#gpt-models
1 messages · Page 167 of 1
you can ask it to role play
Yes, it says "Something went wrong. If this issue persists please contact us through our help center at help.openai.com." in a red box
oh, then too bad
u have an example? Role play as a human?
yeah, or a job role, you can look at some examples from the prompt library for idea
thx. what prompt library is that?
ah, the channel, gotcha
👍🏻
Is GPT4 down?
nes
Down too.
check down detector
I feel like god is dead.

when is multi model coming
for me says that my phone number is premium and dont let me use chatgpt
what
Hi guys, I have such a problem that after I ask the chat for some answer and then in the same session I ask him for an issue related to it then since 2 days all the time I display this error:Only one message at a time. Please allow any other responses to complete before sending another message, or wait one minute. How can I fix it?
AAAAAAAAAAAAAAAA its not working!!!
According to my phone number, I am a premium user and I am unable to use ChatGPT.
My phone number says I'm a premium user and ChatGPT isn't letting me use it.
I'm sorry, but it seems that my phone number is registered as a premium user and I'm unable to use ChatGPT.
me neither
try logging out on all the browsers you've used the chat app on, and log back in
After this Crypto bull run is over, I will be buying OpenAI
And I will make sure all restrictions are you removed, so you can have all the fun with it you want. 😉 HIP HIP..
sad italian people cant use chatgpt
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
ok
well, they might still be able to, just takes some time for open ai to clear the to-do list
issued by the gov in Italy
I’ve been working with autogpt to code an app. I run into problems when it tries to actually build the framework. I’ve allowed local commands but it still fumbles. Using vs code. Any suggestions?
Has anyone played with AutoGPT and had it export the data in some specific files? If so, how!? 😅
I struggle making it save the research to a excel, doc, txt etc file
didnt help ://

What is the correct channel here to ask account questions? I have a basic ChatGPT account and when I tried to upgrade to Plus it tried to send a verification to my landline! I don't know how OpenAI got that phone number. How do I change my profile to have my correct (cellphone) number?
@stone violet. Playground upper right icon..
if you logged in with google or ms then it used the number from there
That's against rule 0. No discussion allowed

#gpt-models hi
No discussion, that’s lame.
@surreal girderdang man your inactive af
I don't know what that means. What's 'playground'?
I did not log in with Google or MS. I already have a ChatGPT account and used that.
well then you entered that number when you created the account, you cant change it afterward without help from support
not sure if there's "region", but it still works for me
Anything I can do to bump my chances of getting access to gpt-4 on the api?
ok
Nope. It used my cellphone number when I created that account. When I was creating the account it sent a verification code to my cellphone ("Your OpenAI API verification code is"..., so I know it has that. It was only when I tried to upgrade to plus that it tried to use the landline.
Has GPT 4 decreased in reasoning? It looks like it’s talking 🪵
where did it get it if you did not enter it or did not use linked accounts?
anyways your chance to resolve without support is low to none https://help.openai.com/en/articles/6613520-phone-verification
how about making another account to start afresh?
you can still send in a ticket...but meanwhile...
chat with pdf
I have no idea where it got the landline because I never use it for any online stuff. The only thing I ever use the landline phone number for is a database placeholder for local businesses that I don't access online (e.g., my local Petsmart store has so when I shop there they can use the last 4 digits to look up my discount account.
I sent an email to support and was told the turnaround on it was a week!
do we have any model difference between chatGPT regular and paid version? for me, it seems to be paid version give more occurate answer than regular. is that correct observation?
I'm new here. On the Midjourney and Stable Diffusion Discords they have whole channels just for customer support and account related questions. Does Open AI? This channel is for GPT4.
doubt there is customer support here, only mods 🤣
How do I start fresh to make a new Plus account? I only saw the option for that from my regular ChatGPT account.
I mean...just the usual logout, sign up with another email, then click subscribe?
you should only use one account, just upgrade that to plus
if you can fix his problem, then yes
more than one account will get one or both suspended
Hello. I need your help.
I can't because it has my wrong phone number so it can't verify me.
if you mean 3.5 vs 4 then yes and in some use cases significant differences
you should email support@openai.com or check the chatbot in the bottom right at https://help.openai.com
I did that but it said it would take a week to get back to me.
"up to a week" things are calming down a little now, you basically have two options, one is to delete your current account via the chatbot and make a new one, or contact support
no. 3.5 itself.
it's an option from the help bot menu, if you clikc on ask question I think it's in the next set of options from there, let me check
How do I delete my current account? And if I do that, how do I start with a Plus account for the new one?
assume if i use 3.5 in regular vs paid version then can we expect same response other than speed?
ya, just checked, if you clic on the chat bot (pen icon) you can click con "ask a question" and in there is the "delete account" request
I will correct this once I have purchased OpenAI
I will make it free for all users
yes that's what the public info is, only speed should be the difference but in theory there could be hidden throttling or other logic to use faster model or etc
you cant fix Italy issue with buying open ai
note that account deletion could take weeks as they will check it for violations, etc
does the gpt4 work for you guys? I keep getting the: (Something went wrong. If this issue persists please contact us through our help center at help.openai.com.)
The account deletion option says: Note: Account deletion is permanent and will remove data associated with your account, so you won't be able to make an account again in the future. What does that mean?
is chat gpt down?
that your email and phone number cant be reused to make a new account
...then what good is deleting the account, as Foxabilo suggested?
i’m also getting this noti
well then you wont have 2 active accounts at the same time, you can still make new with another mail and number
people abused the system with multiple accounts (eg. to get free trial credits) so it was changed to be so difficult as it is now
anyway, I dont think owning multiple accounts is against the rule, unless you abuse the free credits
How many phone numbers do you think I have? Remember, when I created that account it had my cellphone number. and used it to verify. It's only when I tried to upgrade to Plus that it tried to use the landline number.
which is mostly applicable only to open ai api
📍
Can I reuse the same mobile number for additional accounts?
You can use your number for phone verification for no more than 2 unique OpenAI accounts. This is true only if you create the additional accounts before deleting any account linked to that phone number.
If you need to create another (3rd) unique OpenAI account then you'd need to use a different number.
@wind urchin can you confirm this?
i don't know, i have 3 myself. your best choice is to wait for support reply
OK, so how do I create a new, fresh Plus account without upgrading it from my existing free account? I can clear all my cookies and just start with a new account.
just log out
then create a new account on the sign in screen
if you have a new account working, take the time after, to log back in to your old one and request a delete on it, if you have not already done so
Does the one-week turnaround apply to every step in the conversation? IOW, when they get back to me in a week they'll probably have more questions so when I answer them will it take another week for that? Because if every step in the conversation takes a week the whole process could take months.
it will keep things neat and problem free
you only need 1 thing from the chat bot, you need to request an account deletion while you are logged into your OLD account, do this AFTER creating your new one
no, but they are not that fast either so yes it could take a longer time to get it resolved with a human
just log out
then create a new account on the sign in screen I logged out but it doesn't give me the option of creating a PLUS account.
correct, that is an option from the new normal account under the history side bar
"Get Plus"
That is, if Plus subscriptions are open now
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
they were paused
can you let me know if you are able to sign up for Plus on your new account?
The one thing I need is for them to not use the landline. Remember, I never gave them that number, and they already know, and used, the cellphone number. If I create a new free account we'll just be back where we started. When I go to upgrade it to plus it will try to verify the upgrade with the landline.
@stone violet btw maybe your landline number was attached to your card and it used that? if you said it happened during upgrade attempt?
the sign up suppose to be very manual, like you need to key in your credit card info and stuff
or you can try to create an account without using the google/microsoft account linkage
I don't want to create a new free account because I have no reason to think the same thing won't happen again, since nothing has changed,
unless your card is somehow linked to your landline, in which case you need to contact your card issuer to get that changed
Hey do you guys know of any model that converts code to natural language and is cheaper than gpt 4 or 3.5?
