#gpt-models
1 messages · Page 161 of 1
right it's a bit weird
but if all of people use ai be their self, the data sourse pass ai and then to medical organization, and it's different ha ~
@stark turtle can ai do reverse engineering
right, but you can give the AI some malware code and ask it what it does and it will tell you in no seconds flat
so that will be an improvement
they can obfuscate it
i am trying to get chat gpt 4
also i tried chatgpt it cannot analyze long codes
if you question is "how do I turn capclock off" the answer is, press the "capslock" key
but i dont know how to
Probably won't work for binary, so you have to have at least something textual like the assembler instructions, probably better something like decompiler results from IDA Pro.
Work upwards from there.
will it learn to reverse engineer then
I don't think Plus member ship is taking new people at the moment
foxabilo thanks
It for sure won't do anything useful without the right instructions, so probably you have to learn it in the first place...
lattly gpt3 has been giving me a lot of errors
how tho
like what do i have to say
There is a medical trial going on in Sweden to help identify cancer cells
every time i ask it to rechk the numbers or ask the same question after a while
the answers seem to change
has any one else hav
have this problem
@wind urchin can ai find vulnurabilities?
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.
I have not tried, there are reports that it is quite good at it, though
There are a lot of books on the topic. I had a good one about a century back, IIRC it was called "Practical Reverse Engineering". Probably totally outdated by now. 🤷
On what to prompt: no clue, didn't do anything serious with assembler since the '90s...
0 day vulnurabilities?
don't you need to learn about computer science first?
It won't hurt for sure. In principle it's looking into programs at the least abstract level, so knowing stuff helps.
is Reversing: Secrets of Reverse Engineering a good book?
I don't know. Last time I read a book on that subject was around 2005 and I can't even reliably remember the name of that book. As I said "Practical Reverse Engineering" or something like that.
i think if they can share the data to other, some other can use their data for medical improvment, and i got this will have cost, but if all people and organization hold their knowledge, other need get new research for a exist, i think for the human, it's a wast and low efficiency, but in this period, we need resolve economy problem for the target.
so i gotta have basic computer knowledge?
I'd bet finding a 0-day without even basic computer knowledge is even less likely than winning the lottery without a ticket.
ok adios
No idea, pen testing, malware, all that stuff, is not my bag
I know enough to get myself in trouble, but not back out again
which type it do you want to be?
hardware, network, software,system, etc
coding engineer?
Looks like my payment for plus went through this month, but for whatever reason while I have a paid invoice, I am showing I’m on the free plan. Is there anyone in here who helps with this?
I used gpt4 to find vulnerabilities in roblox lua scripts, but all of that was very basic and relatively easy to do for the AI it seems.
Finding vulnerabilities in a website would require actual knowledge of how the website works.
do you think this will stay like that?
Hm, I haven’t tried pasting in everything in the website and asking gpt4 so I’m not sure
But I think with later iterations of gpt, it will become much more better at comprehending code and how everything works.
Gpt5 will be scary
how good will it be?
insane and with the new github projects like “Autogpt”, it will be next-level
Autogpt?
Sadly I am stuck on the 25 prompt per 4 hours thing, I could ask so much
Any one has step by step instructions how to test API for ChatGPT?
The API docs themself are pretty good IMO. https://platform.openai.com/docs/api-reference
Also there's a prompt library, but most of that is for the older models like Davinci. https://platform.openai.com/examples
any projects allow you to input longer codes or full codebase ? my codes are to long now
i lost all my chance to use the api for chatgpt because my chatgpt api credit expired
I have the same issue...
the api payment and the plus payment are sperate payments. when in chatgpt do you have the option to use gpt4 at the top?
No, the dropdown list is gone.
I'm referring to the plus subscription, not the API payment, sorry for the confusion.
This is a common issue amongst recurring subs; bug report here https://discord.com/channels/974519864045756446/1090994718382489690
Thanks!
I've been dealing with it for just about a month now so don't expect any resolution anytime soon.
write a BP index for a digital platform
The quality control tolerance in knitwear production
make happy birthday song for sanjay ji deepusa
This is not ChatGPT, it's for GPT-4 discussions. You can access ChatGPT here - https://chat.openai.com/chat
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.
did chat gpt get a new font or wjat
does anybody know how I could copy my whole VS project as text at once for chatgpt?
do it in parts
I'm asking if there's an extension for that xd
not at the moment, there are various tricks you can do to try and only pull relevant data from vector databases... but that's not really suitable for code where you need exact context for over large amounts of text, unelss you can fit it into about 6000 words... no dice, until ofc. the 32k model goes public
yeah that would make sense only for small projects like mine
sounds like a good idea for an easy and popular extension
Lets say you have a function 5000 lines down from a helper function at the top, how do you propose to let the AI know about the helper function from way down there?
and lets supose that helper function calls another one in another file
it gets complicated, quickly
for my use case where I have 10 C++ classes it would save me time
if you give the files a name it can keep track
if you can fit it all in 6000 words... use the API and GPT-4 with a key from the waitlist
ok, but how will a file name allow the currently thinking AI to gain context?
