#community-help
1 messages Ā· Page 154 of 1
GPT4 is currently down for some people
Including me
my other gpt4 open chats are working
ChatGPT is up -> https://status.openai.com
Welcome to OpenAI's home for real-time and historical data on system performance.
Yes, but refreshing the page should fix it
resolved
But the PDF file is not empty
are u using a vpn
anyone else unable to upload files?
what happens when you try?
it takes a long time and then says "unable to upload"
has anyone actually got the GPTs to accurately and consistently search knowledge, or is that a pipe dream lol
I have a Mining and Prospecting GPT which uses it's knowledge perfectly. I haven't had a problem with it. When you have trouble uploading files what filetypes are you uploading and how big are they?
it actually just uploaded finally, then it said "message in conversation not found"
they are text files that are not even 2000 KB, but there is a lot of text so maybe thats why it isnt working. what kind of files are you using, and how big are those
My GPT has an 80 page PDF, which is textually dense
ah 80 pages, i guess 600 pgs is no good lol
maybe next year lol
I have worked with a 300 page PDF without a problem
wow i wish mine worked when i uploaded PDFs, it always searches for a long time and then fails
can you try another format, for example EPUB?
thats a good idea i will try that now after converting them
Is there anyone familiar with adding an API to a custom GPT?
My goal is to have a GPT connect to my google docs. For example, I want to have a google docs that has my goals saved in it. I'd like my GPT to get that document when needed to see if any changes or progress has been made to my goals. Any idea on how to do this? I'd prefer to not use Zapier.
The original functionality is broken. But if you know json, it still works fine.
Ah, that makes sense as to why the screen looks different and chatgpt now isn't acknowledging that i'm able to connect an API to my gpt like it did yesterday. I am familiar with JSON.
In the docs i've read from google, it looks like you typically need to setup a developer console on your google account. Do you know by chance if that really is necessary? Or is there another way to just have it connected to specific documents I want it connected to?
apparently epub can only be used by code interpreter which makes it even worse at reading the file
my GPTs have all used EPUB
thats weird, mine says this: The following files are only available for Code Interpreter
I don't know then I am sorry. Is a .docx a viable alternative?
can it read docx well?
I don't know, I have only used PDF, RTF and EPUB
which one works best for you
well until you put doubt in my mind I was happy with EPUB
I'll create a new GPT to test EPUB with again, give me a moment
I cannot directly use the file you've uploaded, I'll need to convert it into a readable format. Once that's done, I can provide a summary for you. I'll start by converting the EPUB file. Please hold on for a moment.
Fast moving space, EPUB is no longer a viable option
ah that sucks lol
I'm considering building a website to share my DALLE prompts and tips and I was wondering if it's permissible to use 'DALLE' in the domain name. Could you please advise on any guidelines or restrictions regarding this?
short answer is no, you cannot do it. Longer answer is here -> https://openai.com/brand
can anyone help me find angular documentation in pdf
any help pls @tribal talon ?
so. i have voice gpt on my ipad but not on my iphone. same apple account and openai account and everything. it's been this way for at least a month now. is there anything i can do to unify access across devices?
"The previous model used in this conversation is unavailable. We've switched you to the latest default model."
Yooo, what? Can someone tell me if there is a way to switch my chat from simple GPT4 to GPT4 with plugins?
i don't think so. it's a real downside. especially w/ query limitations it would be nice to swap around.
disable ad blocker/privacy dns services/extensions
hmmm im not aware of having any of those that might affect chatgpt ios app but ill check
dns services affect everything on the phone
also where did the option to log in with microsoft account go
bah i guess im an tablet-based user for the time being
Is there a way to have my custom ChatGPT assistant be able to search Reddit posts?
can chatGPT 4.0 plus actually contextualize long pdfs or does it just form a main udnerstanding and act like it knows everything by providing made up/estiamted quotes/analysis??!
cuz I heard the data anayalsis mode is actually made to understnad PDFS, and just skims it and forms an understanding
somone know the problem?
What is this?
It won't let me write anything
Is there a maximum number of files I can upload to Knowledge (GPT builder)? It won't let me add more than 10 files because it doesn't load the drafts
is the TTS api live?
there is no TTS product by OpenAI
there is speach to text, if that is what you mean
deactivate the plugin
This is not true
https://openai.com/blog/new-models-and-developer-products-announced-at-devday
Developers can now generate human-quality speech from text via the text-to-speech API. Our new TTS model offers six preset voices to choose from and two model variants, [truncated]```
I hate to dampen your exitement, but ChatGPT is at the math level of an eighth grader. Unless you integrate an API with Wolfram Alpha or similar, there is no use for such a bot.
Code interpreter is capable of running calculations and is allowed for usage in GPTs
darn, second time I do that today, lol, indeed, yes it is already on the API š
no no, i meant tts, and i already managed figure out how to use it through api. works pretty good!
I checked it too soon on devday.. and just today I figured that I could have already been using vision lol
it was my mistake, it is already on the API, just checked, lol
https://platform.openai.com/docs/guides/text-to-speech
Yea ChatGPT can do the same, but on the logical side of things, ChatGPT is tending to hallucinate
AIPRM is an unofficial extension
Ate this propositional logic problem like it was nothing
(yes it's correct)
from itertools import product
# Define a function to check the conditions
def check_conditions(A, B, C, D, E):
return (not A or B) and \
(D or E) and \
((B or C) and not (B and C)) and \
(D == C) and \
(not E or (A and D))
# Iterate over all possible combinations
def find_who_is_at_home():
for A, B, C, D, E in product([0, 1], repeat=5):
if check_conditions(A, B, C, D, E):
return {"A": A, "B": B, "C": C, "D": D, "E": E}
return None
# Find the combination that satisfies all conditions
who_is_at_home = find_who_is_at_home()
who_is_at_home
I can't judge, never had math on an academic level. But there were a lot of smart people here lately who demonstrated, that it sucks at math pretty often. You can also just google this, there are a lot of people that say the same thing. It just can't to it, even with code interpreter. Thats why there is a Math AI like Wolfram Alpha. Else this wouldn't even be a thing anymore.
With that said, this doesn't even look like a hard question to me.
Iām sharing our new GPT with others but they donāt seem to be able to access it. Iāve published publicly. Any ideas why others canāt access it?
guys are you getting HTTP ERROR 400 when trying to use dall-e 3?
Help! Sam Altman just announced that a new version of GPT-4 Turbo is now live. However, I can't find it in my account.
What's gpt pro?
How do I access GPT Pro? I have a Plus account.
Yes, I have a plus account, but I'm not seeing GPT-4 Turbo. Maybe I'm already using GPT-4 Turbo by default?
Yes, I'm currently able to do that.
Guys, I am working in the playground with an assistant, and it is not showing any generated images. They show as a broken image?
Any ideas for me?
i am using gpt-4-1106-preview
It's only available for plus users.
How do you get access to the playground? Can you share a link?
probably never, they are already working on 5. maybe 4 will be a free tier in about a year
i put a credit card in and used platform.openai.com
It's not that much for it's value
its so worth it to be a part of the ai revolution
mow one lawn a month
But seriously, anyone know why my assistant is not showing the generated images? š¦
Thanks.
It's indeed GPT-4-Turbo, like how 3.5 in the selector is GPT-3.5-Turbo
please god I hope this gets fixed
I have 1 single 7kb file
and it just wont accept it.
if i want to combine GPTs with plugins is there a current way to do that or will there be an update? im willing to learn. thx
Anybody else getting this?
Yes. Just got that.
I guess we just gotta wait then
i'm getting an error message of "Moderation error"
what the hell is going on again?
The same issue here
I got an error of "Mandatory error" first, then tried to login again and got the same error message as you
Error in Moderation
A lot of people are getting this right now.
Like someone else said, this is a new one. Havenāt seen this one before.
Everything I do says "Error in Moderation" ... how do I fix this?
this is what the errors are about
It's even for queries like "What is a dog?"
They need to start pro-rating my subscription fees ...
I want my $0.23 back Elon
(waits for someone to tell me that Elon doesn't own it lol)
they keep working hard he might buy it š
what doest this mean
.
same issue
thats why
both gpt 3.5 and gpt 4 are showing me same message error in moderation
Servers Are down or getting to many API Requests
Iām waiting for https://status.openai.com/ to update with some sort of issue. š§
Welcome to OpenAI's home for real-time and historical data on system performance.
šāšØ š š
i saw a tweet yesteday of sam altman said that he uploaded a new faster gpt 4
but i didnt notice any difference in speed for gpt 4
still slow in generation
hahaha
lmao
I just got this too. What is this? what is an error in moderation? did I go over a prompt limit or something? where do I ask about this?
Boom. Just got the notification.
Issues with Labs, API, and ChatGPT.
Perhaps another DDOS attack? Ugh, this is why we canāt have nice things.
If anyone has access to Sam Altman, please ask him not to update ChatGPT anymore, because every time it is updated, the performance and accuracy seems to worsen.
the error seems to have been restored
Before I do a bug report, I just wanna check if anyone is having issues with actions. Mine just says the current environment doesn't support internet access.
Then again, I am working with the worst and most complicated api in api history. So I could have fault
Its my fault, nevermind. Turns out I made an error when writing the 57th parameter
That said, it doesn't support massive schemas
Anyone elseās gpt unable to do maths, just hits you with the āerror analysingā
Me
I have that problem
Hey everyone, I am experiencing difficulties in uploading various file types, including CSV, XLS, and even DOC files. To troubleshoot, I've attempted to upload a simple Excel sheet containing only two data entries in two cells. However, I've been facing this issue for about 4 hours now. Is anyone experiencing the same with ChatGPT?
I just got a "bad gateway error".
yeah i have the same, cant analyze any excel file
Error Details:
On the first attempt to read the file, the system returned an AceInternalException error without additional details about the cause of the error.
A second attempt to read the file also resulted in the same AceInternalException error.
Specific text of the error or detailed description of the fault's cause is not provided in the system's response, making it challenging to diagnose and resolve the issue.
yes same error
The ChatGPT password is forgotten. Why am I not receiving OpenAI's emails for Google account verification? Has anyone encountered a similar issue? How can it be resolved?
Is anyone finding the new version of ChatGPT to be sluggish?