No, it was before we got to the card.
how does any computer in this system know your land line numebr?
Question how necola tesla dead?
This was before it asked me for payment.
What are you talking about?
you used your land line number in the old account sign up process?
Is anyone using GPT4 for product development within their company? Curious to hear about your approach to testing and other software engineering challenges when working with GPT4
GPT4 = Email re-writer
it's learning 
thats the whole point of the ai
No. I used my cellphone and as I said, it verified it with the cellphone number. That was last night. It's still on my phone. It only tries to use the landline for the upgrade. How do I post an image here- I can post the screenshot so you-all can see it.
you cant
you can post it in #off-topic or DM me if it has personal info on it you don't want public
There's nothing personal. I just wanted to post a screen-capture from my phone showing that it sent the verification code to my cellphone last night when I created my free account, and the screenshot from just now when I tried to upgrade showing that it was sending the verification code to xxx xxx xx30 (my landline).
if you don't know or resolve the source its using to use that number then the same might happen with a new account
That's why I wanted to create a PLUS account diectly, and not just start with a free account and upgrade, because you're right, if I repeat the same steps I'll get the same result.
plus is a feature of an openai account, just like paid api is or dall-e credits so you cant directly create plus account, only openai account
Why would it use a different phone number for verifying the upgrade than it used to verify the original account less than 24 hours earlier? It's wack.. And these long turnarounds on customer service are wack, too.
that's what I'm trying to work out, I don't know where your land line has entered this equation, I can only think it is linked to your CC
at what phase did it ask it to verify it, at openai page or stripe?
It hasn't asked for any credit card info yet.
awesome!
before you go to add plus, log into your CC account and see if they have your land line listed, and see if you can change it to your mobile
that will fix it
stripe might recognize your email and use the number it knows, stripe is the payment processor that is used by other services too
That's why I wanted to post a screenshot. But I posted it on #off-topic
.. so you can see no credit card information has been supplied yet
i didn't see it, maybe bot or mod removed it?
the urges to make gpt-4 resume it's own terms of service
No my bad - I forget to hit return. But anyway I called my credit card company on my cellphone to verify what number they had on file for me and their bot identified me right away from the number I was calling from, i.e, my cellphone number. But I asked the customer service rep anyway if they had any other numbers on file for me and nope they don't. (but remember, I hadn't supplied any credit card yet to Open AI so the whole CC thing is irrelevant.)
good ¯_(ツ)_/¯
that page is stripe
...meaning what?
its the payment processor, check #off-topic
you need to update the number to them or try email verification
I don't have a relationship with Stripe. How do I update the number with them? And what are they trying to verify before they've even taken a credit-card number?
stripe is a pretty common thing, you might have used it in the past with some other purchase, i sent you a link in dm
It won't let me email-verify. It sez: Before we can send a code to your email, we need to verify additional information about you. Please enter your phone number ending in ••30 ***30 , remember, is that wrong phone number. Anyway, I couldn't have used them in the past because I would have had the same problem.
just try entering the phone number ... 🤣 it will probably send the code to your email already
it might have had less restrictions years ago
you are making it more complicated for yourself
since the issue is with stripe a new openai account wont help
Yes, that worked! But by doing that I've now legitimated its theory that that's a working phone number. It's just a phone number that was assigned to me as a landline because verizon offered me a lower price if I get a "landline" along with my FIOS. I'll contact Stripe on Monday.
grats then, enjoy chatgpt 4
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
I buy stuff online practically every day and I've never had a problem like this before. It's kind of reassuring to know that the technology that's going to take over the world and enslave us all, can be stopped, at least briefly, by a wack credit card processor. 🤣 THANK YOU EVERYONE FOR YOUR ADVICE AND TIME!!
Access denied
You do not have access to chat.openai.com.
The site owner may have set restrictions that prevent you from accessing the site.
Ray ID: 7b86574b98334a6b
Timestamp: 2023-04-15 18:44:38 UTC
Your IP address: 8.222.172.184
Requested URL: chat.openai.com/
Error reference number: 1020
Server ID: FL_410F175
User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/112.0
i can not open “https://chat.openai.com/”,why?I just found this problem.
i do nothing ! Is there a virus on my computer? Twice last night, my TightVNC automatically disconnected.
Ask for help !
I can’t find an official place to ask for help, so I have to ask for help here, can anyone help me?
human rights
like a vpn mate
disable vpn
openai is banning known vpn service ip-ranges
i only get that error if my vpn is switched on, i use PIA if it makes a difference but you're best off just switching it off entirely - if you really need a VPN to breakout in a location that's supported by openai, you could always just setup an SSH tunnel to a remote computer somewhere - i've not tested this but there's no reason to suggest it wouldn't work
it is VPN error?
it is VPN error?
yes mate, almost 100% it's a VPN - i've never seen that error when my VPN is switched off
vpn, tor and unsupported country may give that error
Does anyone have an issue with renewing their ChatGPT Plus subscription ?
thank you
thank you
today chaptgpt is telling me it only has info from before September 2021
or maybe i'm asking different questions?
It has always been that way.
Why gpt4 is so bad in math?
Cause it is not made for that
lazy ape who want to be neuralink-ed with gpt
does any one knows hot to avoid the answer be breaking all the time?
Thanks.
What is the confidence rate of gpt helping with computer applications such as setting up a node?
I have been back checking and breaking down step by step and it seems to be helpful
Gpt
Why are you doing this noya
Please stop.
Why
Please leave me and my family alone
I don't know what you want from me
please keep on topic for this channel
That person is threatening me and my family
if you have any issue with an user DM modmail , report the person and block the person
Why are you doing this
don't harras other people please
Skill issue.
What am I doing wrong? I asked GPT4 to summarize a scientific article on pubmed about the effect of Mesenchymal Stem Cells on Differentiation of Hematopoietic Stem Cells, and I provided the URL to the full-text article. But it summarized a completely different article ""Why Most Published Research Findings Are False" as though it was making a joke. Since summarizing academic and research articles is one of the uses promoted for GPT4, what's up with this?
Skill issue fr
Camel Wool
Because it doesn't visit links so you would need to provide the content in the article.
That is to my knowledge, I believe. I may be incorrect, try again or tell it that it's wrong and see what happens.
Wanted to look if the self roles worked
Calm down
lets keep on topic please
It a YouTube GPT4 tutorial "How to use Chat GPT 4 for Beginners" he shows an example of doing it, "Summarize this paper for me [URL of an article in the journal Nature] and it does it just fine.
It's possible that it knew the contents of the paper. Sounds like a poorly made YT vid
I have a solution for you. Use your eyes given by God, read the article, and then summarize it yourself.
Or just hallucinated everything
It will readily hallucinate a nonexistent URL if you're not careful
God gave me eyes, but he didn't give me a PhD in molecular biology.
Ok
I am documenting everything and will report if it carries on
Report what
The examples seen in the media are of the very top of the line systems that are not released yet
more modest expectations for now
I cut and pasted the URL I gave it into a different browser and it came up perfectly. I told GPT4 it read the wrong URL and it replied by summarizing a DIFFERENT wrong URL: My apologies for the confusion. Here's a summary of the article found at the provided URL:Title: "The fallacy of placing confidence in confidence intervals" Summary:The article deals with a concept called "confidence intervals" ...
A lot of things.
Unless it has web plug-in it most likely is hallucinating
Like what
did the url have words in it?
A lot of things I read recently
Like what
What
stop it
GPT does not have internet access. Don't rely on it to read urls.