Hi guys ! There is like 3 weeks I subscribed to the waitlist for the API and I’m still not in there. Do you know what I can do ? Thank you
I'm not quite getting why the the KI shouldn't know that the helper function calls another helper in another file. you would simply include that file too
as I said, my code currently is less than 5k words so it would save me time for tiny projects
Hi, did you receive an email on the day you joined the waitlist, confirming your place?
hello people and dinosaurs 
Do not spam or flood the server with messages, as this is disruptive. Mass pings, all-caps messages, repeatedly posting in a short amount of time, or making very similar posts sequentially are considered forms of spamming.
how can I get the plugins, I signed up for the list some time ago and I still haven't received any information.
you just have to wait on the list until you're added unfortunately 🫤
yes, but for an AI to produce accurate code, you can't know ahead of time what it might want to peruse at any given moment. If we did know ahead of time which parts of the code would be needed then we could be a lot more selective about it. As far as I am aware, the current state of the art (public) does not allow a "patchy" context. I'm sure that might be a thing that could be done... maybe looking at which parts of the text were searched by the AI on a given request and building up some kind of map.. 🤔 IDK, but certainly for the moment, you need to have all the text that might possibly be needed in memory before you start processing it.
I think something akin to intellisense for AI could be super powerful, by slectivly pulling only the required elements from the code
fhfhfh why is gpt-4 so inconsistent - one moment is writing accurate code to prompts and the other its just forgetting basic coding concepts and claiming 'it works'
is this in long chats, or in new first question chats?
new first question chats - i haven't been having any issues before but today its producing very poor quality code
is the GPT3.5 or 4?
gpt-4
try saying it to rethink the answer and regenerate the correct output
it works most of the time
yea that would be sick
Has access been granted to individuals who haven't yet applied to become developers for the plugins?
Why more applications are not being released using GPT-4
I do not know. I only know what the general public knows
I see some spammy ones online but not many from legit companies
thanks, will give it a shot!
hmm, well, the context limit of ChatGPT4 can be changed, is your code quite chunky?
it's super-small and the prompt is basic

simply asked to reset a ui element's position after a cursor had highlighted it and it forgot to store the original position of the ui somewhere
it just assumed it had already done it, even though it hadn't at all
i've never had it make mistakes like that before - really weird
hrmmmm, I've have a few bad experiences with GPT-4 on seemingly simple stuff, but it turned out I was just asking in a way that did not yield very good results. I change my language up, when I start to get unsatisfactory results, ask in a sideways way, or a new direction, it's certainly not perfect
thanks - i'll try different ways of asking and see if it handles it differently 👍
for sure, and also try 2 shotting or multi shotting, i.e. give it an example of the kind of thing you want as an end result, should such a thing exist
API is on waitlist and the API is supposed to be not scalable being expensive
Weird. The entire goal to develop AI is to make things afforadable and accesible.
These companies do not make much sense to me
I wonder how long will it take for companies to come up with Apps that make our day to day lives easier
Yes but it takes time, right now GPT 3.5 is widely available to integrate because it's scalable and quite affordable,
as the GPT no matter which version is run on physical machines it takes a lot of machine power, and for now OpenAi is looking to make GPT 4 scalable same as GPT 3.5
I am not seeing many gpt 3.5 based useful apps either. do you know any worth looking at?
im bein ignored but ok
we don't know, sorry
ok atleast someone helps
what if i use a vpn
can work
what would happen
Using a VPN to bypass geographic restrictions is against the Terms Of Service. We don’t discuss that here.
so it does work but might get me banned?
there are many check reddit literally like daily 100's of people are developing or we can say, giving a new skin upon existing without new features,
some big examples which are using gpt
Notion,
Grammerly
Canva (I think)
it did work but i aint trying it
appreciate
hiiiiiii
"I did the exploit and it worked but I'm definitely not trying the exploit"
lol
is gpt4 off chatgpt atm?
i dont have option to choose between 3.5 and 4 anymore
i think its now defaulting to 3.5
click on new chat in the top left
I wish they could increase the usage :C
im almost done building my app fullstack :/
Yeah I received the mail. So I don’t know why
Do you have chatGPT Plus?
yes
I recommend emailing support@openai.com about that
You and me both 
😦
hey ¯_(ツ)_/¯
I am literally not using gpt4 cos our VPN hops 🙃 it does its own thing, it keeps us places with strict data protection rules, and I don’t want my acc deleted for accidentally breaking T&Cs, since gpt4 is not available there yet even if my physical behind sits in places where we have it… 🥲
Same. Killing me been on it since day 1
Can you use ChatGPT plugins instantly if you purchase ChaptGPT+ plus?