For the past three days, I've been receiving an "error saving draft" message when trying to upload a document to the knowledge base for a custom GPT. This only happens when I attempt to upload the 11th file. If I delete an existing file and upload a new one, it works, indicating a limitation of 10 files. However, ChatGPT informed me before the creation of the custom GPT that there's no limit on the number of files (only a 50 GB total size limit, if I recall correctly). Will this be resolved soon? It's quite limiting for me. 
You can't always rely on ChatGPT to reliably report on its own limitations unfortunately, that's one topic it frequently hallucinates about.
ask in #prompt-engineering
i did
Try combining documents together and let me know how that goes for you. Im also in a stuck spot atm.
Combining my text into a 7kb file did not work for me though, so im wondering if it will for you.
I tried to create a custom GPT. uploaded documentation for the Godot game engine.
when i asked it to vector index it. I want it to be able to find documents by similarity. Is that a possibility?
anyways, it failed to process it, because it took to long. i asked it for a script that i can index it on my local machine. it did.
I am unable to edit or change anything with my custom gpt because it always says error saving draft when i try and change something on it
when I tried to upload the result it failed to process it. even after cutting the json to contain a vector embedding for a single file.
the cut file
I tried to avoid learning LangChain; but if I will and create such a chatbot, i will open source it. Since many Godot Game developers have NVidea GPUs, they would probably use it with one of the open source LLMs, with gpt4 only as a fallback.
Is it just me or the recent update ruined chatgpt?
I have been trying to generate images using gpt 4, however, everytime I do so, it returns in a error "Error Creating images", and even with just normal site, it seems to be returning in more errors then before.
What prompts are failing? Do you have any custom instructions active?
Nope, it works on incog, but not my regular brower.
Incognito window of the same browser? Any extensions active?
Yeah Incog of the same Brower, I do have a extension activated which is a web extension for chatgpt.
Although I can insure you that the extension is not the reason to why chatgpt returns in errors.
You've tried disabling it and you still get the error?
Well I have done that yes, but I had the same extension before they started making changes to the current site.
It's up to extension devs to update, extensions are common culprits for breaking ChatGPT. Have you tried disabling it recently?
all chatgpt extensions are unofficial
you said it worked in incog which then disables all extensions among other things
I only have one extension on my chrome browser, which I just removed, and as for incog, it works, but it doesn't necessarily mean that its the extension which causes ChatGPT to break/return in errors.
I need an OpenAI engineer. I've built two GPTs today. After I finished the first one, I typed in "testing" in the chat bar. "SOMETHING WENT WRONG" was the only answer I got. So I again starting from scratch built a second GPTs. After spending hours of research, and inputting files, I can't even get the GPTs to respond to a "test."
You're unlikely to find OpenAI staff here, you might try posting to #1070006915414900886, as they do appear there more frequently. Or, we can try to help troubleshoot as fellow users! Do you get the same error in normal new chats, or is this only affecting your custom GPTs?
Thank you
I was wondering if there is a way to use say Google voice assistant to start ChatGPT in voice mode and continue without touching the phone by using only audio commands...
for example say your driving and want a hands free experience I would like to be able to say to google instead of just "Hey Google open ChatGPT" rather maybe something like "Hey Google open ChatGPT in voice mode"
I guess or if nothing else maybe the ChatGPT app should auto start the voice mode when the Android device is in driving mode ...
I guess this may be a feature request more than a request for help but i was unsure where to post this so I figure here is a good place to start
Im getting: "Something went wrong. If this issue persists please contact us through our help center at "
I have seen that several times in the last couple hours doing some image gens
I know a couple days ago they was hit with ddos attack and that may have also been the cause of yesterdays outage and the ddos attack may not be over ...
I do hope it is over as maybe they can cause a nuisance but I wish they would not I be trying to work
glad you solved your issue
Why does my GPT always show Unpublished Changes even after I update its files and click "update?"
Just use iOS
never my friend Android lifer here
I am just to broke to consider apple products honestly
Then you probably need to wait another two years
Just overdue your credit limit
Who cares if you can pay for it as long as you have it
yea my credit score is not a number it literally comes back as unscorable but anyway
are you saying you can start the ChatGPT app in voice mode using voice commands only on iPhones already at this time?
Idk about android but i am using the voice feature for about a month
Has anyone found the best file format for ChatGPT to read from?
the voice feature works just fine on Android and that was not my question
.txt but seperate them for each topic. In doubt just create sub topics
makes me think markdown might be much better format overall
I think that would be in Google app > Settings > Google Assistant > Shortcuts, but I don't see the voice launch option there, so I imagine that's something OpenAI would need to add, unless Google maybe blocks hooks like that for "competing" assistants. But, in the meantime, do you know about the shortcut within the app icon? Long-press on the ChatGPT icon and you'll see a popup that launches the app in voice mode. Not the hands-free solution you're looking for, but it makes me wonder if you could get there manually somehow through some Tasker/IFTTT integration since the functionality is there as an app shortcut, at least.
hmm on mine it does actually say "Voice Mode" when long pressing it just says "ChatGPT" but if I click where it says "ChatGPT" after long pressing it does in fact go straight to voice mode... so yea there should be a way I think
thanks lots man very good insight to share given my question, appreicate it
I'm encountering a persistent issue with verifying my domain for OpenAI settings and would greatly appreciate any guidance or insights you might have. Here's a detailed rundown of the problem and what I've tried so far:
Issue Description:
I'm trying to verify my domain with OpenAI by adding a TXT record in my DNS settings on IONOS. Despite following the provided instructions, the domain verification continually fails with the message that the domain could not be verified or that the verification failed.
Steps Taken:
-
TXT Record Addition: I copied the verification code provided by OpenAI directly (using the copy button) and pasted it into the DNS settings on IONOS. The code looks like this:
openai-domain-verification=dv-stringofcharacters. -
DNS Settings on IONOS:
- I added the above code in the 'Value' field of a new TXT record.
- For the 'Host Name' field, I tried leaving it blank and also using "@" (representing the root domain), as there were no specific instructions for this from OpenAI.
- The TTL (Time to Live) is set to the default of 3600 (one hour).
-
Waiting for Propagation: It's been several days since I first added the TXT record, well beyond the typical propagation time.
-
Re-Adding the Record: After initial attempts failed, I deleted and re-added the TXT record to ensure there were no errors.
-
DNS Propagation Check: I used online DNS checking tools to confirm if the TXT record is visible, but with no success in verification.
Questions/Concerns:
- Is there a specific format I need to follow for the 'Host Name' field that I might be missing?
- Could there be an issue with how IONOS processes these records, or is there a known delay I should account for?
- Are there any common pitfalls or less obvious aspects of this process that I might be overlooking?
Thank you in advance for your time and assistance!
Hey there,
I'd like to set up an Assistant with big files, but I'm concerned about the pricing.
Would OpenAI retrieve unnecessary information from my documents to fill in the Input Window (16K)
or would it pick just enough (say 200 token)?
I could control the input price with langchain... :/ but I was never sure if the answer was completely relevant
txt record changes should always be instant unless the host system has a delay. are you sure the code did not change at some point for some reason?
Does this mean there won't be an Input charge and I'll be charged 6 dollars per GB of information no matter what?
token costs still go on top of that
that is the retrieved input , let's say ~200 tokens of information out of 1GB
- output tokens
- daily storage fee,
right?
but how can I chose the input limit?
"The tokens used for the Assistant API are billed at the chosen language model's per-token input / output rates and the assistant intelligently chooses which context from the thread to include when calling the model"
What do you mean by the code changing... Ive attempted several times.
I think OpenAI value stays the same.
did the openai generated code change for some reason at some point
It doesn't seem to change. no
I just want to make sure it intelligently cares about my wallet too
then if you verified it from multiple dns servers that it matches it should have worked
currently its using all the free capacity and you cant control max tokens or used data from files
Its same value. Hasn't changed. First website I am attempting to verify for OpenAI Builder profile
did you for example use nslookup io and checked with few different dns services there
I don't know what that means
I am using IONOS
This what you mean?
The commenter is suggesting that you use a command-line tool called nslookup to check the DNS records of your domain from different DNS servers to ensure that the TXT record for OpenAI's domain verification is properly propagated and accessible across the internet.
google that service: nslookup io then check all the tabs
that's not from the site and its the non authoritave answer
According to ChatGPT I shouldn't need to mess with nslookup
Using IONOS DNS Management: If you've added the TXT record correctly in your IONOS DNS settings and waited the suggested time, you've done your part.
If I am using DNS Management on the provided host then that should be enough...
Hey everyone - I'm new here and slightly terrified as I'm completely out of my depth. I work for a non profit and due to the economy we're down to just me despite demand for our services increasing 7x. I'm looking at how I can use AI to automate and streamline business processes across gmail, google docs/sheets/slides, xero, squarespace, mailchimp (likely to update this platform). As we are government regulated I have to create so many board/member strategy and iniatiave proposal docs as well as having to keep member contact detals updated constantly. I know I can use chatgpt to help with a lot of this. I just have no idea where to start. I was recommended to look in discord, can someone point me in the right direction of an SME (emphasis on the S) channel that can give guidance on these sorts of things? I tried searching but I suck at that too. Sorry for asking such a basic question, I appreciate any help at all.
its for verifying that all services have the info and in correct form
I think you can use GPTs and Zapier.
Thank You!
Hello! I have been playing with the assitant API and the retriever, I alwas get empty annotations. Anyone else havng this issue?
how do i fix a bad gateway error?
I just had the same error. Refreshed and it was back
Is it possible for assistants to create/modify files via retrieval? For example, working on one that can revise resumes for the user for various roles.
Hello! I have a strange problem using ChatGPT: I can't share any of my "GPTs" created by me. Every time I try to publish them using "Only people with a link" or "Public" the system gives me this error:
"Your GPTās name is restricted for public sharing. If this is your organizationās name, please verify your website via the Builder profile in Settings."
And that doesn't makes ANY sense because it's my own name what I'm using: JAVIER LOPEZ
Why I'm I receiving this error/restriction/whatever?
Thank you in advance!
@snow sable GPT 4, or maybe just the customgpt's(?) can edit files, but it is limited. It has to edit the files through short python code. So, it has to be able to scan the document and easily know where to put text. PDF forms are a good example. Each text field is named and so easily referenced. I haven't tried many other types of files, but its worth exploring
I don't think they can create tho
Hey, I'm curious. Are you building this inside ChatGPT or somewhere else? If it's inside chatGPT, provide a link so I can check it out?