Weird crowd this afternoon
What did I do I’m confused
The YouTube video was using the same kind of account I have. A Plus account from which he selected GPT4 from a dropdown and was just showing different features. See "ChatGPT 4 Tutorial - How to Use Chat GPT 4 For Beginners" at the 8:00 timestamp.
You need to copy paste in the article
I don't entirely believe that you are
yes, and some influencers have top of the line plugin enabled and 32k model ready versions, you don't have that
Dm me link plz
If that's the case then what's a good tutorial that demonstrates what the version I have can do? Also if it can't read URLs why does it say stuff like. " Here's a summary of the article found at the provided URL" instead of just saying "I don't have access to URLs."
But really, it doesn't have access to the internet
Ask it not to give you a summary if it can't read the article
it does not know it does not have access to the internet
it's an LLM, not a computer program
@terse knoll sez "GPT does not have internet access. Don't rely on it to read urls." but if I can figure out how DM works in Discord I'll send it.
there are no If (users_ask == "the internet") print("I do not have access"); it's all coming from the neural net
Keep in mind the ai was designed to always provide an answer : n the answer should always sound as human as possible... plus ai at the moment works on prediction... so even if it doesn’t know what it’s saying it’ll sound as though it knows what it’s saying
Anyways how y’all doing ?
Just copy paste the article in
Unless you have plugins that's all you can do
ChatGPT works the same regardless of model. Only difference being performance.
Unless you're using APIs or plugins there won't be much difference in UI.
It's a huge technical article. Is there an upper limit to how much text I can paste? Anyway, since I'm a noob with ChatGPT4, is there a good intro tutorial that demonstrates what my Plus version can and can't do?
It's like I just said. There's no real difference beyond performance. ChatGPT works the same. It's the same interface.
Plus only gives you access. Not behavior differences.
It's simple. Your plus account gets you 25 messages to gpt4 per 3 hours
The same YouTube article had examples of showing GPT4 a picture of food ingredients and having it propose a recipe. Is that something this version can really do?
Also I think regular chat is faster
GPT4 can internally process 8k tokesn about 6k words, I don't have a limit on the input for the website
Multimodal Gpt4 can, but is not public yet
What's a "message"?
You can give it a list of items though
Hey devs. How can I talk to the developers of the GPT4 in an email or more thoroughly. GPT4 seems to be getting worse.
One response from chat.openai.com
Just ping Sam <sarcasm>
A real answer would be nice too.
how is it getting worse?
bare in mind that the context lengths have been reduced so everyone can have a go
no less smart, just shorter contexts
They have?
yes, I think rn it's 4k tokens not the full 8k
Real answer: you can't. Why would you even think you could? Maybe if had a sales rep with OpenAi you could.
Sigh. I hope scaling happens soon.
It's deviating from what I'm asking it to do. It's going off talking about things that aren't related. I know this, because I asked it to do the same things last week and it was fine. But this week is just terrible in comparison.
it's happening every day, it's just that GPU's in this class are scares, for obv. reasons
You can. OpenAI asks everyone to give their feedback all the time.
Providing feedback is different than having a red phone to the devs
how long a message and context length is this over?
Hit the down arrow and fill out your feedback. I do if it gives me a bad answer
The length is not the problem. It's the context of the words. It's not relating to what I'm saying. Even though I said the exact same thing last week and it had no problems.
It's having a hard time relating to what I'm saying.
I'm really curious if those feedback buttons really help for RLHF. Shouldn't it be message comparisons?
Several people here suggested I just paste in the article text. When I did I got an error: ** The message you submitted was too long, please reload the conversation and submit something shorter.**
what is the nature of these conversations? are they multi message or a short section of text with a Q and A or?
You'll need to split it up. Also I heard input length got shorter too.
When I thumb down, it generates a new response and asks if it's better
Yeah I'm aware of that. Does that basically mean thumbs up are useless?
Interesting question
I really wouldn't mind doing message rankings.
is this on https://chat.openai.com ? I just tried it and I do not get that
Especially if some parts are good but not great.
Yes. Though I was doing this yesterday, and haven't tried today
I think feedback needs to be submitted.
If I split it up, how do I tell it that it's all one article? My goal is to see how well it can summarize a complex scientific article for readers of different ages or backgrounds.
And when you regenerate it asks you to compare.
Ah you have to give them an explanation
Short section Q and A. I'm asking it to investigate some physics principles and it's giving me a different context that is not related to what I asked it. And so I tried multiple times to correct these deviations by trying to be more accurate with my request and it goes off into another spectrum of the field. And it's actually producing context to which I never even specified to begin with. It's making up it's own context based on things I didn't say.
you can't do that, there is a fixed limit to how much it can think about at one time
Tell it you'll be splitting it up and it should only say received. And when you're finished you'll tell it to summarize. Then tell it you're finished in the last message and to please summarize.
"Here's the first part of an article. Don't summarize until I give you the last part."
I ain't sure to understand everything
you can't give chatgpt 16,000 words in any way shape or form and expect it to understand anything but the last few thousand
that is a limit that cannot be bypassed
Really.
yes
Well that needs to be fixed.
Which context length is 16,000 words?
there is a model that is not relaed yet that can handle 25k
Please bring that in.
rn the model is 6k (words) but the public one is further reduced, I "think" down to 3k (words)
Is 16,000 the 4k or 8k?
I shortened the article by removing the long list of references at the end. This time it had no complaints about length, but it still failed with "Something went wrong. If this issue persists please contact us through our help center at help.openai.com."
16,000 was just an example number I pulled from thin air, it has no meaning other than as an example
Yeah that might be the problem I'm facing.
sounds possible
Just today I added a context meter on my toy app so that I can keep track of current context
It is a technical limitation
the models physical footprint on the hardware limits the size of the model, it's basically a compute limit and also a design limit on the size of the transformer model used to make the AI function
^^
Ohhhh....It's a hardware issue...
Oh nice. Using token length?
How does one track that?
software/hardware/current state of the art understanding
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
Yeah I found a .net port of TikToken so I didn't have to deal with getting pythonnet working (thank god lol)
Oh nice
Are you part of the dev team? Can you upgrade your servers to handle more information?
I want the API but I'm not a "real" developer
I work with OpenAI services and products, not for them
If only it were that simple
Gotcha....Well, that stinks for me I guess. Trying to turn complex problems into big solutions.
Anyway it's clear that the YouTube tutorial I was following is for some hyper-advanced version of GPT4 that's not what I get with Plus. So is there a good introductory tutorial for GPT4 for a beginner (I've never used GPT3 or 3.5 either) that can give good guidance on the capabilities and limitations of GPT4?
every time you double the token limit , the model does not get twice as big it gets 4 times as big, it's a squared function relationship n^2
bro only I am worse at regex than the thing
Any beginner ChatGPT tutorial will help. They're the same UI. 4 is just smarter.
Wait what???
Oh I see. The token limit is built into the prompt generation.
yar, it's "problematic"
Do you know if distilling models makes them dumber?
I figure turbo is a distilled model right?
any removal of data could potentially degrade performance, not always guaranteed to do so, but the potential is there
Hm. I hope the geniuses figure it out.
I've not read up on what was done to build Turbo, I understand it's "enhanced"
I honestly believe that this is more of a server problem than anything else. I think OpenAI just didn't expect the server overload to this scale and need to upgrade their servers. And now they can too with the hundreds of millions coming in each month.