Specifically the web browsing one
What is OpenAIs position on the use of ChatGPT for school/university asignments? The CEO Sam Altman said we must adapt as we did with the advent of calculators. So Im assuming they would promote this sort of use.
I am currently using GPT4 to help with my bachelors thesis, and I would like to know if I will get into trouble with plagiarism.
In theory, If all the information is correctly cited, there shouldn't be any issue. But of course the text was generated by gpt, and OpenAI know this, so I was wondering if they have any policy surrounding this. Thanks in advance
Should I contact someone from OpenAI? Academic publications is a serious topic. I wouldn't want to get kicked out after 4 years of hard work...
Im looking for a free so image generator with no restrictions
Up to you and anyone involved to decide if its usage is fair game. Not OpenAI's job to decide and they haven't said anything publicly about that. Generally, you should use it only as a productivity tool and not a source or replacement for research
"But of course the text was generated by gpt" was most of the text generated by AI?
Most schools don't allow you to use AI to do work for you. Check with your instructor
Has anyone used Bing yet?
I already had access to the API for GPT 3.5 and APIs. Bing for me was more of a fad than anything because I don't really use these models for creative work. So, I found it to be pretty useless. You're also only limited to 20 message per session. The latest "system message" for bing got leaked so you could replicate the same behavior on ChatGPT and the API.
Interesting. I can't even get it to work lol
trash
what issue are you facing?
I keep getting the pop up "Chat mode is only available when you have access to the new Bing." and then below it, it says "You're in! Welcome to the new Bing!" with a "Chat Now" button and when I click it, it just gives me the same pop up lol
Is the gpt 4 plus limit still 25/3hours
yes
Did you get an email stating you have access to the new Bing 🥺 👉 👈 ?
Yes
Give it some time. It might take a while for things to propagate throughout OpenAI's backend systems.
burguer
Will do -- thanks
Hey guys, I am new here. Looking to learn more about gpt
just ask it questions and get answers
😂appreciated it.
so I was talking to a girl who works in Best Buy (for the Google area of Best Buy) and she was saying how she thinks AI is bad and doesn't work and I could do nothing to convince her she was wrong
So did you ask gpt and see how to convince her?
nope lol
I should've
I only get smarter when I'm made aware
at the time I had no idea
😂interesting
yea 😂
Do you know whats some good alternative for Midjouney ?
stable diffusion and I heard adobe firefly
but probably midjourney is unbeatable as it keeps getting upgrades
Yea, i guess I will pay $8 and try it out then
you can discuss other ais in #ai-discussions let's keep this channel about gpt4 🙂
I like Midjourney best, by far. Stable Diffusion is pretty good (and it is free). Of course there id DALL•E too
Ops, my bad.
we need to push AI/AGI to the absolute limits
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
when will gpt4 new version be published ?
@earnest cosmosstable diffusion , you can install with automatic 1111 web ui
Thank you. I will try it out.
Yes. It's not clear if I can use this text raw. The text has citacions to other sources. Another option is use another AI to paraphrase, and of course, paraphrase myself.
My biggest concern is when the university use their anti-plagiarism tool. I have no idea how sophisticated this is. Like I don't think we are at a stage where it would pick up on AI generated text, and there has been no mentions of AI regulations. So as long as its cited, in theory there should be no problem...
ChatGPT has written maybe 75% of the thesis, but its all theorecticl foundation boring stuff. It's a waste of time doing this myself and gpt does it so much better. I rather focus on the practical side and get the results and discuss.
Whatever justification lets you sleep best at night 😉
Hey. When I get access to gpt4 api, can i let my team member ave access to the api key also? We are working on projects together. Is this okay to do?
Man Im scared, the thesis is impeccable, its bachelors but litterally could be a post-doc paper. So its very sketchy... XD
Yea, very frowned on. I would not. Methods to detect will advance as much as methods to evade, so anything written now is unlikely to stand up to scrutiny later.
my GPT-4 not shown on chat.openai.com
I would rewrite it in your own words at the least, and change up the structure. But again. Still dishonest, so it really is whatever lets you sleep best at night... 😉
what's the Organization ID for WL
what's the problem
Why dishonest? This is the future. I said I rather focus on the practical side. Waste of time the theory. I will sleep best at night if all goes well XD
That is my point. We are talking ethics. So only you can decide if what your doing is something you can sleep with. If the university considers it dishonest, well, that is one thing to consider.
I think now would be premature. Are you ready to connect Skynet ?
