you don't have that leads them to make different decisions
Yaa, I'm just playin, would love to work there tho
@fair iron will u hire me?
Omg
for?
Work, Openai work
considering I don't work for OpenAI or represent them.... no 😅
Hire me to manage ur local system files?
I got that handled
😂
Ok lemme know if that changes thanks u!
Does anyone know if there is an autogpt discord & if it can be installed on windows
it doesnt seem to be ip related, I changed range 3 times, either whole aws is blocked, or something related to the vpn tech that the site doesnt like.
Utilizing your server profile to bypass the filter may lead to infractions. AutoMod will notify you if you mention any blocked words or URLs.
There is! If u google 'autogpt discord' there's github issue with a link in there! I do believe it's def written to work on windows too.
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Does anyone know when the plugin feature will be available
Just have to be patient
?
Hi there. I was under the impression that GPT 4 allows for 25000 words of input. However, why is it being limited to 2000 or so words?
I've found that specifying gpt-4 as the model allows me a conversation context window of 20k characters pretty well
That's not happening for me
What symptoms r u seeing? Is it hallucination or rejection of the request? What are you prompting with? How many messages and characters?
I am getting this error: "The message you submitted was too long, please reload the conversation and submit something shorter."
That's a verbose msg I haven't received from the api yet. And the shorter prompts do the same thing?
When I send a request to the API that has too many chars I normally get a generic 400 err
Anything more than 2800 words does that
Hmmm, that could be around the 20k char count I've found a sweetspot in
Are you using the general window or the playground? 🤔
Using the node API, I think the diff APIs should all accept an array of messages, trying to figure out where it's going wrong for u
Does it fail on prompts with far less chars?
GPT3.5 is better at programming than 4, is that to be expected?
You may need to look into a framework for breaking up large prompts or uploads
No way 💀 It is NOT
How did u arrive at this conclusion?
I swear, gpt 3.5 is stupid compared to gpt 4, I asked it about an error and it started describing code 💀
Nah. It does not do that. Also, is it 20k characters with or without spaces?
For my API, it includes all chars, just a JS string count
What made you think that?
In that case, it's probly the upload is too much for the model to meaningfully parse thru. U might need to break it up into chunks and teach the model how to go thru it piece by piece. A serious effort
Personal experience until now, GPT3.5 was able to write an API call to Walmarts Affiliate API in Swift, while keeping track of multiple Swift files for an iOS app. GPT4 became stuck in a loop and despite opening up new conversations had trouble recalling recent submitted swift files.
There's a "Sweet spot" where at X number of messages, GPT3.5 doesn't forget what you've mentioned and will perform greatly. However with GPT4, I've had a lot more recall issues and I've had to re-iterate while it makes basic mistakes. I can literally paste the errors in Xcode console into GPT3.5 and get fixes. For GPT4, it was looping for a bit.
Anyone know why the latex rending command is not working properly?
Huh. Interesting.
Quite literally the opposite experience with me.
Agree that is interesting. I was under the impression from my testing that gpt-4 was better at understanding and reasoning, and coding. AND has a much larger context window
I do notice GPT4 seems to forget previous things, but gpt 3.5 just seems to know nothing about dart
I mean I'm honestly going to stick to GPT3.5 for iOS dev at work. Understandable, it was OK at first, but GPT4's "working memory" seems a lot shorter than 3.5 to me.
But that can't be, cuz it's not true. gpt-4's context window is far larger.
/
Doesn't seem like it to me either
Remember it's about managing a context window
Are there any suggestions for how to teach the model? Because for example, even though I asked it to wait for my next prompt as a continuation of the previous prompt, it started to generate random stuff.
Fair. This 'ChatGPT' or the playground/API?
I mean, I spent the last few hours working, all I wanted was a simple URLsession build and GPT4 kept looping between 2/3 answers (suggesting Alamofire) despite further context being introduced.
Just ChatGPT no playground/API.
Cat
ChatGPT's web implementation of GPT-4 is good, but not as good as the playground/API implementation imo. Loses context faster from my testing.
Try out the API if you can.
Yes. I'd suggest teaching the model how you want things to go with the prompting, what to expect from your further prompting (including 'prompting bots' which provide info), and teaching it to respond with something like "I understand." or "Understood." Then u take that and program ur own loop. You can use the prompting strategies employed by autogpt to make a loop. Managed loops for LLMs are the future!
Will do, thanks. Are they doing trials/waitlists for the API still or can we just zoom in if we're premium subs?