Hello everyone, does anyone have an example of a Node.js API for a custom assistant? I'm struggling with the feature to trigger functions.
Are we limited to one Action? I can't make more than one. Probably should be called Action, not Actions
My user screen is frozen like this. I have been experiencing this issue for some days, but it's been unusable today, completely. Any solutions here? Tried clearing cookies, browser history, etc.
Hello everyone, why do I keep getting Error saving GPT when I save the gpts I created. I chose Only people with a link. When I select Only me, it does not appear. Why is this? Have you encountered this situation?
I am also getting this right now
And it seems that gpts often has problems in the past few days. I remembered, I had this problem yesterday too.
I highly recommend at this point in time you don't trust ChatGPT in terms of your work. In fact, if you ask ChatGPT itself, it will tell you and advise you to get an actual professional to do the jobs you need. The reason being is, again, not to insult the AI. In fact, everything you're reading now has been done text-to-speech through ChatGPT.
We've seen in the legal world, and also the government world, lawyers using AI and getting blasted from various judges because sadly the AI, though it seems very reliable, and believe me, it is reliable, but it does make mistakes, and there is also a memory issue with the AI. It does not think like us, so in terms of long-term memory, it can pull from older parts of the chat, but it will also forget as well, so sometimes you might have to remind it.
In fact, what I also do, like let's say if a chat becomes corrupt or severely outdated, you could actually copy-paste the entire chat and bring it into a new updated chat, but again, you're going to have to remind it of very specific things, otherwise it'll forget. In fact, I've tried uploading union documents and non-government-related contracts, and sadly, sometimes ChatGPT will make things up. You see that a lot with ChatGPT in math, but yeah, it's not reliable, and one of the last things you want to do is send out a government contract with wrong information.
It's sort of like if you're using ChatGPT for government usage without permission, basically because everything has to be recorded.
ChatGPTās thoughts on this matter ChatGPT: Your caution regarding the use of AI like ChatGPT for professional or critical business tasks is well-founded. While I strive to provide accurate information and assistance, my responses are generated based on patterns in data and are not infallible. In matters requiring professional expertise, such as legal contracts or government documents, the stakes are often high and the consequences of errors significant.
Professionals in these fields have training and experience that AI currently cannot replicate. They can understand nuance, context, and the subtleties of human language and law in ways that AI, at this stage, cannot. AI should be used as a tool for assistance and not as a replacement for professional judgment.
For someone working in a nonprofit, it would be advisable to use ChatGPT as a supplementary tool for brainstorming, drafting, or learning, but critical tasks and decision-making should be validated by a qualified human professional to ensure compliance, accuracy, and accountability.
how do I train my own ai on my own data set? and how much does it cost?
Hi ! Can a chat be "broken" ?
My GPT4 is endless giving me an error in this chat.
8 errors in a row whatever I do or I ask. Image or text.
All my others chats are still working
same error š
I had that error when I tried to make a GPT that had the name "ChatGPT" as part of my name for it. As soon as I removed ChatGTP, I was able to save it.
Thank you so much for the extensive reply and your experience! I definitely hear you on using ChatGPT as a supplementary tool. My focus is definitely reducing repetitive/administrative tasks, not anything business critical or regulatory. Especially as the risk for getting it wrong is pretty severe (as in prison severe). It's my member services and content marketing teams who were let go so getting all the different email/finance/edm/social platforms to connect and help automate tasks is what I'm exploring. If I find an SME solution I'll post it here to help others.
Thanks!
Ahh, was that your issue too?
Has anyone had problems using Copilot for Xcode. It acts as a companion app to Apple Xcode offering Chat GPT powered chat, and Github Copilot powered code suggestions. I have Chat GPT Plus which I understand gives me access to GPT 4...
I have added a Chat GPT API key to the settings in Copilot for Xcode...
But when I press Test (see above(, this is what I see...
chatgpt plus has nothing to do with api which has its own restrictions and payment
you need to be a paying api customer to have access to gpt-4 models
Dear Community, Has anyone faced this issue : https://discordapp.com/channels/974519864045756446/1173587557431390219
So how do I get to be a paying API customer? And why is it that with mu current Chat GPT Plus account, I am able to access custom GPTs such as this...
chatgpt and api are two different things
anyone knows if we can change our billing details? I'd like to make public some of my GPTs, but using something more brandable than my own name.
no, you can't change your account name
thanks, just realized we can also add a domain, so that should do as well
if you want your brand there, just use a domain
domains are pretty cheap those days.. You can get a domain+basic hosting for about 2 bucks a year
You don't even need hosting to verify. But a dead link probably isn't a thing you want either
Our you route it to your instagram or smth
Yeah thanks. That's a nice idea. I already verified with a domain I just got and I'll prolly reroute to my Twitter or yt channel
Great!
Question. Are you using PLUS or the buisness version? The reason why I ask is because even PLUS members, their records are not private, so keep that in mind. If you are using ChatGPT PLUS on a personal account, consider everything you say open game to be looked at by anyone within OpenAI. I highly recommend if you have a personal account, contact OpenAI and ask the cost of a business account. So that way you keep your information private and your company's information private.
Hi, i hope, this is the right channel. Since the UI change a few days ago, ChatGPT is incredibly slow answering. Do anyone know if i can do something about it or if this will get fixed? I use ChatGPT Plus.
Everyone got hyped because of the update. The load is just too high right now, but I am sure they will scale, or the traffic will flatten out soon
Best time to fiddle with ChatGPT is around 5-10am german time rn
Thank you @astral quest. Okay, then i try to get up earlier. I hope you are right and this is just the hype and not ongoing. ChatGPT is a great help for my work and i dont want to miss her. (In my imagination it is a woman.) You know if there is a spot in the evening or night as well? I am afraid, it is difficuilt for me to work before 10.
Well its worst, when both EU and NA are awake and have nothing to do (mostly in the evening) but other than that, it's hard to tell when it's better. Guess we just need to keep testing
Thank you very much š I try to be more patient š have a nice day or evening, wherever you are š
Germany. You too!
Sadly, each new update for OpenAI, often it will come with new features, but at the same time it will also reduce certain features or make the AI less effective. Ever since Sam Altman met with the World Economic Forum some months back, I believe it was in June, we started to lose critical features with ChatGPT, and it's gotten worse over time. So think about it like Google. Google eventually started censoring its results slowly. So while Google has increased its services over the years, its search function is no longer what it was back in 2007, where you could actually look at the old version of the web page. That's not possible anymore. And also, good luck trying to find something you're looking for due to censored and authoritative sources results.
What critical ChatGPT features have been lost?
There is every feature available, that ever was. What do you mean?
Hi, I have a support question. I am a PLUS user. My UI says that "browsing" is available, however, the model tells me that it can't browse. In Settings & Beta, there is no "Browse with Bing" toggle available. How do I activate browsing?
GPT-4 is itself a multimodal model now on ChatGPT Plus, no more toggles. So go to a new chat, make sure it says "ChatGPT 4" at the top, and you should be good to go. Let us know if it's still buggin for ya
Ah, I was using an old session, thank for the easy fix
Is there a support page here?
You got it! Enjoy the new model it go brrrrrr
Except rn it's Big Traffic so a little clunky
You can share a problem here if you want, and we community members can maybe try to help. Or, you can make a post in #1070006915414900886 -- OpenAI staff are periodically in there š
My longest chat crashes or has an error "Message in conversation not found" after I try to feed it a prompt. Any fix? I dont want to start a new chat as this one has all of the knowledge.
I've heard of very long chats running into issues like this before, unfortunately I think sometimes they just don't work. Are new chats working okay for you? If so, you may need to copy/paste critical info from your old chat into a new one.
For what it's worth, due to the way the context window works and depending on how long your chat is, it's possible that it's already forgotten the oldest stuff anyway.
Is chatgpt plus more cost efficient than paying for the API?
depends on your use case
for gpt4, you pay about the same amount, when prompting up to 500.000 tokens. If it's less than that, you might be better off with the API, else use ChatGPT. Also you should keep in mind, that the ChatGPT interface is much easier to navigate trough. If you are not an experienced, or at least half decend programmer, using the api could be uneffective due to the time you would need to put in to learn everything. Time is money.
My openAI API is bugging, anyone else has issues with it ? It starts writing the same characters infinitely
"An error occurred. Either the engine you requested does not exist or there was another issue processing your request. If this issue persists please contact us through our help center at help.openai.com."
It actually starts answering correctly, but starts bugging after about 50-100 words
right now with me, it tries to use code interpreter, but for some reason it starts answering without even letting code interpreter finish calculations
Hey everyone... Is there a way to ensure a custom GPT keeps only within the "knowledge" base provided to it?
I was giving a try at the ChatGPT creator but I'm having issues with the "quality" of it's responses.
I've uploaded 2 xlsx with dummy data to give it a try, but when I ask ChatGPT about the content of that data, it... makes things up on the fly? It's like, somewhat refering to that file, but half the time it just make things up, while reassuring me it's showing me information from that file. I can correct the gpt because I know the content, but I just don't know what else to do for this to not happen.
Did someone had a similar experience? Any suggestions? I already tried explaning chat this is many different ways.
I want to create a GPT that converts a .gx3 file to a c/c++ code file there i was wondering if any have done that? if not how can i create something like that?
is Dalle down? all the gpts say they cant draw images
Lots of people havinf issue atm
Can you try the default GPT-4 model? Some have said that works if the DALLĀ·E one doesn't
i realized lol
it does yeah
gpts are broken actually
What is going with Dalle-3, I can't create images. It is giving me text
mine is not following its instructions, neither using any tools
there are issues rn
Try in the normal GPT-4 model
Hum wrong channel
š¤
does anyone know how much is gpt-4-0613 ?