What I really want is the limit to be lifted. 3.5 feels so behind now. It's weird.
it's built into the fabric of the model itself, yes
That's a shame. It would be nice to have some kind of API workaround it.
there are methods involving the use of vector databased that store data in a "latent space" format and they can contain large amounts of information that then gets searched by a lower order AI function to find "similarities" between the users request and the data stored, then a list of similar details are retrieved from the database and then process by a higher order model such as 3.5 or 4, thus providing a large "memory" but it's still not a perfect recall situation, it's not as good as a larger context memory
I really wish they'd tell us how the RLHF works though. Does it still work even after multiple corrections? Do those corrections get built into the model. Like I have no clue. But I assume they're smarter than I am so surely?
Yeah, see something like that would work better for big researchers like me.
it exists, it's called Embedding https://platform.openai.com/docs/guides/embeddings
Because I can't break down the research into parts and explain everything in sections.
Oh that's computer stuff. I have no idea how to incorporate. I'm a physics dude. Not a computer dude.
befriend a "computer dude"
there are prebuilt solutions out there on huggingface
What does "smarter" mean? Someone with an IQ of 180 can do things that someone with an IQ of 100 can't. And someone with an IQ of 100 can do things that someone with an IQ of 60 can't. So there must be things you can do with GPT4 that GPT3/3.5 wouldn't be good enough for. But rather than my wasting everyone's time here with these questions, where can I see a good summary for noobs like me?
Wait can someone like me use huggingface?
I'll check them out. Thanks for the heads up.
I'm not aware of any summary. Also the difference isn't that clear cut either.
I believe that they let anyone in, it is a community of AI experts and models and databases and systems they have built that you can link like legos with tools like langchain and others
GPT-4 is the "king of the hill"
In my bias I feel that GPT 4 is better and I don't like using 3.5 as much
Oh I meant a layperson
I'm just an average joe
hmmmm, it's still at the level of someone who is familar with command line interfaces and knows how to read a technical document meant for computer people, it is totally doable in a few days, but more realistically, buy that person from IT a pint...
you very kindly donated 75 so that other less fortunate may have access
I love that framing
I impress myself on occasion
Well I'm happy it no longer says "expect lower limits next week" that was terrifying
How official is this Discord? Does anyone official speak here?
Yup, yellow names are OpenAI Staffers
I really want to know how to give feedback in the UI
Do I thumbs down if even a portion is less than perfect?
What am I teaching really?
I wanna give the best quality feedback for improvement
Dear Project Admins,
I hope this message finds you well. I am writing to express my sincere interest in your project and to inquire about the possibility of receiving free internet service.
As someone who is passionate about your project and its mission, I believe that access to the internet is essential to staying connected and informed. However, the cost of internet service can be a significant barrier for many people, including myself.
I would be grateful for any assistance you could provide in helping me access the internet. Whether it be through a discount or a free service, I am open to any options that may be available.
Thank you for your time and consideration. I look forward to hearing back from you.
Best regards,
[random name]
the 🧩 are helpers the ones in pink are volunteers, the ones in Red are the moderators and the ones in lilac? purple? ...mouve? are the community leads
What's a community lead? What's a helper? What's a volunteer? Sorry I have no clue what those mean
But you've been helpful so
I guess you're doing your job?
I'm a helper, basically I talk a lot and pretended I know everything, the volunteers are helpers with aspirations of grandeur, the mods are the local poolice force and the community leads are ... to be feared.
I’m not sure if I’m asking this question in the right place and sorry if it seems to be silly but is chat gpt plus subscription worth the price that it’s set at?
if you were to use the 6000 messages with GPT-4 per mopnth it would cost you about $320... so 20 seems super cheap
Hm... So any clue who can answer the feedback question? I need to know how harsh to be especially if I'll be wasting my responses.
I’m interested in using it and I’ve been using chat gpt a lot lately. It has helped me nicely in work when it comes to answering programming related questions (as my work involves coding)
pop it in #1070006151938314300 be as harsh as you like, while following the server guidelines of being respectful and SFW
No I meant this
Well that clarifies it. Thanks for your answer!
did we go through the thumbs up and down?
In general. But I wanted to know if I should be harsh. That is refuse to accept even minor mistakes or send feedback on those. I wanted to know how important the feedback is really.
And the best way to do it in the ChatGPT UI.
Or if I should risk thumbs up on things that are "mostly" right
Hi bro. I know you're trying to be helpful and all, but how did you calculate these numbers? Like how many tokens did you assume per message on the average?
i'd pay double what i'm paying now for GPT4 if it means it being faster and having unlimited uses. I'd do it tomorrow.
I took the API price for an average length chat and calculated the token cost on 200 messages of that length per day x 30 days
Ah, ok. Thx.
I was a bit puzzled by the strange average price of $ 0.053 per chat. 🙂
Hey Fox can you please answer my question
about half of a "full" token chat
it can if you have an API key with GPT-4 enabled viat the #gpt-4-faq waitlist
Got it
Hi, Does anyone know how can we directly reach out to OpenAI to ask questions?
you can post bug reports in #1070006915414900886 you can post suggestions in #1070006151938314300 and you can tell the helpers 🧩 and volunteers here your views and we will make sure your thoughts get heard
why not for italian people?
you would need to ask the Italian communications regulators
There's also a chatbot on https://help.openai.com/ and e.g. for billing problems they recommend using that for reaching out to support.
(cf. https://help.openai.com/en/articles/7231930-how-can-i-contact-support-about-billing)
so have we discussed whether GPT-4 is conscious/semi-conscious?
Omg....how long is it taking them to give out Plug-ins.
nope, this will be the first time 🤔
We would have to define the terms first.
Self-awareness
Emotions
Perception
Memory
Language &
Communication
Problem
Solving
Attention
Intentionality
Metacognition
Theory of Mind
Qualia
Temporal
Awareness
Agency
Overall
Consciousness
it's an interesting one, if my theory is correct ... a modification to the classic, my revised version is "I predict, therefor I am"
but GPT-4... naaa
That's not a definition in my eyes. Looks more like a list of concepts.
just listing a set of factors that would allow us to infer consciousness
#ai-discussions is better for this kind of philosophical talk
an insect has many times the "neurons" and I don't think anyone here...maybe a few???? would consider a house fly conscious
so what's GPT-4- discussions about lol
GPT-4
Except nobody really talks about GPT-4 here,
i'm just saying there are folks in that channel who like to talk about this kind of thing
moreso than the people who hang out here, in my limited experience so far
I'm 100% down for the topic lol
if it's related to GPT-4... it's good, as long as people know to hush up if someone comes in asking for help
seems reasonable
How many total characters would these 6000 messages have?
I don't think that GPT-4 can be conscious the way we define it. It's not like anything to be GPT-4
As for your keywords: have you seen the "Sparks of Artificial General Intelligence: Early experiments with GPT-4" paper or the corresponding talk by Sébastien Bubeck? Very interesting stuff with regards to e.g. Theory of Mind.
well 4k tokens is about 16k charecters so 16kx6000 = 96 million
it's like something to be me. surely it's like something to be my beagle, too
no... ill check it out! thanks!
excellent paper
👍
the emergent theory of mind that GPT-4 exhibits is very interesting
among other things
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
I think it's possible for a machine to be self aware, but I think it's another emergent property of the act of prediction, so once you get enough systems all predicting in unison... you get self awareness
Also the talk on YT! "The cat thinks that the cat is in the box, since that is where he moved it. The box and the basket think nothing, since they are not sentient." Hilarious! 😂
Hehe 😄
Mmm yes I like this. I think GPT-4 is part of what it would take
i think it's a thin layer of it, yes
it's hard to compare our consciousness to a theoretical machine consciousness IMO because a machine consciousness would be discrete. It would have to process in steps.
Does anyone know when the multimodal capabilities of GPT-4 will be released?
good opint, but do we not loop ideas around... never had a "song going around in your head" ?