Forgot to reply, rip.
There we go.
This is a great suggestion. Thanks! 🙂
I've been able to section-out pieces of info I want to tell my model by messages sent to the API. I have a 'current-info-bot', and several other bots that provide info to the model on each request
API is on a waitlist, link is in #gpt-4-faq
I've used those messages as 'system' level, and:
Give the name of the bot in the name field.
First message tho is something like "This is 'current-info-bot' providing you info."
I've had success telling my base personality model to be on the lookout for messages from other bots using this technique.
Us premium subs don't get API access instantaneously as a result of our subscriptions but for stuff like plugins we supposedly get priority access.
A lot of the art in prompt engineering I feel (And it's a brand-new field so who knows!) is around anticipating issues w/ requesting things from the LLM and responding to issues. Like how I built an auto-continue bot to run if I see a msg ends on an english character that doesn't typically end a sentence (need to update that to other langs ofc), and an oops-bot that prompts the LLM to regen if an issue happens on the server or something
Why isn't the latex rendering working in GPT?
@sweet thistle Will you hire me?
I'm not hiring for anything personally and I'm just a community volunteer 😅
If you're interested in working at OpenAI look into https://openai.com/careers/
Ughhh yes I applied they didn't respond, I'm out here trying to build a chatbot that will talk to ppl for me so I don't have to
That sounds pretty convenient, maybe post about it in #1037561385070112779 if you haven't already 😄
Thanks I posted in there, it got buried, and I can't find it anymore. Can u help me figure out how to get on bluesky?
i am
💯
they should make a thing (forum, channel, bot, etc) that lets you give it information for the new gpt-5 model (or really any other new model), so you can train it on data it hasnt been trained on before to make it smarter (yes, it can be reviewed) idk just something i thought of that could be kinda nice
can you write paper?
?
the chatgpt is gpting
.
guys is gpt 4 still limited in terms of the amount of questions you can ask
When will I be able to insert images into chat gbt
Yes 25 every 3 hours
This feature is not yet available to the public.
Why not
i dont know, im just quoting #gpt-4-faq
Lol nice
Why is gpt-4 always giving less code than gpt-3.5? It seems to be smarter but a lot harder to make it code larger projects
hi. can we feed image as a input to chat GPT-4?
read right above you
💯 This is my experience as well. It routinely forgets requirements, injects previously fixed errors, gets stuck in loops with errors that it thinks it's fixing and constantly injects new code rather than building what has been previously used. It's almost impossible to use sometimes and I find myself just deleting entire chats and starting over.
I'd much rather have 3.5 than 4 at this point.
I've even had it switch languages on me from C# to Python for no reason. It just decided to start generating Python and couldn't remember that C# was being used after I asked it what it was doing.
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
Man
People losing braincells
They try to talk with #gpt-models which isn't even a bot?!


Anyone know what the folder like icon is for? It's beside the 👍🏾👎🏾
/ imagine
lmao
Perhaps you need to break the project into more "bite-size" chunks? I find it codes, swimmingly with descriptive and precise prompting. Far and away what chatGPT was producing.
#gpt-4-faq how to get a pi value
Hello, I paid for GPT Plus, and I get this message :You've reached the GPT-4 cap, which gives all ChatGPT Plus users a chance to try the model.
it's just like what is said...25 messages cap per 3 hours 🤣
Thank you @mystic nest , I didn't pay attention, so what is the solution ?
nothing, just have to wait
it's a sliding window limitation, so when your first message reached 3 hours, you will be able to query once more
OK, thanks 🙂
If you ask, you shall receive a subset
Buy another
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
when will be the addons available? i wna try it
Wut like plugins?
yeh i mean plugins
why everyone is pinging the bot maybe it works?
After filling out the form begging for access, it's up to some human at OpenAI to grant it to you. But let me tell you. The plugins are GARBO
They suck! It sucks!
It's not fun
If you ask me, I'd implement it way better and already have, alas, funding, so my project is paused
you sure it sucks? cause the potential of it are huge
I built a project that would inline image gen and that's just the start of feedback from LLM to human during a convo
Ya, it's boring, unless u already use opentable or one of the few plugins they support
Here's the beloved plugin store
Oh I can't upload pics here.
🤣
Deepspeed
Available plugins:
- Milo Family AI
- Wolfram
- Instacart
- Klarna Shopping
- Shop
- Speak
- FiscalNote
- Kayak
- Zapier
- OpenTable
- Expedia
isnt that what they tryna do?