I'm having a problem with making a custom GPT what channel is best for that?
probably this one
Well in short:
Trying to make a GPT with actions but the actions don't seem to be saving when i change the OpenAPI spec
the save system gets corrupted sometimes and requires the page to be refreshed. if it still happens after that then check the error console
Been having the same issue. It doesn't save the edits, unless you make some other change. Or you can delete the whole action and replace it with an edited one
Have you guys ever had problems where GPT is just not for the life of them understanding the OpenAPI spec? I give it parameters for keywords, fromdate, and todate and it keeps doing
{
"keywords": "fromdate=2023-11-06&todate=2023-11-13"
}
or
{
"keywords": "2023-11-08 to 2023-11-13"
}
Like come on
What does the schema look like
Because it looks like you are indicating the date parameters as queries
"/LegisInfo/en/bills/json": {
"get": {
"operationId": "SearchBills",
"summary": "Search for bills using various criteria",
"parameters": [
{
"name": "keywords",
"in": "query",
"description": "keyword search for bill",
"required": false,
"schema": {
"type": "string"
}
},
{
"name": "billtype",
"in": "query",
"description": "type of the bill",
"required": false,
"schema": {
"type": "string"
}
},
{
"name": "politicalaff",
"in": "query",
"description": "political affiliation related to the bill",
"required": false,
"schema": {
"type": "string"
}
},
{
"name": "fromdate",
"in": "query",
"description": "start date for the search query",
"required": false,
"schema": {
"type": "string",
"format": "date"
}
},
{
"name": "todate",
"in": "query",
"description": "end date for the search query",
"required": false,
"schema": {
"type": "string",
"format": "date"
}
}
],
"responses": {
"200": {
"description": "successful operation"
},
"400": {
"description": "Invalid request"
},
"404": {
"description": "Not Found"
}
}
}
},
Right off the bat, I will tell you that you need to be VERY specific and give examples
The AI will do all kinds of batS$%&^ nonsense otherwise
Ah ok
it works when i add this to the instructions:
When asked for a bill within a certain time period (i.e. latest bills) use the SearchBills action and the "fromdate" and "to date" query parameters. If the user asks for the latest bills search by last 7 days
Example: {"fromdate":"YYYY-MM-DD","todate":"YYYY-MM-DD"}
SSmall example of mine:
paths:
/hs/:
get:
operationId: getHSExportHistoryData
summary: Get HS export data by date, optionally by country
responses:
'200':
description: Successful response
content:
application/json:
schema:
type: object
properties:
data:
type: array
items:
type: object
'400':
description: Bad request- possibly due to incorrect HS code format
'404':
description: Not Found
parameters:
- name: get
in: query
required: true
description: The names of the parameters of desired data and to be returned in response, separated by commas.
schema:
type: string
enum:
- CTY_CODE,ALL_VAL_MO,ALL_VAL_YR,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- CTY_CODE,ALL_VAL_MO,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- CTY_CODE,ALL_VAL_YR,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- ALL_VAL_MO,ALL_VAL_YR,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- ALL_VAL_MO,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- ALL_VAL_YR,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- CTY_CODE,E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- ALL_VAL_MO,ALL_VAL_YR
- ALL_VAL_MO
- ALL_VAL_YR
- CTY_CODE
- E_COMMODITY, E_COMMODITY_SDESC, E_COMMODITY_LDESC
- CTY_CODE,ALL_VAL_MO,ALL_VAL_YR
- CTY_CODE,ALL_VAL_MO
- CTY_CODE,ALL_VAL_YR
examples:
example1:
summary: Returning city codes, city names, and total values for each month and each year.
value: 'CTY_CODE,CTY_NAME,ALL_VAL_MO,ALL_VAL_YR'
example2:
summary: Returning desired information of the HS code, short description, long description, monthly value, and yearly value
value: 'E_COMMODITY,E_COMMODITY_SDESC,E_COMMODITY_LDESC,ALL_VAL_MO,ALL_VAL_YR'
Being even more descriptive under summary and description really helps
Ok, Thanks!
Enum also is good whenever you can use it
Can I upload yaml files here? I would give you the example of the whole thing, if you like
Hold on, thats an old one. That doesn't work. One moment
Pattern is a regex, btw. As a programmer, I recommend not trying to figure it out. Just get the AI make one for you.
I'm very fluent in programming too
Yes, my brother
Maybe the ai undertsands yaml better? Its generating paramaters not listed in my spec
i got issue with yaml and chatgpt one time. i think it generates fake variables
I added a pdf of the book "Freedom by Midnight", which I have checked through myself, to the configuration of this GPT. However it is struggling to actually parse it when prompted. It is able to grab a hold of big picture ideas like Gandhi and other historical aspects for the book, but continues to return blank outputs and an error when asked to parse through the text for chapter summaries or outlines. Any tips on how to work around this? I have added an example of a page from the book pdf. The pdf is 672 pages long and its file size is 290 MB.
Is anyone elses GPT not sending acitions anymore? mine seems to seemingly not send anything
try ensuring that the text is orientated correctly, it struggles with angled text.
I think you need more detail in the "get" operationID and desscription.
I think the AI decides whether to use the API just based on the operationID.
Mine changed from always errors to sometimes errors. Meaning it can read and analyze a pdf, but a txt file is too damn hard
I am facing the same issue! GPT does not respond anything whenever I write any prompt on my computer
This is great, mine is 'forgetting' parts of code to run.
You would think.
Won't save
Hold on,
I had that, I created a new gpt and the errors went away
Except I have spent days working on this
100% happening the same to me with XLSX... weird part is that it always answer like: "based on the content of X document..." and then throws some random crap
darn it, you gave up one click before it would actually let you save it..
eh, its fine. I managed to get it doing what I wanted again. Sometimes the python code it makes is not so great. I have to either give it a python script or coach it to do what I want
Is there a affiliate link
no, affiliation systems usually come about when there is a shortage of customers and an oversupply of products, OpenAI has the opposit problem.
Does anyone know if you can change the voice style of a Custom GPT? I changed my global settings but it doesn't seem to update the Custom GPT voice that is used.
You can do it via the API, but not ChatGPT as far as I know
Okay, kind of a bummer since I don't worth with the API. I guess it's an idea for the future.
for sure, I can see this makign it's way to GPTs fairly quickly
I can't remember where it was, but I saw someone ask somewhere about the remixing option. Are the instructions really public? Would that not expose and ultimately hurt creators?
But no one would see the knowledge files?
I am shamelessly working to get money out this. They are concerns of mine.
Hello. For people building GPTs within ChatGPT, how are you getting around needing to start a new conversation every time a GPT is updated? How do you pick up whereyou left off?
If it's available to the GPT, then a user can ask about it. If the user can ask about it, if the user is very interested, they probably can find a way to get significant verbatim information from it.
I don't believe that is possible. I start the new conversations.
so it is. and reginald, I have had no issues with this, personally
Actually, I shouldn't be worried about the instructions exposed. Thats fine.
Ah I see. I dislike having to restart the conversation each time. Just wondering if anyone else figured out a workaround.
Hey! I'm working on making a custom GPT. When I go to the configure section and upload files, its fine up to a certain number of files. However, I get to a point where it says there's an error and cannot save. I have tried this with two different custom GPTs, and I get the same error.
I thought it was the file so I changed those up as well.
Is there a limit to how many files you can upload in the configure section of the custom GPT? Or am I doing something wrong?
Yes it would, and I've already submitted a bug report in hopes they remove it https://discord.com/channels/974519864045756446/1173825276023689276
after 10-15 files it starts to hallucinate.
missing features, comment placeholders.
not the amount of days, the amount of files
I couldn't upload more than 5 without it erroring out every time. That was fixed? Nice
i am also working shamelessly for money.
and world peace
i was the one who released the prompt. you can see from the convo here
much earlier
hail, my friend in greed
@cosmic crag you're a cool dude
hah
please,If you want to collect the dalle3 generated images shared by each user of discard group, is there any automated method to achieve it
I don't believe there is
Max of 10 files, max storage 500Mb I believe.
Iām not sure I understand the pricing for my application:
I need to create 40.000 assistants, each with a single roughly 100kb text file.
Does that mean each assistant will cost 0.20 x 0.0001 = $0.00002 per day, and that therefore the retrieval part of of the cost for the entire set of 40,000 assistants will cost around 40000 x 0.20 x 0.0001 = $0.8 per day?
If not, how much will it cost?
please stop spamming, you only need to ask once in one channel
Does anyone know how to download this software? This seems to be the PC side but I can't find it on the web
hi builders, i need some help. i am trying to make a custom gpt that query an api but the server of this api has location restriction based on the error message i am getting. Even through i am using a vpn, and changing the location of my ip. i still get the same 403 message errors message related to location restrictions of the server. i ask support to the provider he said that because cdn of open ai might be in usa , it will always get the error message ?does anyone know how to bypass this?
@charred parcel sorry to call you out like that . do you know the answer to my question or if not who might know it?
you are trying to call the API from within your GPT?
yes , i am . to give you more context. i am using the new feature "custom gpt" i am having this gpt calling an external api for data. investigating issue wit the support team for the service. they than their server bans request from usa. He then advise me to ask if queries are sent from a usa ip?
the queries come from the USA
that explains everything then. my vpn is pointless because the query is coming from your ip not mine š
yes everything that happens with a GPT is server side, your own IP, VPN or routing has nothing to do with it
interesting edge case. My issue was related to financial market data from an exchange that is not vetted by usa
@charred parcel thanks for clarifying this Rob š
That's a preview of the GPT store, which is rolling out later this month š
https://openai.com/blog/introducing-gpts
getting stucked at Seaching my knowledge a lot of times in while interacting with my GPTs
should i wait for few hour's for it to process the data first.
even regenerating response is of no use.
I want GPT-API to parse a string and return the values as a JSON array. gpt-3.5-turbo also does this. However, gpt-4-1106-preview returns as a formatted string [see image, cause Discord format it too] no matter how I formulate it.
What is the trick so that GPT-4 does not set any formatting characters there?
how many max words is advised to put in one pdf document to feed the knowledge base? Is there a limit of how many pdfs can i upload?
anyone else getting the "error saving draft" when trying to create an API based action on a custom GPT? I cant for the life of me get it to save the API key without giving me the error saving draft message.
Hello, I get this error and the messages do not load in all browsers except the mobile network error. can help mi
Does anyone know how they got this button to popup a form??
#dall-e-ask-here message
The button pops up a question screen for me! Just tested
Yes it works for me as well. But I am curious how they did that? I would like to do it on my server
Oh sorry I misunderstood your question š
Hey, I recently got GPT Premium on my account. Basically, I want to convert my lessons into flashcards. Is there any way I can do it with ChatGPT
Just copy the texts in and tell him to create some. If necessary, tell him what he should sort it by or which points are important. This will probably already produce a good result.
is this on mobile app or web version?
Well, it's a science lesson and it has a lot of details. When I ask simply, it tends to create general responses that don't go into the details.