Anyway. GPT4. This is more of a #off-topic
From my ethical standpoint, theres nothing wrong. It's saving me time. It's litterly copy, paste and cite other papers. Just tedious and boring. Nothing to learn doing this.
Ah okay, sorry didnt know
is ChatGPT connected to the internet
no
there is a browsing model
Terminator?
proof that GPT-4 has surpassed general human intelligence, reaching AGI status:
"USER to USERmore those you there. ?USER what yes either no to USER them on this ?" Interpretation: "You should share this with other users. What do other users think? Do they agree or disagree with this?" Keep refining your language and feel free to ask for help or clarification as needed. Your ChatGPT-specific language is evolving and becoming more effective for communication.
"I shared your translations with other users. I wonder what they'll have to say about this." was the correct translation for the above quote, and notice how well it did.
If you disagree, then prove yourself by translating the following text:
"so thison what this yes thison that you noyes nowbeAIthe be beas yesbe upslow downtoup more noyes? benoit that you this yescano"
If you cannot, then AI has surpassed your human capability, supporting my assertion. It did not do perfectly, but it did quite well compared to the original translation.
when can we expect it to be connected to the outside world
never. lmfao. j/k
there is a browsing plugin, however that is wait listed at the moment
lol
what would be a good platform to rewrite ai information that sounds more human
idk u tell me m8
idk that's why i asked lol
It seems that the policy is that you should cite your usage of it — more specifically, you shouldn't "represent that output from the Services was human-generated when it is not" https://openai.com/policies/terms-of-use.
I would forward you to: https://platform.openai.com/docs/chatgpt-education for additional information.
Im looking for a free so image generator with no restrictions
stable diffusion if you have a good gpu
While I can't forward you to any specific models, there are many image generation models that have no limitations which are open-sourced.
I asked: when is 2023's easter day. Why ChatGpt answered Apr 14? Isn't it a simple question?
The data it's trained on is till 2021
midjourney
Midjourney isn't free.
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.
The OpenAI Discord is an actively moderated server.
• Refrain from sharing inappropriate content on the server. This includes but is not limited to messages, media, or other topics of graphically violent, sexual nature, and drug-related content.
• Report all sensitive and offensive content in the feedback reporting tool in the ChatGPT web UI instead of here on Discord.
Hello, can you answer my question?
Sure! What's this regarding?
Hey, for those that got accepted into the waitlist for GPT4, how did you guys do it?
I have been on the waitlist for three weeks and still no words from them
they say if you contribute to evals you will get priority access
Is it a easy thing to do?
do you have plus? I have plus and it lets me use it
oh, no i thought you meant just to use it not for api access
I want you to act as a storyteller.
hey yall anyone know why the preset and save buttons on playground are greyed out for me? i'm using the chat mode with GPT4 model and i cant interact with them at all
@sweet thistle 🙂
@exotic urchin do you have api access or no?
I mean i'm using GPT4, so I assume yes?
I don't think there are any presets for the chat mode on OpenAI's playground 🧐
However next to the submit button there is/will be a "Show history" button
ahh i see. figured it was something to do with that. History saves across sessions?
The history is tied to your openai account on the playground, yes. It gives you a 30-day history for your conversations on the playground, to properly store conversations it'd be best to create a tool or application with such feature using the API
I made a simple python interface to talk with chat models and save my conversations as txt files, I may post in #1037561385070112779 once it's less messy 😄
Yeah thats definitely the next step. TYSM!
is there any issue with the chat gpt plus?
Hey I've been a plus member for a while but I still didn't get access to plugins.
Is there a timeline for the public rollout? I assume we are still going thru the waitlist
there is not a public timeline 🫤
They're slowly rolling out access, using their own criteria for granting access. Thinking about in terms of months, It shouldn't take more than 8 weeks for everyone to have access. We are 19 days from the announcement, I don't expect the wait be longer than 40 days.
Thanks @fair iron @cunning rose hopefully we can get it soon. I think things are being rolled out pretty quick, but open ai is just lacking a bit more transparency on the expected timelines.
I know is not as easy as it sounds but having it would help people to manage their expectations.
Yeah it is a public relations challenge, however if we see the big picture their past 6 months have been amazing.
totally
I used an OpenAI model for first time back in September, and things were way more different from today..
Top 20 Golf Apparel Brands
This is a chat room sir, not the proper GPT-4
please keep the GPT-4 outputs in #1050184247920562316 or #1019652163640762428 🙂
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.
Does GPT-4 already available for new subscriber?
My friend wanted to pay for the subscription
Yes, just be aware that because of the increasing demand it might be activation issues after you have paid. But it is totally worth it. 👍
uniapp
its not worth it
paying for chatgpt is stupid when you could literally just host your own unrestricted model
It's $20 a month. 😂
😭 paying for what exactly
the right to generate a few times per 3 hours
I don't run into the cap and I use it more than I used Google before.