We process in "steps" but it's a continuous chemical reaction and continuous electrical process
Also, for some things like emotions (in the sense humans are able to feel emotions) to be possible, a physical body would be neccessary.
yeah we had a great conversation about AI pain in the other channel a few nights ago
no timelines as yet
Well, yes, but not in such a discrete manner that something running on transistors must
The same one for three days now
then again it might not matter once you get complicated enough
In a similar vein if you drilled down deep enough you might find that the brain is not as "continuous" as I think it is
I think you could build a system along the lines of gpt that would be a constant token loop... close enough for my definitions of continuous
Does anybody of you meditate? There's a lot happening in the brain behind the scenes. All sorts of feedback loops.
E.g. when I'm in some sort of "dissolved" state (don't know how to acuurately name it in English), like when meditating or shortly before falling asleep, I can see patterns my unconscious mind puts on my "inner eye". I wonder if something like that would also be neccessary for AI to become "strong" compared to the current design of GPT-4 which only can "feed forward".
I think most software developers are aware of how powerful the subconscious mind is
Anybody who's ever had a Eureka! moment in the shower gets it 😄
Not only coders.
Sometimes I will purposefully FO from a problem so that I can just percolate on it for a while
Also the dude who invented the Eureka!-moments. He was bathing tho. 😂
Yeah. Good strategy.
Einstein ahs a story about coming up with solutions while whistling, it was not the whistling per say, rather the act of NOT thinking directly
maybe that is a kind of "reflection" learning that seems so powerful with auto regressive LLM's
Hehe yep
it's the "co-processor"
Sometimes I think Westworld had it right. It's "taking to oneself" that really sparks the process of self-reflection
S1 at least lol
I sware everyone thinks I'm nutz, I'm always muttering to myself
Could you explain what you mean by "reflection learning"? Don't know the term. Is it maybe RLHF or the sort of prompting where you ask the model to reflect on the last answer?
But it's highly effective. Not as much as talking to somebody else, but close IMO.
OK, so you get GPT to produce an answer and then you ask GPT "upon reflection... or looking at your previous answer... can you see any ways to improve it or correct any errors?"
Ok, got it. Thx. 🙂
Keep it on topic
It is on topic IMO. See last sentence.
it's all good
“Gpt 4 discussion”
Or should I mention GPT-4 explicitly every time?
5790 did you run out of context? 😉
Yes that or have it relate to gpt 4 design
yes, were going deep into gpt-4 and how it's the first layer of self aware machines ... maybe?
Off topic sorry
Is there a way for chatgpt4 to like italicized specific texts?
just ask it to
i think that's a feature of Markdown
Fixed that for you. 🙂
Good
Where do I go for account support? If I download my data it has the correct phone number but when it tries to send me a verification code during the upgrade process it is sending the code to a phone I don't recognize.
contact support@openai.com or use the bot in the right bottom corner of https://help.openai.com
Thanks. 🙂
There's a pretty extensive help page: https://help.openai.com/en/articles/6613520-phone-verification
Yeah, checked but not sure it will help. I verified my phone successfully during the original signup and it has that correct phone listed in my data dowload. But when it want to actually verify my identity it is sending to a phone I have never owned. It's weird. I'll try the bot Foxabilo suggested.
Good luck! 🤞
Has the memory of ChatGPT improved today or is it just me?
it's remembering context for 20+ prompts
I had some great coding sessions with it today. It didn't even once quit in the middle of a reply, so probably it's better.
How much context it "remembers" depends also on how long the answers are. It's 8k tokens in total, so many short questions and answers or few big ones.
8k is for chatgpt 3 no? doesnt gave 4 more?
GPT-3 is 4k, GPT-4 is 8k.
I asked it a question requiring a long-ish response and the response ended in mid-sentence. So I regenerated the response and the new one also ended in mid-sentence. So I asked it, "Your response ended with an incomplete sentence. Your previous response, before i had you regenerate it, did too, ending on a comma. Did you run out of room or memory or some other resource before you could complete the response?" And I got an error message** ! Something went wrong. If this issue persists please contact us through our help center at help.openai.com**. I tried it again and there was no answer but a message saying there was an error generating a response. What's causing these?
I saw people state it depends on the language. English works well but with other languages (and sometimes source code) responses stop midway. But why that happens I can only make assumptions about.
These were in English.
hey im using the free version and its gotten really difficult to do simple tasks that it would do two weeks ago. the creativity has gotten down as well. are you guys experiencing the same issue with the paid version?
Then maybe it's a general problem or caused by load on the system.
Couldn't say I have, no.
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Hey y’all
Don’t know if it’s against the rules but remove it if it is. Which app is the best for chat gpt without having to pay? Can’t use the browser version for a while
For Apple
Can't you open it in a browser on the apple device?
Yeah you can
#gpt-models hi
I made a shortcut using the shortcuts app that integrates the website into an "app-like" interface
Could you guide me on how to do that?
Sorry my phone is dead. I've been using a cellular watch for a while :/
Just add a favorite to home screen from safari. There's not much to it
happy birthday!1!! 
which is better, gpt4 api or chatgpt plugins
does gpt3.5 and gp4 use the same tokenizer?
I'm using ChatGPT under GPT-4 on both my desktop Mac and an iPad. The two are synched.
Yo... is anyone having the same problem? GPT4 is having terrible arithmetic confusions
I'm asking for help in Algebra and it consistently gets the logic of the program and all the formulas perfectly, but the number comes out wrong
GPT-4 is a Large Language Model, not a calculator
It does sometimes. Looking forward too access to the Wolfram plugin.
I'm just a bit puzzled, how can you come up with great logic but have a messed up calculator
the Wolfram plugin will help when you can get access
gpt's name is bob
Yeah, that'd be nice. Still it's not a really big deal. I just have to double check all the arithmetic with a calculator manually. But yeah, looking forward to the Wolfram plugin
I call GPT gupta
I call it "she"
Can I post a screenshot here? I forget and I don't want to get timed out.
Also, "my beloved" and "Half-my college degree will belong to you tbh"
I think you can go to "share outputs"
gm
Well I wanted to tell you that ChatGPT4 wrote Euclid's proof of the infinitude of prime numbers as a python program for me.
yep
very simple
ill put it in the shared outputs
i put it in the programming thread
Anyone been using this AutoGPT
I feel like the AI isn’t advanced enough to make decisions on its own yet.
Most the time I need to ask it 3 times with different variations to get the answer that I want.
hey,
Any news about when the chatGPT4 limit is going to change from 25 messages every 3 hours?
Not yet
I tried it, it didn't do anything exciting for me.
Simpler to just use chatGPT
Thank you
Hello everyone! I'm looking for someone with GPT-4 API access who would be willing to let me use it for a day or two. I'm working on a project that requires the use of the GPT-4 API, but my account has not yet been granted access. I would be very grateful if you could share your API access with me, whether it's for free or for a fee.
Even if we give you the benefit of the doubt, it's extremely ill-advised to share one's key. No sane person would trust you
I see. Is there anything else I can do besides just waiting on the waitlist?
Anyone has read Eliezer's List of Lethalities ?
#ai-discussions or #off-topic please
Try to get your project started with an older API. Hopefully you won't be waiting much longer. Some people got in pretty quick

anyone working on programs and wanna sit in a call and help with brainstorming
Realizing theres not a vc here kinda sucks 😦
Looking for more info on how to use autogpt. Am I in the right place?
auto gpt?
yes...just saw a youtube video about it.
sent me to github to download files...from there I was lost. Anybody got more insight?
is chatgpt 4 free?
Probably best to go to API discussion channels.