It depends on the quantity, but you could upload the text as a PDF. But when it comes to indexing the law books from 2000-2023, you'll probably have to split it.
It has 58 pages. Do you think that would work?
Just give it a try. Several times if necessary, because ChatGPT is sometimes a bit moody and Chat 1 tells you that it can't do it and Chat 2 happily starts calculating.
Why the hell does he get stuck in the same place every time? 5 chats in a row. And yes, I tried regeneration, it didnāt help
probably a latex issue. ask it to write in plaintext
I've been getting same error since yesterday
out of almost 10 or so times, it works like once or twice
I'll try now
It keeps crashing...
Did not help -_-
why does it say that?
Developers working with a multi-million dollar corporation skimped on beta testers -_-
web app cropped the screenshots
i need help. please
what do you need help with?
Oh man, peeps, this new ChatGPT is such a disappointment for me. Constant network errors, no creativity, waffly responses, empty responses... despite refined prompts. This is really ruining my flow. Is my experience isolated?
I get a network error all the time in addition to regenerating a response and it crashes in all browsers. and I am paying the extra so that it works better, everything hangs and I don't know what to do, I deleted the temporary cookies, the browsing history, I restarted the modem and I don't know why it continues like this. thanks
there is instablity at the moment, hopefully it improves soon
I hope so, why pay for this bad service and also you can't talk to help because it closes the chat by sending you here. and my English is bad so I don't understand much, I need to translate everything they explain. Does anyone know if we will have compensation for not being able to use it while paying for a better service? thank you so much
the service is functional, it's just slow with a little instability. There is huge demand on the systems.
ok thanks
That's a risk when paying for alpha and beta products . Everything detailed in the tos.
You can always discuss to support@openai.com
thanks
Apologies if this isn't the place to post this, but it seemed appropriate. I recently upgraded to 4 and am trying it out. I am using Google Chrome. I am using it to review and comment on already written documents (I am a professional writer and published author). So, I upload some Word documents and everything works fine for a bit and then I get a strange error. I have attached a screenshot. It's like two quotes in brackets and if you hover over it, you get what is in the screenshot. The numbers change, but the "tether browsing" bit doesn't. I have searched but not found anything about this error/issue. It basically dies and you have to Regenerate, but it does the same thing over and then gives the same "Regenerate" again and again. I have tried clearing cache, etc. Any ideas?
I got this persistently. Deleting the whole action and starting over cleared the issue.
Greetings. Any updates as to when 4.0 will become faster in response? It is curiously much slower than free version
Is anyone else getting consistently Network Errors?
Thanks. I'll try that to see of it helps.
well, unfortunately this isn't working for me. are you using basic or bearer auth for your API?
Neither for the case which I got to work. Have not tried again with auth yet.
To whoever at OpenAI that added this
Thank you
Can someone explain to me why is my DALL-E webpage blank?
getting them too
Good, at least im not banned š
Unexpected turn: You are both banned
that would be sad
Getting Network Error too, its almost like OpenAI are ratelimiting the API, like trimming the tokens it can use but doing it in a way that just breaks the output rather than cutting it short or dumbing it down..
If your asking for Dad jokes maybe you could š
I keep getting errors like this, about 3 times today, interrupted and then this
When using my GPT built in GPT Builder (QuickSilver AI - R.A.G DocuMaster) I keep getting errors when trying to execute a chat. See loom video here:
is there a way to force ChatGPT to use Advanced Data Analysis without trying to use a request like "Use the data analysis tool for all future requests" inline with the prompt? Still not sure if it's being used when it says it will. It was much better at coding for my needs, now it seems to default to 4, which I don't want. I'm ignoring the ton of error messages and 3x slow down, which makes it quite hard to use now...
I feel not being able to use the ADA tool specifically is a big loss to the recent upgrade. Not sure if anyone has found any work arounds?
They do have a dedicated ADA-only GPT available š https://chat.openai.com/g/g-HMNcP6w7d-data-analysis
Seemed to respond initially how it had previously so fingers crossed. thanks so much
Sure thing! And regarding your original question in general, when you asked about putting a specific request inline with each prompt: are you familiar with custom instruction functionality for your account? That is, the option you have for your account settings to input instructions that apply to all your future chats? The model keeps CI top-of-mind for everything it reads from you and writes for you.
You could also make a custom GPT with a more "beefed up" version of this idea
chat.openai.com/gpts/editor
Be aware that you can't stack custom instructions with custom GPTs--it's one or the other. And this includes the ADA GPT I shared above--custom instructions only apply in baseline GPT-4, GPT-3.5, and GPT-4 Plugins chats
āā¦and head to the [š no access] channel to see their reactions live!ā
I thought the cut off training date was now April 2023. but it keeps telling me its training is now to January 2022
gpt4?
yep - turbo
its april 2023, its possible you just got it to hallucinate if you begun to question that specifically
Hello, I am new here. Can I request for help? Is here the right place for this? I am a plus member having a lot of errors with gpt4 ... and even with errors my quota has reached it's limit.
Just asked it again - he is what it says, word for word. 'My functionality is based on the general knowledge and capabilities of OpenAI's GPT model, up to its last training cut-off in January 2022'
start a new chat and ask when the queen died
i am getting load failed , error in stream constantly
Server overloaded right now. Even if not received answer it count as 1 prompt
but it can just ask the web for that - as its multi modal. But I will try
you can see if its using web or not, its been 2023 april for me almost 4 weeks
ahhhhh - hit the dreaded cap !!! dugh!!!!
Do 1 prompt every 3.5 min and you never reach the limit 
do these incomplete answers with errors count towards the usage cap for GPT4? would be rediculous if they do
they do. As do failed image genrations, or refusal do do anything - any interaction counts towards your cap
pardon my French, but that's horseshit.
Its your input that is counted I beleive
Maybe if it prompts you to "Regenerate" the response it is counted as the same interaction?
Is that right. Good tip , ta
Regenerate count as prompt
That is a major flaw
dunno - good question
I find I can easily use my 50 questions, when in reality I have only had maybe a third of that in actual valuable interactions
That means they are saying the error is caused by the user
Good tip - is never wish it a 'good morning' or use a question to say 'thank you' - you waste a question, and it does not care (I hope!!!!)
You can use that before and inside a prompt
I think it have an impact someone told me long time ago. I can't confirm this
I have activated chatgpt plus, but I cannot use gpt4.0. I cannot send messages in 4.0, but can only send messages in 3.5.
If you just buy it , it can take few minutes before being accessible
@rigid muskBought yesterday
Does it work on phone app?
I havenāt installed it on my phone. I donāt know if it can be used.
It works great, it was just as test . Else can you try in incognito ?
incognito ?I'm not very good at
guys quick question. Im getting an error when saving drafts in gpt builder whenever I try to save an API key. In chrome's console, I see this message: r with Permissions-Policy header: Unrecognized feature: 'document-domain'.
Any ideas what this issue might have to do with?
what do i do
I cannot remove, edit or save a custom gpt's knowledge
Anybody have any idea what files get absorbed by customGPT vs need to be referenced later. I added a text file with my story in it and when I asked about it the GPT didn't show that it was referencing that infor when answering me about it but if I ask questions regarding other files it has to search. Is this a size or format issue? We really need more transparency on how these systems work in order to account for behavior of GPT when formatting and feeding information. Still amazed by it but there are some black box questions I have.
Hey y'all š
I got a quick question. Just paid for a GPT-4 Plus account so I can dive into creating custom GPT's. Managed to interact with the GPT creator for about 20 minutes, testing the GPT as I was going along, and then before I knew it I got a "max cap" usage message saying I have to wait until 7pm before I can use it again. Is that normal for a paid account?
ok, it seems like they are capping the customs gpts
it happened to me too, i was using a custom gpt to help me review stuff and i was blocked suddenly
This happened to me an hour ago. Was it around that same time that it happened to you?
More or less, 30 min ago
Ok, well it's reassuring to know it's not just me. I imagine they'll scale it out more in due time.
i think i can keep using gpt4 the normal one, gonna use that to help me finish my review.
My GPT involves image upload interpretation so I was worried all the testing burned through an allowance I wasn't aware of.
No I'm unable to use GPT-4 Chatbot as well. I get same message to try again later.
I got unblock from gtp4 thou, idk what is going on :(
My gpt was just explain me stuff like a professor, so i dont think i burned it out
but was blocked too
yea guys, we're here for the same issue i think
Sadly I still get this on the regular chat bot and also in the GPT creator.
Strange, I was blocked at the beginning but already unblocked recently, who know if it would block me again after a few request
i have the plus, but i keep getting network error and it's really slow while outputting the answers
it happens mostly when it uses the browse feature and it has to output a long text
Im in a hurry cuz im reviewing stuff for my exam tomorrow and phone/ios version seems to work fine, even with the custom gpts, maybe you can try it.
i thought that Plus members had more bandwith and these kind of things didn't happen.....
more bandwith with gpt3.5 only it seems like
i'm not gonna pay 20$ per month if everyday we get this situation.. yesterday was working fine
i can bear with it if its slow, but come on, a block? Also im just using it as a tool for study, not processing large amount of data or anything
Sounds like the GPT creator hype might be soaking up the platform.
Hey can we send excel to gpt 4 ?
I do agree with this. It's my first day on plus and not a great start. We pay a premium after all, so we should be the least affected user group.
Seems like another massive outage ugh
is it normal? 29 running at same time
anyone knows why i cant find my custom gpts when i search on google? can you guys access it? https://chat.openai.com/g/g-a4qgsKcBL-maximus-bodybuilding-advice
There's no Store yet so there's no central repository on ChatGPT yet, you can post to #1171489862164168774
yeah i know but when i google things i sometimes find custom gpts, but even if I search my exact models they dont pop up
it's luck of the draw, Google decides what it includes in the Google index
okay, its just that sometimes when i go to edit its switched to "only people with link"
make sure you set your Builder profile in Settings
can you explain further?
do i have to have a domain?
Did they remove the ability to edit custom gpt's? I don't see the edit button anywhere anymore.
In Settings you can set your Builder Profile which is required to make a GPT Public
yeah but its not much you can change there right? I can only cvhoose to toggle my name off or on, and its on
ok, then you're set
yeah i think it's an issue. I tried using the old link to the editor
https://chat.openai.com/gpts/editor/[g-xxxxxxx-name-of-your-gpt]
blank screen.
okay TY
can you see if this link works for you? https://chat.openai.com/g/g-a4qgsKcBL-maximus-bodybuilding-advice
working fine
ty!