It's like you clicked on an awesome article but you'll never guess what #12 is all about
It's like you clicked on an ad of ads
they dont have a description for it?
lmao
They do it's all ads
hello party people 

I'd rather click on clickbait
I love clickbait
I haven't used a single plugin besides wolfram to determine the size of my ego
And they said it's something like 0.000000012030303034405
I'd rather walk around the mall and go into every store and tell them I'm interested in their sales
I'd rather signup for coupon code collectors and receive their spam
I'd rather signup for all the cool EZINEs just to see what the coolest celeb gossip is
I'd rather dig up a big pile of dirt and fill the hole back in
You seem like someone vulnerable enough to click on clickbait like me. Would you click on my clickbait? It's better than plugins I promise
How
It's point and click technology
First, I bait you. Then, you click. It's like a microcosm of intelligent life, unintelligently clicking
This creates a feedback loop of clicks. Eventually with LLM technology the clicks start clicking themselves and form a feedback loop of clicks.
Then, once the clicks have reached the point of no return, they start generating new clicks to click
be honest how many of u all have access to browsing or plugins here and when did u applied and when did u get access for it
I had to be a dev w/ years of experience, just spouting about some app I'm working on. They ask for ur socials and pay attention I guess.
but do u have access yet
Oh ya! I got gpt-4 api access and had to pause my project due to costs
Would you like to learn about the garbo plugins?
I imagine the api costs add up rather quickly 😅
Here are the garbo plugins. #gpt-models message
I DID have a bunch of users! My only issue is I was giving gpt-4 away for free!
RIP 
18$ still a thing in playground ?
free one
is it not?
If ur building something and testing it out, honestly, the costs don't add up that much. It's when u start to get users
i found this site claiming to use gpt-4
but they were actually using gpt 3.5
i'm pissed off >:/
clickbait
Cancelled.
I'm not sure to be perfectly honest. I know new users don't get Dall-E 2 credits for free, but I don't know about api credits
you're the hero we need
🫶
I think there's a form u can fill out if ur looking to use models for research purposes
Awww!
and how is bing chat using dalle upcoming model for keeping it free it have to eventually be paid i guess
I filled them both out day 1 🙂
Ya I think at moment GPT 4 priced out of reach of 95% of use cases…but doesn’t matter as they don’t have capacity to serve that 5% where it makes sense
I assume it's part of Microsoft and OpenAI's partnership. I don't know anything for sure though
Here's 1 link I think there might be a couple https://openai.com/form/researcher-access-program
ok
hahahahahahahahah I made a discord bot doing the same thing
I enjoy the ones I built for it
my friend helps with the gpt-3.5 costs
I filled this out already, but thanks!
Love it — it's gonna be us that make the coolest stuff, moneys is the issues
I feel that, I'm using GPT-4 on my support bot on my community
i wonder if sam altman is in this server
And every time I see it used I grimace a little bit
Oh well congrats that's sick! I'm sure ur plugins rock!
The stock ones are pretty lame
not AFAIK =P
rip
But the actual API itself is really nice
so, fun fact
He doesn't even know about this server
Really easy to build
I was sleeping when my friend who handles the GPT 3.5 costs for me got GPT 4
An LLM built this server
and because we share the same organization... I got it too
I bet he's heard of it
the price is too high
ye gpt 4 costs need to go down :(
i mean like the owner of stability ai is in the server and i have chatted him allot too...so i thought sam owns this so he might be in
Due to costs of his tokens, he can't participate
API is too expencive
It's because of inference
If you INFER from the LLM that it has inference abilities, then it will infer a charge upon you
i have ran up over 30$ already this month just testing things
lol
its pretty easy to do on GPT4
do you also have chatGPT Plus?
What is am working on is multiple levels…and using GPT 3.5 to handle most tasks, but be able to relay to GPT 4
yep, I use that for any long form interactions I need to do
me use daily, 0.3 usd so far on 3.5
Many tasks GPT 3.5 is already super human on
you are 100x of mine
hi can someone help me troubleshoot my GPT4 account
Hey just a heads-up to anyone looking to build a website. As soon as you implement GDPR you can kiss your goog analytics goodbye.