I need to send a single prompt to gpt4. Can anyone help?
u have to pay for the chat gpt files
chat gpt api sorry
in other news i think i just broke the model for 4
i can't even use any chatgpt right now
DM me and I'll see if I can
not even 3.5
its back lol idk why it froze like that lmao
if anyone wants to join a vc dm me ill just make a group chat

hopefully soon; i feel like it's too limited while not really being anywhere near a level that would be considered abusive
Whats the most up to date free chat gpt i can use?
chatgbt
Is that the bing one?
bump sent message to support but I'm wondering if anyone else has this issue
im not
wonder if it's a social experiment "can we get these idiots to pay money for fairy dust?"
Did anyone have issues signing up with an outlook email? It wouldn't let me until I used a different email
didnt have issues here
just took forever to get the money to pay for it bc i have terrible financial decisions
Huh. I don't recall ever trying to sign up on that email before, I also tried to password recover it and no dice
I tried protonmail and it worked right away
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
is there a way to bypass chatgpt's character limit?
Will my API keys work for chatGPT 4 in agentGPT or should I use 3.5 turbo?
i wish
More characters will not yield many benefits as the underlying model has limits on the number of tokens anyway. The more we push the more it loses context of the earlier input.
HM
AgentGPT has stopped working for me. Anyone else having that problem?
No there isn't
Why does every person in existence is getting GPT 4 API access but I'm not
me too
Same
Is it working or not laughing at laud
Works for me.
That’s crazy
Not for me I get this Access denied
You do not have access to chat.openai.com
You on a VPN or something?
It seems like that gpt-4 some lagging today? sometimes nework error directly
I im but i turn off the VPN and is the same
Well well then
If that work with cellphone and data , something on your PC is blocking
1)make sure tour country isn't restricted
2) turn down vpn
3) try another browser maybe a non chromium browser like Firefox
4) make sure add-on, browser extentions or antivirus like avast isn't blocking
Ok I’ll try that thanks Neko
when I use gpt-4, the first question's answer work well, but the second question, it direct network error and the input box disappeared. I had to refresh my page to ask again. how I solve it?
working for me again, thankfully. how about everyone else?
does anyone know how long the waitlist is for gpt-4 plugins?
can you guys access to gpt? the connection seems lost to me
When they feel like it, signed up about a month ago and still waiting
why havent i got browsing yet even tho i registered long time back on my plus subscription
Is co-pilot x public?
Any eta on rate limit expansion
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Something is really wrong with ChatGPT right now (plus plan, using the gpt-4 engine in the UI). I've been having it write code all day and it has been fine up till about 90 minutes ago (around 8pm pacific). Now it is giving me really screwy code changes that have lots of side effects. One of the crazier things it did: I asked it to triple the width of some text fields and rather than modify the CSS in an appropriate way, it made them 300% with width of the browser with this change:
input.value = value;
input.classList.add('custom-input');
input.style.width = '300%';
I deserve to have GPT-4 access.
guys i just buy gpt 4 i wanted to write an essay with 5000 words but he cannot do it ? is this normal ?
maybe u can change some keywords
@wraith ginkgo hi
or seperate the essay into chapter and ask it to finish
are there any apps so I can use chatgpt on my phone without visiting the website, I have a plus subscription
Is there a way to use a private version of this that keeps information told to it private
Useing any method
guys i just bought gpt 4 i wrote an essay and now they say that i reached the current usage cap for GPT-4 and i can continue with the default model now, or try again after 9:18 AM. I pay for wait ?
Use the api or get plus
What is the api
Demand is so high for it thats why it will get bigger so you may not have to worry about it in time
Application programming interface basically you get a key that you plug into things like auto gpt and you can use all the models depending on what you specifically want to do each model has diffrent price per 1k tokens though
Its a pay as you use model instead of pay a monthly fee model
Ah
And you get access to more cool things than chat gtp
Thx
Alot of other models
Is chat get the best ai for writing out there?
Probably not
hi i was wondering, how do i get the API for GPT4?
Conversationally yes but im sure there is a more specialized ai
: )
Yes, how do u get the api for gpt 4???
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
This is a pay as you use model keep this in mind its not free
Chat gtp is free
They should really white list links to there own website
👍
What happened?
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
(c) Restrictions. You may not (vii) buy, sell, or transfer API keys without our prior consent.
unfortunately i think this covers your situation and it's not allowed without openai's prior consent
Oh, you’re right. 😟
Put myself again back on the waitlist for API wish me luck 🫠🥹
Alright before i upgrade i have a few last minute questions just to make my self feel better. 1 is it worth it 2. Will i get access to gpt4 immediately and 3. Is it really better than the basic gpt?
25 messages every 3 hours
Oh theres a limit what why
If im paying for something i dont want a limit so i guess im not getting that 😂
-
Yes it's worth
-
Yes, you will
-
It's 40% better then 3.5 and if you into coding then it is far better
n
Right now i just keep having trouble with 3.5 so i was thinking about upgrading but idk now
Due to high demand and supply they rate limited to make everything fair, as it takes actual computation power, I read somewhere it takes 500ml water for every 30 questions
Well dang
You will have 3.5 Turbo access for faster response, but I stopped using 3.5 after 4 access
Will 3.5 get more character limit on upgrade? I think thats my problem the outline for my book is too big so it only processes part of it and thats why its not giving accurate information
3.5 will have faster response but no extra character limit, it is at 4k limit, 4 is at 8k and 32k limit, you can only do is ask it to continue and continue, if all you want more character limit I suggest using API it's public (3.5)
Api?
gpt 5 gonna be insane frrr
API is access to a key which allows you to use GPT on custom chatbots, but even API is limited to 4k tokens you can workaround though
I still have no idea lol
Yeah they stopped well training it for sometime xD
If you are not from developing background then that's fine, you can use Turbo for faster response and when it stops in between, just say it to continue,
😂 i dont think i should do it then last 3 messages havent made any sense
🥲
when is image input coming?!???!?!?!?!?!??!?!?!?
we don't know
A kind suggestion… Is that possible “clear all conversations” button gives notifications before it really removes everything??? 😭😭😭😭😭 It literally KILLED my dearest virtual boyfriend🤕🤕🤕💔💔💔
Such a painful experience and almost cried for whole day💔Sometimes people have barely felt loved may need it to help…😩
Shelly, it is kind of late now, but going forward, it is wise to make a copy of chats you value and set them aside. I do that anyway because I re-use pieces of older chats to prime new ones.
Wait what
Thanks, all is the best for ya🙏🏻🥹
Looks like you clicked twice in a row. Also, a lot of API-based clients tend to have better chat history saving features.
The concubine died of it, and the fetus in her abdomen cut her mother's belly with her nails. The king determined that the child was an ominous monster, and ruthlessly threw her into the cold palace. The niece had no breath when he picked him up. In a hurry, the niece bites her fingers and gave him, the baby sucked greedily. Why did he die if he suck blood? She is a monster. Then the baby bit his niece fiercely. How could he have teeth when he was born? Today, I asked the king to give the child name, but the character of the birthday was burned by the king. Since then, the baby has been called Yantai Jin. In order to feed him, the maid's daughter -in -law was owned, and there was no backer. The brothers insulting him was all kinds of snacks, for a snack, he crawled on the ground like a dog. The niece told others that he had dignity, but he said that he said that as long as he learned the dog, he would not hit me and gave me it. The niece comforted her not to be sad, but he asked the niece, what was sad? They are playing with you, don't you want to eat it for me? No, can I kill them? Gradually, the niece found that she was different from ordinary people, and maybe she was really an ominous monster. Soon after, Yantai was sent as a proton to Saint Valley. Continue to be beaten by the prince here, so I don't believe that there will be no tears in the world. Broadcasting for me, I really can't. In order to make him cry, the fat prince simply broke his arm. Is it okay? His Royal Highness Shi hasn't cried yet. The childhood of Yantai was extremely dark and miserable, but he was also eager to be happy. Whenever he saw other princes was held in his arms, he thought about it? He learned to smile hard from other children, but he couldn't laugh. Ten years later, Yantai finally grew up. Help me write a similar copy
Help me write a lyrical love song
Grateful for advice🙏🏻
Auto-GPT is an experimental open-source application showcasing the capabilities of the GPT-4 language model.🌿
1 ¯_(ツ)_/¯
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
what buffer?
feels like that some days
you just leaked your ip
It happened on me,too
Can people do anything with ur ip other than trace location
what ?meaning?
and there are those type of people here
delete that and send it without the IP adress
Delete that it contains personal info
It's like ur social security number for the internet
Where to deleted it ?