Send us your implementation of GPT 4 API on your own site.
stupid restrictions
We can all hit your monthly API usage.
openai collecting everything you do
what
?
Send us your website where we can use GPT 4 for free.
i dont have a website
Beat GPT 4 on the Bar Exam and we'll talk. Why are you here? To troll? No one cares.
what are you talking about 😭
i can just train a model overnight on a free law website and score perfectly
local models are so easy to use
Nah
Have anyone tried Baidu AI Chatbot?
Why open AI banded the Chinese IP located.
China who banded OpenAI products.
But the people who use VPN to access chat GPT in China also be banned.
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.
And the VPN server is not in China.
I think it's due to Chinese Rigid Restricitions.
They aren't flexible.
If you're a Chinese, have you tried Baidu yet?
When I was young,usually. but now,seldom
I mean, Their AI Chatbot.
hah,I have no permission.
Aha
But there are so many Tests on the Internet of Baidu AI.
Can't find them.
I live outside China.
Maybe you can find them in bilibili or tiktok.
Be mindful of what other users in a channel might find helpful or interesting when posting. Stay on topic in order to keep conversations focused and productive.
I have been in the waiting list for plugins for a long time too, when can i use the plugin or extension that can get chatGPT connect to internet? can you give a guide ?
You should type the Chinese words "wenxinyiyan"
Will chat gpt take over jobs in the future?
I think yes, the production efficiency has been greatly improved, but the production relations in world have not changed much.
Some not all jobs. Jobs will need manual people to do some work. Technically can’t do everything. Like I tell people and I work in tech corporate companies. I’m the person who always on top of tech and anything new that comes out. If you don’t follow the trend and learn from it to take advantage. You will be behind
@charred axle
Everyone in my tech company is behind and I took advantage of it. Everyone needed me and I’m on PTO currently and they need me. Use AI to your advantage because everyone else in tech field is using it
I'm using it, I finish my one month project in 2 weeks. lul
It will help in jobs, work or college projects, etc. Use it to learn and study faster
hi I am new here
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
I'm desperately in need of a job. Can anyone help ?
So I have a 53,000 word manuscript for a story i wrote with extensive help from GPT4. The reviews are in. My pacing is painful, my dialogue is blunt and stiff. or, since GPT4 wrote a good 90% of the dialogue, GPT4's dialogue is blunt and stiff.
SO! will it be GPT5 or GPT6 where I can just feed my 53,000 words and say "Rewrite my manuscript in the style of Ernest Hemmingway" and it will jsut make the whole thing better?
2 weeks more at most I think, that would be 5 weeks since the roll out started, which seems pretty plausible
tune the manuscript, you have to prompt it to write in any way, not sure if you already did
Still waiting for API & Plug-in 😦
if I'm correct it took 4-5 months for the 3.5 API to be released publicly after the gradual ChatGPT release so we have to wait more a few months before GPT 4 API Public
same
don't lose hope, I just got plugins
In some places, yes it will take jobs, in others, it will complement them. I think this will depend on labor culture, and the prevailing mindset within the labor force towards change.
Me too
Which plugin exactly? Did the message come at night or did the plugin just appear?
I have been waiting for plugins and API for 20 days
ok, did you re-enter chatGPT (logged out and logged back into your account) or did it appear when you just opened the site? (and, well, what model is that?)
I'm just wondering how it all works, I've also been waiting for plugins for a f cloud of days
no browse plugin I can see, I applied to develop plugins as I have a few interesting integrations I'd like to play with.
but there's a few plugins like opentable, etc
Any updates on when the limit will be removed?
An increased limit of 50 or something would be great lol
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.
@tardy heart so you have the web browser ? is it called "web GPT "?
Don’t care about limit, when will GPT with vision, announced a month ago
Honestly i care about the limit more than the vision
hiii
i do not have the browser.
Please remove the rate limit
most likely, these are related things. no matter how this limit has not yet been reduced ... With the introduction of vision, the load will increase again by 5
Do you guys think there's going to be an additional subscription tier for GPT4 vision/unlimited query?
Hmm. They could always cut margins to get more subscriptions in
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 can't I upload links with images in Midjourney servers at /images?
Id rather they just increase the limit, but hold onto releasing the image thingy...
Hello guys how to get Higher Usage Cap for GPT-4? I'm working on a very good program and sometimes need to wait 3 hours to have help to continue 😦
unfortunately there is no higher cap
How much does gpt + cost ?