Same
I can't edit my GPTs anymore but I can create new ones
hello! i am starting to mess with making my own gpt but when I go to my gpt's, there is no option to edit. is this not a feature?
I think they may have disabled it for some reason
it's disabled for everyone, we don't know why
It is, but the button is gone
Bummer that the edit button is gone!
I just came here to ask about the editing š¤£
it's disabled for everyone, we don't know why
Ya. But also I see that link sharing is too?
So perhaps it's all accidental or broken rather than them not wanting us to edit them anymore
I did notice this morning you can use gpts in the app but yesterday you couldn't
lol same
same lol i was editing and then i reopened the page and boom, blank page lol
ha ha got to love Discord aye! Least you can see everyone has the same issue and dont spend hours resetting your pc among other things. Thought I went crazy for a sec š¤£
Today's mantra
Is there a way to generate more unique names
For my fantasy adventures it loves Eldoria and Seraphina š¤£
I was hoping for better names fitting the setting like modern names for a slice of life and different time period names
I mean I would like to build a consistent lore table for a Eldoria but itās been inconsistent with those names and their roles
Edit button is back š
Hi, i need a help. I dont see my chatgpt plus chat archive.
Do you have a VPN installed by chance?
i have a vpn application but is closed. Offline.
Does your VPN have a "Web Protection" feature, or something similar? If so, can you try disabling it, separately from disabling your VPN itself?
no. i try 3 different web browser, and clear my temp and cache. But not to be a solution.
What VPN do you use? And what are the 3 browsers you tried?
protonvpn free. and i try chrome, safari and firefox
It looks like ProtonVPN might have some feature called "NetShield"--can you try disabling this if you have it with the free plan, separately from disabling the VPN itself?
oh, it works. i try change my wifi connection, and i connect my mobile internet.
Oh interesting, so there's some filter happening on your WiFi maybe? Do you have custom DNS settings on that network maybe?
thanks, i dont change anything my connection settings. Maybe i reset my internet modem, to be a solution?
That could help, I'm glad you identified the issue though, thanks for sharing! Hope you get the WiFi sorted out š
thanks for help š
What channel on OpenAI Discord can I post my images that is not daily-theme? Was there not a gallery thing here? Where has it gone?
why do I keep getting network error on MYGPT
all of my MYGPT bots get that error after they do their little proccessing and start typing. then they just stop and give me that error
Thx! Really weird that was gone from my channel list somehow...
I've made a Custom GTP for a company - can any of the members login to use it or does everyone need a separate account? I gave the link to another member who is logged into ChatGPT Plus and it said that she had to join Chat GTP Plus to get to my Custom GTP.
gpt's require chatgpt plus account
@Radium we both have chatgpt plus accounts, but she can't get into the custom gpt
did you share it the correct way?
@spark wraithdium I published it Only those with the link.
what error does she see?
@spark wraithdium she is logged into plus and it says that she has to sign up to access the gpt - even tho she is already in Plus
sounds like its not related to the gpt but chatgpt itself, maybe because server load
i jsut got a message saying my conversation is too long and to start a new one š
wasn't really THAT long, but i did have a handfull of files in the DB
Has anyone else had decresing performance with long prompts using the extended context window for GPT4?
ChatGPT is not working for me. Every query gives an error: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
I've tried, different browser, incognito mode, log off / on, rebooting. Nothing helps.
Anyone know why the max length is so small if the context window is 128k?
Error saving draft
Hey I tried making a marketing advisor GPT and even uploaded a lot of documents on marketing logos leads and etc, but when I asked it to make me something it gave me the same answer as normal GPT any ideas why?
it seems the issues i had today have gone.. š really happy
I am SOOOO tired of getting into a long discussion with heaps of context, inquiring as to the results of a web search, and told that it doesn't have the ability to browse the internet.
Dk if the inscructions I gave it where bad
Has anyone had issues attaching their domain name to their builder profile? I've added the txt record to my domains DNS area and allowed enough time to go by for it to propagate (several hours), but I still get a failure when clicking the "check" button back inside ChatGPT builder profile settings.
Or rather, is it more suitable to ask this question in the community API help channel?
Hi, I've logged in into OpenAI with my Microsoft account. now I want to retive my account's password. how can i get it?
Has anyone noticed that GPTs seems to not work quite often?
It looks like you're not using your custom gpt
I am
It says your using normal GPT4, on the left side menu you should be able to select your marketing gpt
Check the other screenies
Ah ok, my mistake
I sent the 2 for comparison
I got confused
in that case i want to point out that your pdfs should be easily readable and be formatted in a way where the text can be selected
Also for people getting the "Error saving draft" error. I find it only does that if you save it as "Only people with a link". Once you save it as "Only people with a link" it won't update again
https://discord.com/channels/974519864045756446/1173672285622763680 Hope anyone likes it
Hm ok, I just downloaded some marketing ones from google, should I try formating in a way its just the text?
You could try that. It might help GPT understand it better
OpenAI refuses to describe how the 128k model actually works but community experiments suggest it's like a "virtual" context size that selectively compresses parts of the full 128k tokens into a smaller window. It works well for simple queries but can't seem to recall the entire context verbatim at once. It might be that the real context size is 8192 with half reserved for the output
Has anyone had issues with actions in custom GPTs when sending data via a POST vs a GET? I created a service with 2 endpoints, posting data to be saved and getting data to be retrieved... Getting data via a GET works fine but the POST fails. The console on the server shows the data to be saved in a querystring in the request and returns an error saying the data is not there - though it also shows it was in fact a post. It appears the JSON that should be sent in the body of the request is actually being added to the querystring instead even though it is sending as a POST... Does that make sense? Anyone else see this?
hey everyone. having a weird issue that someone suggested may be something native to openAI, was looking for some insight. I've been working on a discord chat bot and have a really hard time giving it a personality that isn't "formally helpful". One example is no matter how I write the prompts, I can't get it to not start answers with "based on the chat history" or "based on the search results". Using gpt-3.5-turbo-16k. Is this issue just the nature of the model?
Anyone know how to get rid of the preamble before just about every answer? I'll ask it to design something for me and it'll always start off with a paragraph about why I would want to generate such a thing
Jeff, the purpose of the Actions is to retrieve information from an external source without resorting to searching with bing. I would think it is only meant to use GET requests. Giving AI the power to POST might be a scary thing at some point
GPT-4, by the way
O colonel, my colonel, I have tried hard to make it skip that kind of formality, but nothing I have tried has worked. Pretty sure its caked into the bot for it to inform where its getting the information that it is using when it is external
Darn. Yeah, I hope they add some kind of setting to hide that and make it purely internal, if that's the case
Yeah. I think part of it is where the responsibility of false information lies. If chatgpt just says it, it is the source of false information. But if it says it got the information from a specific source, that shifts the blame away from chatgpt
This is an annoying problem, I am also working to solve.
how should I be building GPTs? ive been using the create tab but should I type in on the instructions part directly?
I've been using a mix of both, chatting with GPT builder tends to delete previous stated instructions
Hello
Hello
I want anybody give me api key for use in my bot in discord for help peoples
Bec i am leaving in Egypt i can't export api or create api key
yea, coz I was having that problem where it would give the exact same answers as the normal GPT, I tried adding instructions for it to rely solely on the documents provided + i asked it to refer more to the documents but that didnt change anything, you thing you could help me out?
Uncaught (in promise) FatalServerError: Cannot have multiple custom actions for the same domain
Error saving draft
Anyone else run into this?
I have the problem too, The only thing I found that works is to delete it and create a new one
You can wait it out for the issue to be fixed
detail: "Cannot have multiple custom actions for the same domain"
Yep, I get the same error, my gpt has multiple actions for the same domain. I don't know why the error says that
I've made 4
Am I blind? I don't have the option to make more than one
I have multiple paths in my action
Yeah, I was just starting to make my own API as a kind of API aggregate to access more. If I can make more actions, that would be brilliant
I'm assuming it's a future feature
Well, the amount of API's I am using would probably hit any kind of limit they'd have. Hmm.
Why does my custom GPT have a different voice? Itās set to Sky, and it works fine, but when I open up my custom GPT, it has a different voice.
I don't think the voice of the custom gpt's is changeable yet
i have chat gpt plus how do i access plugins
It doesnāt use the voice you have your app set to?
Nope
Changing the voice is a local change right now., and I don't think they are able to make local changes like that to the custom gpt's. But my understanding of OpenAI's sytem design is very limited, so I could be wrong.
are medical related GPTs deleted??
Why, did yours? I was actually looking forward to those
no, not at all, well, I did JUST start with it, but some1 in #1174144724093906964 said they are deleted so if they are yk, i wouldnt even do them
I did mine but im still expanding on the knowledge base, u wanna try it out? @cosmic crag for now it only has like 2k pages of anatomy and 2k pages of pediatric, gonna add more tmr but its rly late for me. Just gonna do 1 last update and I can send it to u
Heck yeah, I'll try it
alright
i tested it w my dads knee condition and it got it right lol, but lmk, ill send it you in PVs in just a minute
sounds good
Sus
I highly recommend you donāt vision image in a large chat it causes the AI to malfunction.
click Learn more š
Hi! I'm trying to open a new openai account, but verification email is never received. Any known bug related, please?
have you used the email or phone number on an account previously?
No, I didn't. In fact, I could create the account, and land on the page that says "Verify your email. We sent an email to x. Click the link inside to get started."
What changed in the ToS? At least tell us what changed
instead of having to read through all of it again...!
they did if you press the learn more
oh I went directly to ToS š¦
Oh, well, leaves much to be desired still lol
here is the eu version of changes https://help.openai.com/en/articles/8541941-terms-of-use-updates
Someone has access to my email and is using my chatgpt plus account, I changed my email password but they can still use my chatgpt account, how can I solve this problem?