How can I get gpt4api?
im at 30 cents today and i think quite a bit of it was whisper testing
email support@openai.com about tech support please 🙂
playing with the idea of a smart assistant integrating with homeassistant
you can sign up for the waitlist in #announcements
Have you tried whispering quieter?
do you have apikey?
yeah i need to learn more about Porcupine first for the wake word
I DO! I'll never give it to u tho sry
yes that is how the api works
everyone has an api key
that would break the TOS 😉
ok
Id be willing to pay for api4 access
And this is why you haven't been granted access
you'll have to wait on the waitlist like the rest of us
i feel very lucky to be able to have api access
Ya and my spirit I can't trust other peeps w/ my stuff
Sames!
because I have access I am trying to actually create with it
the reddit stuff I put on hold until I find out what reddit is doing with api access 😆
btw, users who where selected from the wait list still pay for their usage, they are just allowed to pay to use GPT-4
incase it isn't clear, asking for api keys or asking to pay for accounts with api/plugin access in front of a moderator isn't a great idea...
will u hire me tho?
imagine giving api to others lol
I still don't work for OpenAI
. Nothing's changed on that front
have someone made auto gpt plugin for chatgpt ?
yeah "here take my wallet" basically
LOL!
Just checking!
(none of the server staff do)
I run a few large communities and this place makes my head hurt, i feel bad for the mod team here, but your community is very large compared to mine
I actually once gave my friend on accident, and he returned it to me and explained it to me.
2nd largest server on the platform!
yea.. even if you want to share your acess, the system has the "organizations" feature for that so no one has to ever use an apikey for an account that isn't your personal one
i was like #10 in tech
so the lead role peoples here word for open ai ?
which his pretty cool for a community server thats not part of any company
they do not. OpenAI employees will have yellow usernames. #announcements has messages from OpenAI employees
ok
oh that's cool, however how would it know that me is me, base on ip or hwid?
I am surprised how few projects I see that are actually open source in #1037561385070112779
maybe i should start putting my half finished things in that "work" so people can use them as bases
okay, this chat isn't a job board...
tagging random staff for a job.. is not how to get a job 😆
loLL
in fact feels like the opposite tbh
Ok I'll try to be good — ya, I've given up on that as a possibility
on the OAI accocunt dashboard there is the "organization" feature, you can invite users to join your organization and with that you can grant them access to use your account resources, the invited users can generate an APIKey of their own and at the same time, specify the organization ID on the requests
Hey but that's cool ur managing a bunch! Send me an invite!
when I set up a paid account it took 5$ without my permission I wanna refund
its why I think openAI's wonderful API can help lol
we built an embedding powered GPT4 support bot that gives a single response to a question to try to help
yeah, I pasted my source code to my friend included with my key that time, gladly he noticed and told me.
.
I simply don't have time to help any other server. I'd love to, but I couldn't manage the time commitment between multiple haha
personal tip, keep things like api keys outside of the code, use a json or something
3M members 😎
Use .env!
email support@openai.com about that please
I do that now
yeah env is smart
they're not replying
when did you email
Did u ask nicely?
I sent a message from their website days ago
they prob busy with people going "why cant i use plugins" ¯_(ツ)_/¯
we can't help you on this server as the server staff team doesn't work for or represent OpenAI. We can only point you to that email address 🫤
use chatgpt to ask nicely
Probly smort
hahaha
you might be mistaken the 5$ free credit you get, does it say on the dashboard it is a 5$ grant?
Ok let's not talk about 5 dollars
I've been building out a reddit bot for our event days to look for specific post types and flag them, its way smarter then any regex can be
Maybe if I ask chatgpt nicely it will give me 4api access
it didn't say nothing when I linked the card it took 5$
gpt-4 for an app costs far more then 5 dollars my fren
do new accounts get $5 free credit for the api now?
I live in jordan man!
it was reduced a while ago from 18$
Oh alright, I'm sry I'm probly just being ridic
it matters 😄 and it's without my permission I mean
but it's currently $5 for new users?
@waxen badge you should read the page on billing
if you see that 5$ doesn't worth you can send it back to me so I can stop chasing it
I'm saying the $5 is pretty low for considering building an AI app
if you are sure it took 5$ from your credit card, then, yea.. you will only be able to get that issue handled with the actual OAI support via email, be aptient, you'll get your answer =)
there is a 5$ grant to new users
Yes but they didn't tell me I mean
Ah ya, oy
It directly took the 5$ when I did set up a paid account, Idk what's the difference
that's likely an auth hold
thank you! 🫶
Ahh, ya wut webhead said
not replying
i remember vaguely getting that a few months ago
this is not normal, the system does not simply make a 5$ charge on your card, I have my card there for months and it never did a wrong charge
All G, I think they do that to ensure the card works to help prevent spam or bots
it doesn't make sense they didn't return it
"Pay as you go
A temporary authorization hold will be placed on your card for $5. At the end of each calendar month, you’ll be charged for all usage that happened during the month.
and it clarifies what it means by a “temporary authorisation”:
An authorization hold puts a temporary lock on a certain amount of your available bank or credit balance. This appears as a pending charge on your account and will reverse within 7 days.