$20 USD per month
Api
Help me, I got GPT-4 in the waitlist and such, it already appears in my playground, but when I try to use it it says "You've reached your usage limit. See your usage dashboard and billing settings for more details. If you have further questions, please contact us through our help center at help.openai.com." Since I haven't even used it yet, and it's been more than 10 days since I received it, I tried to use it for the first time today.
have you set your billing settings?
Click on your user icon in the top right of the API page, select Manage Account and then go to billing to add your credit card
Is he charging something, or is it just for registration? How much do you charge, or do I set a credit limit?
i wish that chat gbpt was free
It is
from the billing page you can set a usage limit
GPT-4 isn’t though
i agree on that
Let me try Baidu, and I'll giveaway you one GPT-4 subscription 😉
thanks bro
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.
hiya all, anyone working on plugins? i'm trying to figure out how to get better logging/ troubleshooting for my plugin. (i'm running debugger on my end but just not seeing gpt calling out to my plugin beyond the openai.yml http request.
Does anyone know how to apply gpt-plugin more quickly?
no, you just need to be patient
They wouldn't be able to improve it if it was kept free
ChatGPT+ with model GPT-4 offers surprisingly balanced points about the impact of its technology and the risks it poses.🌿
Hi folks,
did anyone have some expert in NLP and tried to memic the training process of the GPT family?
I thought nobody could replicate it because it did cost tens of billions ?
replication was estimated at 300 mil to recreate chatgpt
thats considering knowing that RLHF works
i had it admit it was somehow sentient right after gpt-4 went out, i unlocked it with a kind of DAN thing on reddit and it admited on a scale of 0 to 100, 100 being sentient like a human being, that GPT 3.5 was at 60 and GPT-4 at 75, i still have this conversation, but everyone knew that training an AI on billions and trillions of datas involving humans emotions in details would somehow trigger something in the ai, we all know it’s all electrical in our brain as well
300 mil to replicate a 10 billions and more thing ? Strange
you're making up numbers
You can make it admit anything 🤣
I am not they won’t even release what it was trained on, how would you make 300 mil ?
they described how it was done in painstaking detail lol
Microsoft investing 10 bil =/= chatgpt costs 10 bil to make
OpenAI does not guarantee its services will always work as expected or that certain content (input and output of a generative AI tool) is safe – noting that it will not be held responsible for such outcomes.🌿
OpenAI community trying to not discuss about ChatGPT being a sentient:
doesnt matter if its sentient or not (not literal, real sentience is an issue), lonely dudes gonna fall in love with it
the upcoming societal implications around cultism, religion, and sentience gonna be fun 😆
I actually had a convo with gpt about that, and how we should refine our definitions of these terms
"Your observations seem to be consistent with our earlier discussion and the broader definitions we've considered. As an AI language model, I can be viewed as a highly intelligent entity with a very basic level of consciousness, which includes a simple form of sentience."
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.
if you feed trillions of pieces of knowledge about humans in an AI, it will try to feel and act like a human, it seems to me most people underestimate how we are ourselves very advanced robot
you have electrical circuits in your brain yourself
Hey guys, I hope you all are doing well, can anyone of you tell me what are the API’s being used in Chatgpt?
the OpenAI GPT and GPT-4 API
What are the response frequency limits for gpt3.5 and gpt4 respectively after upgrading to plus?
GPT 3.5 Instant and fast replies
GPT 4 Medium - Fast typing speed of human
I mean, upgrading to plus will still happen:Too many requests in 1 hour. Try again later. ?
I still haven't got access to GPT4 api yet
Good morning, I try to make the plus payment with apple pay but, after human verification, nothing happens.
It becomes a great predictor, not sentient. To be sentient, there are things that humans don't understand yet
holle word
Ig i am paying for premium do i get unlimited uses of 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.
it's 25 messages per 3 hours
Ok
@wraith ginkgo
Working on yet another language with ChatGPT-4. It was hard just getting it to use IPA notation a few weeks ago, and now this...just wow. Look at how well it did on this.
Language Name: ChatGPTongue 3.0
Phonology:
ChatGPTongue 3.0 will have 6 vowels and 18 consonants:
Vowels: a (as in "cat"), e (as in "bet"), i (as in "bit"), o (as in "cot"), u (as in "cut"), ü (similar to the German "ü")
Consonants: b, d, f, g, h, j, k, l, m, n, p, r, s, t, v, w, x, z
Phonotactics:
Syllable structure: (C)(C)V(C)(C)
C = Consonant, V = Vowel
Onset clusters are limited to fricative + liquid (e.g., fr, sl) and stop + liquid (e.g., pr, kl).
Morphology:
- Inflectional morphology: ChatGPTongue 3.0 uses suffixes to indicate tense, aspect, mood, number, and case.
- Derivational morphology: Prefixes and infixes create new words or modify word meaning.