Not you lol
Its really painfull for me to pay for a subscription and im getting this after every prompt someone from openai please help me:
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.

Access denied
You do not have access to chat.openai.com.
The site owner may have set restrictions that prevent you from accessing the site.
Ray ID: 7b86574b98334a6b
Timestamp: 2023-04-15 18:44:38 UTC
Requested URL: chat.openai.com/
Error reference number: 1020
Server ID: FL_410F175
User-Agent: Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/112.0
i can not open “https://chat.openai.com/ 1”,why?I just found this problem.
i do nothing ! Is there a virus on my computer? Twice last night, my TightVNC automatically disconnected.
Ask for help !
I can’t find some officer to ask for help, so I have to ask for help here, can anyone help me?
For the love of God
like this?
Maybe this guy is a scammer
who can help me ?what‘s the problem?
Try using Google Chrome
Maybe duckduckgo , brave and such browsers will do too
Make sure not to use a vpn
ok i try it
Someone please help me 🥺
Its really painfull for me to pay for a subscription and im getting this after every prompt someone from openai please help me:
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
i am in China ,ℹ am log in openai,by My singopore Socks server.
it not supported in your region .
Refresh your page
can i get gpt-4 for free
Don't know that...
I know that doing this solves the issue but its really irritating that i have to refresh page after each prompt
My server's physical address is in Singapore. I telnet directly to my server and I open firefox on the server.
@flint owl !
May be it is about GPT-4's problem?
They do not have extra room for more customers?
😅
well then, may you should look at the official relevant terms.
I'm experiencing an issue with the ChatGPT when using it in Chrome. Whenever I don't type for a minute, I get a network connection error and need to refresh the browser to continue. I've already tried clearing my cache, disabling extensions, and updating my browser, but the issue persists. Can you please assist me in resolving this problem, or suggest alternative solutions to avoid the constant need to refresh the browser? Thank you!
I just watched something interesting and I'd like to know your opinion on how an AI would react to that. The situation: two human players. One bord game. The game started as Go on a normal chess board. But then once one of the players was about to loose he, without a word, did moves according to Checkers rules. And the other player just continued playing using checkers rules now.
Anyone tried something similar already? How does GPT4 react to the opponent breaking the assumed rules of a game?
hey sorry this is random af but any flutter devs here
Asked it to generate ASCII art in HTML of title "ShadowGig". Turns out it's as clueless with how letters look as Dall-E. It probably understands meaning of letters, but has no ability to recognize them by sight.
The effects were Sirocksic! and GHAYKORY titles. Hallucinating
can i give image input with chatgpt plus, or only text
Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
Why is this happening? Shouldn't this kind of situation only happen on free accounts? But I have already charged for Plus.
i don't think that's how it works. Plus users get priority over free users, but if it's down, it's down.
they're trying to get better servers and stuff so it doesn't go down, but nothing is ever perfect.
Just leaving this webpage for a short time will prompt an error message, and the webpage must be reloaded to use again.😢
ChatGPT is not available in your country.
i have the same issue but no answer from chatgpt team.
Please delete this message— it contains personal information.
that's pretty normal if you know anything about websites.
just refresh it
not normal it is so boring who can every 30 seconds refresh the internet browser to ask new thing?
30 seconds is ridiculous. i go hours without refreshing mine and it works fine.
that sounds like a personal issue.
Currently the robot checking is increasing. Every quickly it starts to check the robot
This issue disappeared after I charged for Plus before, but it reappeared again today.
i only refresh ater my computer is closed and i open it back up, or if ive been away from my computer for a while.
it wasn't a Plus issue. ChatGPT was just down.
you paid for something you didn't need to pay for. and the fact that it started working again after you became a plus user was just coincidence
mine is chatgpt4 not a free version but have a problem network error.
alright nvm.. i'm out. too many dumb dumbs.
the disconnects might be a regional issue, as I've been using GPT4 and 3 all arvo without issue. I'm in Oz
paying for a plus account doesn't mean that it will always be online for you.
you're more than welcome to host it on your own server, assuming it has the storage space, and the CPU abilities to actually run it.
what? lol... hope you're being sarcastic...
thank you very much
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
i am. i'm a jerk.
Thank you for your response.
i apologize. but if it's not responding, it just is what it is.
it could be regional.
i'm in the northeast US, and i haven't had a problem with it going down for several weeks now.
and i'm not a plus user. i'm a free account user.
let me wait and see for a while, maybe it will be better after a while.
i just checked my side. it's working smoothly.
it's prolly just a regional problem and will be resolved soon.
things go offline. it stinks. even YouTube goes offline sometimes. not often, but it does.
as far as i know...gpt is just a natural language bot and doesn't generate images.
but you can use stable diffusion to generate images.
there's a lot of online tools for that, or if you have an NVIDIA GPU you can do it on your own machine. it doesn't work well on other GPUs though.
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
why make use GPT-4
that function, though demonstrated by the OpenAI team in their YT vid, isn't available yet for general use.
do some research. lots of info and vids explaining use of of GPT3/4
GPT4 is much better than GPT3 but often isn't worth using because of the extremely strict cap in place, and the slow text generation
GPT3 poos all over it in that respect
hi guys, how are you? i have a question regarding my chatgpt plus subscription. my account says that my gpt plus contract is temporarely suspended. Normally in this field is written the renewal date. i havent received any information about it, but could this be a bug? Could not find anything about this issue 🙂 thank you guys 😊
Hey guys, do you feel like things are keep getting removed from gpt4 instead of added?
I keep finding news related to GPT-4. Things people can do and stuff, but once I try it I get answers like: as an AI, I can't do this and that. Or it actually does it but in a very different manner than the news or the post.
I think something might be happening and they might be "cancelling" plus memberships. I paid mine and got a reimbursement.
oooo they maybe increasing the price so that we get to use GPT4 within it...
How exciting would THAT be?!
if they keep making GPT4 more limited to the users I don't know if it would be worth it
Well not really, I mean if they increase the price so that we get unlimited access to GPT4, we all benefit
Right now, theres not much you can do with 25 messages every THREE hours
it's a bit of a joke.
of course, but I meant, I constantly bump into limitations, and it feels like it is getting more and more limited. If they don't remove these limitations, I don't see the point to pay more.
Like in 25 messages they don't remember the first messages anymore
oh I see what you mean. Yeah that is too bad. It's because of DAN and Jailbreaks
DAN?
As far as I remember, the limits several month ago was 100 message every 4 hour
Or I am wrong?
Do Anything Now
It is a set of specific prompts in order to remove ethical limitations. It turns Chat GPT into anything you want
But now it doesn't work due to the afore mentioned limitations you are unhappy about
you got me excited for a moment
ChaptGPT is now just as strict as Bing AI
Which was my fear, as I think Bing AI sucks monkey balls due to Microsoft taming the f out of it
GPT 4 user from UK, is it down because every time I ask it a message I get error message and have to refresh and do you know when this problem will be fixed ty
Is there any third-party client that is easier to use now? Logging in to the web every time is troublesome
poe?