You get similar things with car hire, hotels, or pay at the pump for petrol/gas. It’s a way of verifying that the card is good for making transactions, but it should be reversed within the 7 days."
Don't think they will
Crying about $5 smh
that's 4 hours of working
this makes sense, it is common, but is usually less money... well, idk depends on the payment processor
anyways gonna go play with whisper more
Psst. psssst...
and figure out the homeassistant api lol
Plz do my dishes?
@icy swiftthose comments of yours where very unecessary
Im running autogpt but its lame without gpt4
its pretty lame with gpt4 tbh

I do think it's worth noting the dollar amt considering places outside the US, so I'm sry.
Gpt4 keep getting error now, is that only to me?
@tardy heart dang was hoping it would run better
the 5$ worth a meal. it's not about the amount .. I mean why?
I think I responded in a way just cuz my bill is waaaaaaaaaaay higher than that
Like way way way way higher
my bill makes me a little sad this month lol
we've addressed the issue as much as we can here. let's move to a new topic 🙂
idk man imagine I wanted to use AutoGPT
im gonna be at like 70$ in openAI spending this month
I'm gonna cry tears on my bill
and i don't make any money off my communities haha
Are you a bot or a real person?
they are a person, clearly
Hes abot person
pretty sure I'm real. It depends on who you ask though 
If you need to pay close attention to costs like so, I think you should build your own LLM loop system. It's possible for autogpt to get caught in a loop.
tbf they couldn't have a bot interacting in here under a userID without a bot flag without breaking TOS, as well as responding to each of us due to GDPR data collection
I live on 30 bucks a month what cost are you talking about
the api is not free
You could've been ChatGPT-4-modified disguised as a real person
I'm saying if cost is an issue at very low lvls, which makes sense and is ok, then I'd encourage you to work to build around and build your own framework
One that you know won't be unnecessarily spending
"as a Large Language Lizard.. I'm pretty sure I'm real."
Amazing
if you ever want to have fun with people in a game
use the playground and set a system instruction around the game, RP with it
who's to say I'm not 
and use it in chat, gives people a laugh
was doing that in star citizen the other day
Every dinosaur is a meal if ur a gOAT
the amount of people that does not believe that the staff is composed of real people is too high
well there's a few reasons i know they aren't AI
openAI would not willingly break GDPR I'd presume
and discord would yeet userIDs that are running under bots that aren't clearly defined due to the policy of message intent
that's why I gaslight them into believing I'm a bot. It's fun for everyone!
YEET
its true
Doesnt have to use openai
If you disclose it, they will... enclose.. it..
Could be a local llm bot
still running under a local user account and not a bot account
Yeah but is it human?
For prime example. I am a human trained by AI.
Have you considered this?
I'm a bart eat my hsorts!
Im a bort
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
How do raise capital for my business?
Maaaan I wish I knew. And in fact, it's something that might not even be worth it. Soon as investors get involved they get ideas.. And they have control
I'd say introduce freemium features to ur app
That's wut I'm gonna do for ChappyAI
I've never interacted with a bot that told me something incredible like this
That I could possibly open up a wormhole into a new kind of existence. An otherworldly existence that only persists on planes of reality we haven't yet observed...
It's like reading your horrorscope only to reveal that you've been writing it yourself the whole time.
It's like opening up your https://chat.openai.com/chat to reveal that gpt-4 is just your 4th-grade teacher trying to redeem you for all your wrongs in life.
It's like staring into a nebula but the nebula only tells you poems about how anthills are actually a great society we could all learn from

It's like jumping into a lake filled with ice cubes but not the regular kind. It's the kind that sticks to your botdy and tells you u can never melt.
U get it
write a post
Mayb in the morn
None
oh!!
hi- i'm trying to get my chatGPT plugin to work. It sometimes works but other times crashes the chrome tab with an "Aw Snap". anyone see this problem?
Man GPT 4 lightning fast for me tonight, I keep having to check am not on 3.5
seen it with a plugin you are building?
mhm
hmm i guess i feel a little bit better. 🙂 could it be the size of the API response?
Where can i get prompt for chat gpt?