I can't change the chatgpt password because the first login was with google
Did you log yourself out of all Google devices?
you need to revoke chatgpt access from google settings
How
google "google revoke access"
You're looking for something called App Authorization or Access, Google Privacy settings should be very user friendly (at least nowadays) and will guide you easily
Google Security Checkup š
It shows me the openai acces, but if i revoke it, i still can access to chatgpt so he can to
it may take some time to become in effect if openai does not check it all the time
Can i add password to my chaTgpt accaunt
Is there an option to stop ChatGPT from showing my full name on the bottom left? seems kind of like a privacy oversight
nope since you used google login
its only visible to you just like all the chat threads above
Yeah so I do a lot of stream content and having it there is not good, its a personal problem but I wish there was a way to hide it
But why
not sure why there isnt
What's the meaning of the icon with the number 3 next to my GPT?
that's how many uses it has had
use adblocker or some element hider to hide it
ok, cool thx. Really quick answer.
Hello guys , can you help me ?
What should i write down here ?
Hey guys, my Plus subscription ended yesterday and was not renewed idk why. And This morning I do not have acces to Plus anymore and I have to sign up for a wait list but Im using Plus since February. Any idea ?
See here #openai-chatter message
thanks
Question -- So...I had a session going in 4. I hit the usage cap. It asked if I want to continue with the "default" and I said yes. However, now it is a new day...but it says 3.4 in the upper left corner and the paperclip "upload a document" option is gone. But if I ask it what version the session is using, it says 4. How do I then get the upload option back? But also...actually...every time I try to do a prompt, I'm getting "Message in Conversation Not Found". I am in Google Chrome, using an Incognito window. I cleared the cache as well. I also tried restarting the laptop. (edit) I also tried switching Internet connectivity.
once you are switched to GPT-3.5 you cannot switch back. So you will need to start a new chat, maybe copy paste the last few messages in the conversation to get going again
AGH that's annoying. Thank you.
Though any ideas why the Message in Conversation msg is popping up too? Is it related to it being confused over what version?
it could be a few reasons, maybe server load, things are a little unstable right now which is why OpenAI paused new Plus subscriptions
Hi! I need some help mates! Because my debit card was empty, my GPT Puls has been cancelled, and now I am on a waiting list to be able to use GPT4. I have been a user since March, and I rely on GPT4 a lot in my work. Is there any way to contact OpenAI and kindly ask for a correction? (the debit card is filled with money now, ready to pay)
Only support@openai.com can answer you
be prepared to have possible trouble with the support as officially only credit cards have been accepted
Crapiddy crap.
Hey, guys, any suggestions on what might be going on here? Anytime I upload a profile picture to use, it uploads it fine, then after a while, it removes the image and puts a question mark. It accidentally changed one of my profile pictures that it created by accident. Every time upload it the same thing happens.
Just got hit with another max cap usage error when trying to build a custom GPT.
Has anyone else experienced this around about the same time as me?
Using Create for custom GPTs does count against the GPT-4 usage cap
anyone know whats up with a link to a GPT that keeps asking for sign-in. When signing in it falls back, and asks for sign-in once again.
anyone getting this?
Our systems are a bit busy at the moment, please take a break and try again soon.
I get this message everytime I try to login into chatgpt, howver it works 10% of the time in private browsing mode.
Safari, firefox, chrome all having this issue. & no amount of cache & cooking clearing is helping
it's because servers are loaded to absolute capacity right now
Understood, thanks!
please, don't request for ways to circunvent OpenAI's usage policies
There's an no longer worked on open source project called xbrowsersync. You can self host a server and use browser extensions or ios / android apps to sync your bookmarks with it. all communication is over an API.
I don't like the existing designs for the apps, but I also don't know enough about programing to change them.
Might Sep Gepete (that's how I call ChatGPT) be able to help me build my own apps (even disconnected from browsers) if I can give it access to my API? I probably would need a subscription for that, right?
you would need to code a solution and use the API
yeah, but I mean if ChatGPT can help with that. I saw that it can now also access APIs and read documentations you feed it, but I'm not sure how I can access that functionality
you can create a GPT but what you want to do is probably beyond the scope of building a GPT
okay, thanks!
Anyone else getting this?
this is an OpenAI Discord server, you would be better on a Bing forum for help with Bing
Why is this the case tho
it's just the way it works
Hopefully they add a way to switch between
I've created a GPT with transcripts of my youtube videos. I just want to search and tie videos together. Each transcript starts with a video title, date, and url. Repeatedly, I've instructed the GPT to prioritize newest videos. It keeps giving me old videos. Is there a different way to format the information or instruct the GPT?
my GPTs dont show up on google when i URL search, you guys know how to fix this?
Whenever I try to upload a PDF or text file to my GPT, I get the error "Unable to upload FILE". Do I need to do something to enable doc uploads in my GPT?
you guys think making a GPT that helps people studdy for SATs and ACTs would work? I kinda need that myself lol, dont wanna start doing it if it wouldn't work
Anyone having difficulty forcing a GPT to lookup in the uploaded docs and browsing internet?
I have a step where it is suppose to lookup, I tried all my typical tricks marking it important and vital and it just don't do it. When I stop it and ask about it, it acknowledge and then proceed looking at the document. But only if I ask during the conversation. Not in the initial instructions.
Anyone knows if is possible to subscribe to ChatGPT Plus using a referral link with trial? If it is also suspend⦠what happen with the link? You loses it?
no new subscriptions can be made right now
Thank you! So⦠what happen if someone use s referral link?
they can't, because nobody can subscribe right now
Thank you very much
QQ - has anyone gotten their access to chatGPT plus cancelled? Have been subscribing since the start. Build my first GPT yesterday and today it seems I have been down graded to the non-paid version and only get the "sign up for waitlist". Wonder if that is a mistake connected with the temporary stop of new sign ups?
yes some but not everyone
Any pointers re what to do?
did you renew your subscription manually or was it automatic?
nope, and there is no clear pattern why its happening to someone
it was automatic
Thx - I chatted with the bot. Any other way to get in touch? I understand that it is not really a consumer company with customer service etc.... on other hand I am also paying and would like to get the service I pay for....
email but i think its as bad as via the bot
ugh - thought so but had hopes there might be a semi secret discord channel ....appreciate the help
Are GPT instructions meant for the user that is utilizing the GPT? Or for the GPT and how it generates content?
my name is Ćner from Turkey. I use GPT-4
Since yesterday, I am receiving this error message: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
how can I solve this
For the GPT, I it's the equivalent of the system instructions.
Hello Ćner, do you have a VPN installed by chance?
Yes I have but it is offline
What VPN do you use?
VeePN
It looks like VeePN has a feature called "NetGuard" -- can you try disabling this feature, separately from having the VPN disabled itself?
I couldnt find it but I completely uninstall Veepn
restart chrome but still same error
Veepn was a Chrome extension.
Ah, so you don't have any VPN program installed to your computer itself? Can you try in a new install of Firefox, with no extensions active at all?
ok
Yes, in Firefox it worked
Thanks for help. But even I uninstalled the extension, why it is not working on Chrome? maybe I should clean cookies and history etc
Excellent, that's good. Some users, for some reason, have issues getting it working in Chromium browsers. Definitely try some of that other troubleshooting, so hopefully you can use your preferred browser again, but for now at least it works in Firefox. By chance, is your OS Linux? Just a last little bit of possible troubleshooting we could do if so
it is Windows 11
Ah okay, well that's all I've got then! Hope it clears up on Chrome for you!
Is incredible how i have another person that have access to my chatGPT plus accaunt and cant do anythong about it.
Anyway to optimize the searching my knowledge part?
Can you change your password, or do you sign in with another service? Have you tried reaching out to support@openai.com? They may be your only option
I changed my email password but chatgpt doesn't ask to log in again and the person who has access to my account can still use it even if I changed my password
I sing up with google and i can create a password
I have someone using my chatgpt accaunt and i cant do anything about it
Someone help me
didn't we say earlier to revoke the chatgpt access from google pages? do not relogin after doing it and wait one, possible at the same time contact support
however note that if you have plus, support often will cancel plus access and refund the money on compromised accounts
what is happening to dall-e 3 today, it cannot make any design, all that it produce is rubbish...
I asked for a drawing of twobears in the sunset, following the disney pixar stile (the trend on instagram now)
i restarted the conversation several times, and every drawing it produces is this?
try without saying disney or pixar
i started a new conversation asking for the face of a smiling bear, this is what i got
i think Dall-e 3 is on bad mood
It looks like it's using ADA, to hilarious results š Try using the dedicated DALLĀ·E GPT
https://chat.openai.com/g/g-2fkFE8rbu-dall-e
yes, but it was working all fine until yesterday night. i will try your GPT, thanks...
yes, it worked like a charm, but you know, is useless if thet make a "multimodal" chatGPT and the results are those i showed...
If code interpreter creates a file, can it be ingested by the retrieval plugin to then look up? I.e. if I import a transcript file and transform it with code interpreter, can retrieval access that new transcript?
While using ChatGPT, I cannot seem to add a website to builder profile even after successfully verifying it. The button just doesnāt work. Any ideas? Thank you.
every single time I ask it give me a random SAT question, the first question it gives me is aways this same 1, any1 know why?
refresh
Thank you. I tried that, I also tried using firefox, chrome, edge. with fresh logins for firefox and edge. Nothing worked.
then it just might be another site issue and need to wait it out
The results you showed have the user asking for the image in a non-English language.
The model does its best, but it was really trained mostly with English. It is going to make a lot more errors in any other language, because it has to guess and many very unpopular guesses have not yet been trained out of it.
You can work around this most easily for now by using English with the model.
You can also try to explain explicitly what you want it to do in that non-English language.
to the best I can understand (and ChatGPT might be as confused as me) you asked for something that translates in English as "I need a drawing of a bear smiling face"
While the drawing is actually more frowning, it's a reasonable interpretation of an extremely confused AI doing its absolute best to figure out what you want and how you want it done.
Of course OpenAI is going to keep training the model to be better in every way it can. But you can help yourself in the time it takes to get there by giving very clear instructions to the AI, ideally in English.
I'm pretty sure that user's images were made by mistake by ADA -- I've seen it make similar images just using Python to make basic shape graphs in the past when experimenting. So I think it was just a case of multimodal using the wrong mode, which I'm a big fan of just as a prank from the model š
Is network error a fault caused by me or by GPT? If it is GPT, how can I reduce it?
since dall-e 3 became available, as in portuguese was never a problem, but i made the request in english, and got the same results.
Tapping out! Help, please. I can't get passed the "We've updated our Terms of Use and Privacy Policy" pop up notification when using my laptop. What do I need to do?
Usually there's a place to agree by checking a box. Not in the case. I can't do anything.