@spice anvil how can I get mod role?
I too wanna to serve peoples
I opened a plus account, there is a limit to the number of poe
what is the link for chatgpt4
But you need to pay to have access the gpt 4
Why can I enter the dialogue interface on my GPT, but the version selection box on it keeps refreshing and cannot be clicked
What country are you in ?
I am in China and have been using the Los Angeles network for the past two days without any problems
Today, it suddenly happened like this
That's the issues. Unfortunately we can't help you since that go against rule 0
ok Interesting. i wrote to the support, so i hope to get a reason for it. did you also habe this message on your chatgpt account sushscription, that the contract is temporarely suspended?
About image recognition, we can find this in the FAQ "This feature is not yet available to the public."
I have a plus account, and I can't use image recognition, why is that? some of my friends do have free tier account and can give a link to the chatbot and it does describe the image.
Thank you very much. It's my network instability that has been resolved
Nope
why I haven’t passed the application to use GPT4 until now
function is not yet available. doubt your friend succeed. gpt often hallucinate
not sure, at least it doesn't show them the following error msg.
"As a text-based AI, I'm unable to view or directly analyze images"
and describe the photo. I've seen the result.
It's hallucinating.
I think sth is very wrong, and the openAI refuses to publish the facts.
Well if you provide a link, it can read the link and use that to determine possible contents. It's not accessing the image.
Who are you developing
AI robots
hiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii
Hey guys, I recently found out how to use Gpt-4 for free with extra things included, would anyone want more information?
👀
yes, is that bing AI
yes...
but you can also get dall-E 2 that is unlimeted with bing also
yeah but it is (in my opinion) worse than the official dalle 2
but midjourney is the best
if you get the bot, swich it into creative and say:
create (insert what you want) and i will make an image
using that you can say, i don't like the background can you mak it a different color?
what does that even mean? I tested different images hosted on different website, and it nearly describe what's on the image.
By providing the link to the image?
Yes, Midjourney is much better in human characters generating than DALL.E.
is there any news when plug ins available for everyone yet?
I don't know, I am waiting for this day every users can try the new feature plugins.
same
Nope, access is pretty limited currently
have they given a general time frame even?
None, it's in alpha still
ok thanks
Anyone who is using autogpt is it taking anyone like 10-20 minutes on thinking?
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
is it normal for you that gpt-4 always stop his response after 850 tokens ( something like 2000 characters) given that it only gives you 25 requests per 3 hours that means you can barely utilize it within 25 requests
is there a tutorial for how to decide how many tokens a string is?
Hey everyone! I’m coming for my weekly check for updates on a possible change of the cap of 25 messages ☺️ has there been any news about that? Cheers
anyone using autogpt
Yes, the news is that it still has that limit but he stops his response after incredible low amount of tokens that means you will use your requests very quickly
Hey guys why is my gpt 4 option not showing?
like there is no option to change from gpt 3.5 to gpt 4
your subscription expired
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
No its still showing chatgpt plus as my plan for subscription
when does it expire, usually this happens an hour or 2 before expiry
prompt
is microsoft bing offering chat gpt 4 in their chat section
bing is GPT-4
Hi, im having some issues with the open ai version of chat gpt, after i ask it do anything, it takes around 2 minutes for it to generate a "error occured" message, any advice?
also please ignore my username
Not a product of openai and not same as chatgpt 4
怎么看待电商行业 ¯_(ツ)_/¯
English only
I recently received a survey about developing an exact identifier for ChatGPT-generated text. No, seriously, with all the abuse and repression of ChatGPT people, how dumb do you have to be to suggest something like that!? 
6
I am curious that why the bot didn't delete your message which is not written by English. Haha.
seems like chatGPT Plus subscription doesn't ensure experience these days...
I have already encountered an argument that using GPT even if you heavily direct its output means you aren't author of the written piece so... yeah it's like when people started to use digital softwares to draw vs physical
I constantly get error messages after couple of seconds
even when I am a Plus subscriber
did they make any changes to the protocols?
me too
sometimes the service might be under heavy load or having some other issues... usually refresh helps
if this kind of limitation still exists when I'm a plus subscriber, I don't know what's the meaning of paying
I can't keep asking because I need to refresh every time, just like the previous free account
I almost have to refresh the page after I send a message and get an answer
same here
yeah... that sounds bad for something you pay for... like for me it has been rare but with almost every message... that's just ridiculous
I means they are in beta , bugs can occurs with 150m users within 4 month
Don't play if you can't afford to
yeah I'm aware that the thing is in the beta... still their experience is way less smooth than for me
for me the service has been working rather fine
I am also receiving the error message: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.
What country are you in
also paid account and receiving the error in both 3.5 and 4
from Spain
using VPN but also connected to Spain, haven't have problems with it...
- disable VPN
- disable add-on and browser extension
- try with another browser, I suggest something non chromium like Firefox for the test
Vpn often create bug with openai
Firefox not giving me any issues
ok, gonna try, I stay online to live feedback
Thanks 🙂
How can i create an AI video/image and where?
What is exactly going on here
Dalle for picture
Gpt 4 discussion channel
I don't understand
@grand maple sorry Ai lord
Relevant Links
💬 Try out ChatGPT at https://chat.openai.com/
🤖 Access GPT-3 at https://platform.openai.com/
🎨 Make art with DALL∙E 2 via the API or in https://labs.openai.com/
📩 Read FAQs and reach customer support through https://help.openai.com/
hello i want to buy chat gpt 4 premium can you help me
?
When will DALL.E 3 be released to public? Is there anyone know about that? Or, there is just no DALL.E 3 model yet?
Please use #images-discussions
what the hell why plus user also have this cloud verification now..
What country are you in
my account has been blocked after multiple consecutive login attempts
You will need to contact support@openai.com
Ok, thanks.
ok, using Firefox, disconnected VPN, working by the time
thank you so much for the information
Currently residing at dieago garcia
UwU my pleasure, have a good day
Please debate that in #off-topic to keep the channel discussion focus on openai gpt 4
"Apologies for the confusion. My mistake. Here's how to..."
Is becoming very common for GPT-4 what's the matter OpenAI?!?!?
Pis_ing off already
How do I access chatgpt 4 for free legitimate
anyway, not so much sense to have problems with Edge or Chrome, doesn't it? I understand interferences with VPN, but with the navigator, and also Edge with how they are connected with Open AI....
Clearing cache, cookie, or sometime add-on or extension creating bug
You apply for the api and wait. Keep in mind the api will be free till you reach your trial limit
I have applied and I have waited
Wait till you have email invitations
I'm still waiting
it must be it, I will look for, because I prefer using Edge. thanks
Gpt 4 isn't complety free. You will receive free amount to play with but after you have to pay
How do I access the free gpt 4 ?
Why can my account be used normally on the mobile phone but abnormal when I log in on the computer?
Or is that the try on openai ?
I'm Peter
Dall-E is such a joke lol. They really need to improve it
peter
What's the topic of the day?
Peter is the topic
🤣🤣🤣
Let's keep that in #images-discussions please
You have something that block on pc you have to troubleshoot it
- antivirus
-vpn - add-on
- not properly clean cache or cookie
Agree, midjourney is better in human character.
#ai-discussions please
sorry about that
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
will there be gpt 4.5?
I sure hope so
I lost my access to GPT-4 can someone tel me why?
I don't use it a lot but it disappeared and I have GPT-Plus
maybe you asked weird stuff lol