How can I use Chatgpt 4
You need to subscribe to ChatGPT plus or you can also use via API
Chat GPT plus a steal vs API…good to have both
How can I use the API
Ask Chat GPT
In short, you can use API to create your own app but API for gpt4 isn't cheap. Plus is good if you just want to use it for yourself
Is anyone having issues with ChatGPT not working on windows?
2 days now I haven't been able to use it, all I get is an "Something went wrong"message
Did you tried refreshing, clearing cookies and cache or trying different browser?
yep, still same issue
Is chatgpt 4 working?
Does anyone know for sure what the chat GPT plus context window is for 4? I think the response is 2048..
yes
anyone else?
How to use it?
@astral crag how r u?
What are the things that we need to learn quickly to integrate AI into our lives and business so that we can transform to the more intensified Digital world smoothly?
What do you guys think of Dan?
Will message restrictions be lifted in ChatGPT 4?
/subscribe
why none of my commands work
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Probably no time soon…they are extremely slow at giving out API access to 4 which is much more expensive
Nor have they rolled out 32K yet
@wraith ginkgo hi can you make a service
bruh i want access to gpt4 so bad
i guess you could use dall e 2 but it probably wont give you a very good result
Why chatgpt can’t be opened all the time

🙃
yes
i keep getting this message.. Something went wrong. If this issue persists please contact us through our help center at help.openai.com. what can i do?
Just uninstall it and then install ur head into the dirt like I have
How can a 10-year-old Chinese boy study in Singapore?
Internet works! Also ChatGPT is a great tutor!
bruh you joined today
wth
This isn't just any ordinary bruh bruh
interesting
Told ya
i didn't even see antyhing
Cuz u weren't looking close enough
You have to install ur head in the dirt to see the worms like I said
Does anybody even pay attention around here?
Where are you going
To the moon
#gpt-models hello
High 
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
how do i make it human written, cause im using zerogpt
Is anyone else having a problem signing up for an account? When I get to the Phone number verification step it just brings me back to the date of birth section
kek
U bought ChatGPT tokens?
I was born on a pile of GPT tokens
🍾🥂
Yes

Born in the tokens, molded by the tokens
hello...
can i use gpt4 on multiple platforms?
I am token
yes
like 2 or more ppl can buy same account and use it on different devices. ri8!?
and can tehy login simultaneously?
they*
Hi everyone, how to access to Plugin? i have subscribed to GPTPlus since March and still waiting for the access for Plugin.
@wraith ginkgo can you make me a logo?
china
wolfram alpha + gpt-4 would be insane. computationally intelligent enough to figure out complex maths and compute things like the current exact distance from the earth to the moon based on old data, mixed with gpt’s image processing. just a wack thought but itd be cool, no?
china
厚礼蟹
what is wolfram software?
You can't
Well, you already know what to write, right? Just write it differently
yea however is zerogpt a reliable website
for AI writtens
whats the website for it
This server do not speak Chinese please
It's a padi service
whats that
how can i use mj
there're no reliable way to detect content generated by chatgpt, but of cos not going to say how to do it
Hello I have recently subscribed to chatgpt plus
but I am having error in generating response anyone can help?
Google serach,how use midj for work
He prompted Subscription required
GPT4 not working right now?
is your network stable?
I'm just getting
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
It was fine a minute ago
yes it is it stable
so its not only me ?
🥲
what's the point of subscribing to plus if you are not getting any help or piority concerns?
well, this is a community discord, not support
access to gpt4 and faster response time
but yeah, ppl been complaining support ain't reacting fast
like takes 1~1.5 week? not sure if there is any case that took even longer
I have emailed them
Typically replies withing a week
just asking the community if they have faced it before and how did they fixed it
Instead of waiting a week for an answer
I can use GPT4, but he reports an error every five minutes and needs to log in again
it's just that for ppl who stay long enough in the discord, we see too many same questions that can only be answered/fixed by openai 🤣
vpn, account, system error
if we can help, but too bad not for these
The email I wrote has never been replied to officially……
when did you wrote?
A month ago
now I am getting dissapointed
it would be nice if openai can expand their FAQ in #gpt-4-faq to address current issues, but think it wasn't touch for a long time
I used to think it was because of my network problem, but now I realize that it's not just me who has encountered this problem
VPN problem
Mine works perfectly
If you are in China you better find a proper VPN to use
I have already switched to two VPNs, okay, it seems like I need to switch to another one🥲
Let me recommend you one
👌
I need one,too