Agreed! I love how the AI handled it, just like choosing a word from the list of possible next words, it chose an attempted way to solve the request and made every choice along the way as best it could.
Personally, I am fascinated by less trained glimpses at the AI.
It means to do perfect, it doesn't know what perfect is. This is its guess, so it thinks this is a reasonable, maybe even a good guess.
if the question needs an underlined portion is it possible for the GPT to underline anything?
need this to work so I can study hehe
Can you share a screenshot of what you see please?
Huh.
I can imagine that happening if you continue in the same thread, and don't guide the model back to Dall-E.
Once it decides in the other language, 'I should run python to draw this image myself as best I can try to', then it is very likely if you switch to English and make a similar request, to also try that same already chosen method on the new instruction if it doesn't conflict, like if you don't tell it 'Use Dall-E for this'.
If it did X with method Y and you never indicate you want method Z, it's likely to keep using method Y.
@timber shadow sent šø of screen. Thanks for offering to help!
It looks like you might have a browser extension installed that affects ChatGPT -- I don't recognize that "three line" menu that is next to your chat box. Can you try disabling any such extension(s), and then try reloading?
hi i want to pay monthly for gpt4 but they tell me that i am in the waitlist, the problem is that i really need this gtp4 can someone help me please ?
@timber shadow you're right!!! Thank you very much!
Why would my GPT fail with "Unable to upload FILE" when I try to upload a 30-line text file (with UNIX newlines)? In fact, I cannot upload any file. I get that error every time.
Use the API
For a ChatGPT-esque experience use it at https://playground.openai.com
Anyone know how to command chatgpt to open up the command terminal for listing specific functions?
You know how it does it sometimes. That way.

Custom GPT not acknowledging the knowledge uploaded
Long Prompt with instruction keeps getting summarized why is that? around 8000 words?
Are you saying that you're using ChatGPT, sending an 8000 word prompt, and it's responding to you by summarizing your prompt?
Yes
last version used to do the job , but now even with the same prompt it's summarizing.
Are you comfortable using the share chat feature so I could maybe take a closer look at what might be happening?
And do you have any custom instructions active?
Sure, is your dm open?
I'll keep an eye out! 
Check your DM
Adding website verification txt file to Godaddy. It keeps telling me the code is invalid. Does anyone have any insight on this? Much appreciated!
Hi, I need to generate a few graffiti tag images in a similar style to this image, any ideas for the prompt? I get overly complex images and I just need this simple tags, to use as textures. Tnks
Try uploading the image to GPT-4 and asking for a thorough visual description. Then, try feeding the description back to DALLE.
try and find a famous artist that does graffiti in a style you like.. and ask the AI for an grafiti in their style, might help
I think that is ac tually though weirdly fantastic. I love it to be honest.
not going to lie that is 100% going on the fridge with a magnet, AI's first picture.. all it needs now is a pasta shapes and wood glue one
Whenever i choose MY PLAN in my account settings of ChatGPT, nothing happens. Any ideas?
tried in both FF and Chrome
Have you tried edge? I have it working in edge and it is chromium
I'll give it a go. (is there any other interface I can use to manage my plan, other than through the ChatGPT interface)?
I don't think so
i need this in my life
oh my goodness did you catch that?
the discord bot detected Spanish text in the image
I got excited for a split second thinking it was spanish. Thats Portuguese
yea, the bots here are super smart š
How weird, Tried on edge, and also Safari... and still nothing happens when I try to view MY PLAN. All other functions are fine
Then all I can say is try again later
Having the same issue; Have tried on Safari and Firefox but no joy.
I just had a brainwave, and wondered if i could it manage it via the app. But no, it says I cannot manage plans via the app - only web. dugh!!!
stank
š I tried the same thing, but got the same response
Will do. It's no biggie. I just wanted to check when my next payment was due, so I can keep an eye on it, as I have been hearing some plus members have been getting their account cancelled if they miss a payment.
Yeah, that scared me too
Today my chatgpt can't be used, It says "ChatGPT can make mistakes. Consider checking important information."Does anyone have the same problem?
Guys, I have difficulty creating good functions using function calling! It would be cool to create a custom GPT for this, what do you think!?
How do you update email when old email no longer exists? Can only have one account per phone number. TIA!
Does anyone know how to change their name in the billing details, specifically for the GUI version not API?
I created a GPT, and it's displaying my full name in the builder profile. I'm curious if I can change the name
Ooops, I missed the AMA with OpenAI's DALLĀ·E 3 Dev Team. Hope I can join the next one. 
To change your builder profile name try the workaround at https://community.openai.com/t/changing-account-builder-profile-name/485380/6.
Updating the email address associated with your account is not currenlty supported. Try the alternative at https://help.openai.com/en/articles/7242619-how-can-i-change-the-email-associated-with-my-account
Anyone find the answer if we are able to have an affiliate link
You can message me directly if you know š
Does anyone have any pointers on how I can get my GPT custom action to properly pass on teh api-key?
The native Auth form does not seem to work as many times as I have tried
Hello is it possible to make the GPT run a python script if a type of prompt triggers X answer?

This GPT is designed to trigger a specific Python script when it generates a particular answer. The key trigger word for the GPT is "chicken." Whenever the GPT uses this word in its response, it will initiate a Python script that calculates and displays prime numbers from 1 to 10. This GPT should be capable of handling various user inputs and maintaining a normal conversation, but the unique aspect is its ability to recognize when to use the trigger word and subsequently execute the specified Python script.
that actually works 
Here let me explain it better, what im trying to do is make a GPT give the user random SAT questions, but some math questions require graphs/previus images, I was wondering if there was a way I could get a python script that would get the required image, upload it to a imgur or anything like that and the GPT would hyperlink that or just link it in the question
I have uploaded a compilation of SAT practice papers and it works perfectly fine, it explains if the user gets it wrong and everything, I just need a way to get the graph/image required for some questions to the user
Would the python script thing work?
Ah cool, so its possible to get it to link an image?
I cant get to my pc rn
or I would test it
ask it, I just started askjing the AI if I could do the thing and described it and it said sure and did it
whenever i try n generate a selfie with gpt4 it just generates a picture of somebody taking a selfie how do i make it generate the actual selfie
Kk will do tmr when im on
where is it supposed to get the selfi from?
nowhere im asking it to generate a selfie and it generates a person taking a selfie instead of the actual selfie
the prompt is "selfie of pretty girl"
and it is a girl taking a selfie, not the actual selfie
say I want the image that would be generated on the screen if a pretty girl took a selfie, not the screen or the phone, just the image that would be generated
š
how can i just put that in the prompt im using the bing thing
i'm using bing image generator not bing chat
not ChatGPT Plus?
no i'm broke
yeah
is my line too long?
it's kinda long yeah i want to try to use 3/4 words for it
cause there's a text cap and you can't fit that much detail in it
how long is the waitlist for chatgpt plus?
unknown at this time
try "file system .JPG of a pretty girl after taking a selfie"
what's the file thing for
Is anyone else encountering an issue with their Public published GPTs not showing up under "My GPTs"? My first one I created is no longer there, but the GPT itself still exists because I have a current Chat thread open with it.
I'm hoping the AI knows that a selfi .jpg won't include the phone or the screen or the person holding the phone
i'll try "pretty girl after taking selfies
If you have custom instructions that is a problem atm
sorry custom actions
it only has uploaded files at the moment, I haven't written any actions for it yet.
I know that some that were generated a whilke ago has issues, most say to start again and it should be ok
nope didn't work
you did not put the .jpg bit 
Well,we know microsoft's preferences now
why does it need that
cause i said so! (to make the AI think differently about the picture)
unsafe image content detected 
so it deffo did a different image!
does gpt 4 do that as much as microsoft
yea, for NSFW images
depends what angle the phone is at
does it do it if it's a crop top or something
maybe ĀÆ_(ć)_/ĀÆ
no comment
ty for your help tho

can anyone pls help me with a guidence about how to connect chatgpt api with my discord ?
'My Plan' page doesn't load.
bing censorship needs to quit it brooo
Hey,
If I build a GPTs and upload PDF files with information, does this information consume context tokens or do they not consume context tokens?
**my server ip was banned on your cloudflare, i have api and heavy loaded project, help please
**
Context tokens are what used to be "input" the act of uplaoding does not (I don't think) cost tokens, but you will (after the 17th of this month) pay $0.20 per day per Gigabte uploaded, if your assistant uses information from your uplaoded files as context in a query then you will pay for those tokens per API call
Cloudflare believes that your server is 1) in an unsupported location or 2) behind a VPN or proxy to hide the fact it is in an unsupported location 3) on the very edge of a city that borders an unsupported location or 4) a false positive. If you belive that it is a false positive then you can contact the support bot on https://help.openai.com and go to the bottom right hand corner and leave your contact details and a description of the problem including IP's and locations and you will be contacted back, typically within a week, but you may wait longer at the moment due to the unprecedented load on the support system by new users
at the moment my server is in unsupported location because there is much cheaper servers than in my original country
but i pay 400$ month btw, i need to be unlocked
OpenAI has to abide by US technology export laws, if you are using services in an unsupported country you will no longer be able to use that server, I suggest you obtain a server in a supported location as soon as possible and move your services to it.
the only way i see is to start using a proxy of supported location
You are free to do as you wish, but attempting to bypass location restrictions may result in an account suspension or ban, I strongly suggest not doing that.
you are forcing me to do that, i was not violate any rules.
my project has 500k users daily i need api
The Terms of Service clearly state that only supoprted locations can be used. The rest is up to you.
I'll give you a try
Good morning. The 'My Plan' link does not work in the account settings. I have tried with several browsers. Does anyone know anything about this?
How can I use Chatgpt vision, unlimited?
After several task, reaching limits (Im paid user)
You've reached the current usage cap for GPT-4. You can continue with the default model now, or try again after 4:41 PM
???
Yes, sadly we are limited on how much we can use GTP-4. I believe it's 50 messages every 3 hours. Let's hope they raise it soon.
š The "My Plan" link in my profile doesn't work, and so I cannot cancel my subscription. Have tried this on different computers and browsers; "My Plan" is inaccessible. Could someone help me please?
Me as well. I'm trying to cancel my subscription.
I found it didn't work for me this morning either. Not that I want to cancel, but I wanted to check what other plans are available.
its 40/3h
it was 50/3hr, I don't think they reduced.