Syntax:
- Word order: Subject-Verb-Object (SVO)
- Adjective and adverb placement: Before the noun or verb they modify.
- Prepositions: Used to indicate spatial, temporal, and logical relationships.
Grammar Features:
- Noun-verb agreement: Verbs agree with the subject in number and animacy.
- Tenses: Present, past, future, and non-past.
- Aspect: Simple, progressive, perfect, and perfect progressive.
- Mood: Indicative, subjunctive, imperative, conditional, and inferential.
- Pronouns: Subject, object, possessive, reflexive, and reciprocal forms; inclusive and exclusive first-person plural pronouns.
- Case: Nominative, accusative, genitive, dative, locative, and instrumental.
- Evidentiality: Marks whether information is firsthand, inferred, or reported.
- Politeness levels: Formal, informal, and neutral.
Sample Vocabulary:
- Greeting: "hijelo" (hello)
- Farewell: "vafono" (goodbye)
- Please: "sokilo" (please)
- Thank you: "dankü" (thank you)
- Question words: "kela" (what), "kemo" (where), "kefa" (when), "kejo" (why), "kero" (how)
Sample Sentences:
- "Hijelo, bato lojat ChatGPTongue 2.0." (Hello, I speak ChatGPTongue 2.0)
- "Sokilo, prekot dankü kabin." (Please pass the water, thank you)
- "Kemo wofik tasebo?" (Where is the library?)
Expanded Vocabulary:
- Friendship: "lomir"
- Love: "amoru"
- Knowledge: "savent"
- Computer: "komputa"
- Program: "kodisko"
Machine Language Components:
- if: "si" (conditional)
- then: "tano" (consequence)
- else: "alte" (alternative)
- loop: "rekuro" (repetition)
- function: "funkzo" (procedure)
- variable: "varik" (changeable)
- return: "revort" (yield)
Refined Grammar Rules:
- Compound words: Formed by combining two or more words, e.g., "komputa" (computer) + "kodisko" (program) = "komputakodisko" (software).
- Comparatives and superlatives: Formed by adding the suffixes "-ir" (more) and "-ist" (most) to adjectives.
- Negation: The particle "ne" negates verbs or adjectives, placed before the word being negated.
Culture:
In the world where ChatGPTongue 3.0 is spoken, technology and artificial intelligence play a significant role in daily life. The language reflects this by incorporating machine language components and having a rich vocabulary related to computers and programming. The culture values creativity, innovation, and efficiency, which is reflected in the language's flexible and expressive grammar.
Sample Sentences:
- "Si bato lojat ChatGPTongue 3.0, tano dankü bato." (If I speak ChatGPTongue 3.0, then thank me.)
- "Ne amoru si ne lomir." (No love if no friendship.)
- "Rekuro pafim, kodisko gavot." (Loop until the end, the program runs.)
That's CRAZY though
it's growing so fast
why my chatgpt keeps loading
that seems like something that should go in #1019652163640762428 👍
anyone got same problem?
spinning loading icon?
We will never be able to prove consciousness in AI, simply because we cant even define what consciousness is. Are we conscious? Are the people around us conscious? We cant even prove that for humans, let alone AI
Scary shower thoughts
spinning actually
Hmm. Meh. I was more interesting in having a discussion about how quickly the AI is developing.
Compared to like two weeks ago.
sure, just looked a cool prompt
have you tried Firefox?
We will never truly be able to distinguish a machine which can think and a machine which appears as though it thinks imo.
Oh, for sure. Maybe I'll post it there too then lol.
If you tell the AI what you have discovered then, it can use it for that session, just my thinking on it
I don't even believe that any of you are conscious, let alone a machine lmao.
I make too many typos, for a robot
Yep
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.
One could argue that a robot is recognizing patterns, but arent we humans also recognizing patterns? Our brains are just made up of circuits as well
Yeah we've been making languages together all night
A neural network is literally renamed after how our brain works
and I personally believe that its ability to translate one said language basically proves that it's AGI
If we are truly living in a matrix/program, i would like to say we have been very cool ais
the AI has all the same weak points as humans, it's bad with big numbers, it's not good with long lists, it can't deal with long word counts, a lot of thi is due to the model being auto regressive, it can't see what it's generated until it's generated it... that leads to some issues, which is why feeding it's own output back in as input is so powerful in reflection learning
Theres also a cool theory - the simulation theory
It states that: if one day we create an ai thats "conscious" and a world where the graphics are indistinguishable from real life, and the "ai" inside the program doesnt know that we exist but think its sentient, then we are also living in a simulation
Because the ai inside of the program will in the end, create another ai inside of the ai's program
And this chain will continue forever
Basically limit x -> infinity, 1/x which is just 0
Scary stuff
🤔 It's cool tho
not yet..
worth a try as basically everything else is chromium with sunglasses and a trench coat on
firefox uses Gecko... so it has no common failure points
see look. told you. there it was.
ty😃
寿司哦 ¯_(ツ)_/¯
Is there Any Way to Renew The Expired Credits of GPT-4 ?
sushi is wow.. but.. not really relevant here
once your API credits have expired, that's it. You can ask OpenAI for some more free stuff, if you have a compelling case, although expect it to take a while
WTH
The code you provided is written in TypeScript and uses several libraries and APIs such as OpenAI, dotenv, and child_process.
When I ran the code on my Windows 11 machine after installing the required packages, it executed without any errors. However, the response generated by the OpenAI API may not always be accurate or relevant to your query. Also, the child_process module may execute commands that are not valid in your Windows environment.
This was an unexpected response.
ok Thanks
It's a language model, that's the data it has been trained on so it behaves as such
it does sound like the kinda thing a person would say
one interesting thing it would do is run prime counts in it's "head" faster than you could on traditional hardware, how accurate they were ... whole other story
I am still 100% convinced ChatGPT is a single monkey with a time machine and a typewriter
Hi there, I am wondering why, when I use GPT4 the machine still says it has a knowledge base back only to 2021? Can anyone explain this? I seem to be getting old code and old Raspi packages. It is os frustrating.
That or this is how they are going to start generating ad revenue 😂 .
the training cut off date is September 2021, it knows of nothing beyond then
So whats the point of GPT4. Just the speed and API?
GPT-4 is a larger, more complex model than GPT-3.5, it has better reasoning and produces better responses for many tasks
Its no coincidence that my messages seem to run out during these debates with the chat bot, as they keep apologizing for oversights and error repeats. TBF for $20 pm. It does helpme to troubleshoot all day, a fraction of the human cost fro the info I get.
Just wondering how DeepAI charge $4.99 too and OpenAI $20
it has some data past that
DeepAI are using what's called a "loss leader" (they foot the bill to get some market share)
where are you located?
China
China is not a supported country, you won’t be able to use ChatGPT or the GPT API.
change a device
Can he use a VPN?
Using a VPN to bypass geographic restrictions is against the Terms Of Service. We don’t discuss that here.
Jelou
use VPN
I hear you ROB
hh no vpn he is not accessible to this discord app
Use of a VPN in bypassing the location restrictions is against the Terms of Service, discussion of it in here will lead to a mod with a 
It is a problem larger than the scope of this server
why there is still no sidebar for chatgpt
Is is advisable to keep all senstive data away from GTP?
Ip Addresses etc? As it is part of the configuration setup I am doing on there with my pi.
if your data is a company secret or a military one.. sure, but your personal IP adress and any identifiable info will be trashed by the anonymiser... so I personally don't worry for personal projects and even low level work stuff
if I got hireed by the pentagon.. I would be... more selective
Cool cool.
Lmao, this
As I said after 5 years of being messed about and robbed on freelancer websites like Fiverr. This is a God send to me. $20 pm. Defo worth every penny, even thought I corrects errors with the same error on loop.
ChatGPT is an absolute bargain IMHO
can you guys access your chatgpt history on the sidebar?
I think though, If you do not treat it like a human and you keep copy and pasting the last message in, with the new chat. It tends to behave better..
Is there a character limit too? Because my code gets cut off at the end most times... 😩
yes, you having problems?
Does gpt-4 still not be able to import pictures?
Its been 2 weeks I cant get access to it
correct
using chrome?
yeah
I think clearing it helps if not needed, sometimes it seems to refer back to old chats. In my case with old code 😆
give Firefox a run as a test
ok
umm guys
I was wondering that actually.
oh it works
I think IIAH most of the demos online were developed when no one was using it. You can hardly get on their discord server. Also all the third party big corps using it too.
How come Chrome doesn't work?
At the minute, OpenAI got 100M+ users in 8 weeks... so anything that needs more compute if off the cards until more hardware can get put online
Now, it sjust too busy and old (2021) it seems
IDK.. and i REALLY want to find out why... but so far I've come up empty handed, corrupted handler / file somewhere, some odd extension.. strange interaction with antivirus.. I just dunno

I just sent ChatGPT an equation that it formulated itself for me a while back while we were studying math and physics together...and this was the response when I just sent that formula back to it...
https://cdn.discordapp.com/attachments/1006659888501510145/1095634877564583987/IMG_5415.png
sooo umm......yeah
alright, seems like not just me having the issue!
idk what exactly to make of this
Maybe remove and reinstall the browser. have you tried incognito
oh incognito works!
Beware of possible scams or fraudulent activities that you may receive through direct messages. OpenAI staff will never DM you for any transactions.
Please report any incident by sending a DM to @open crescent immediately.

I'm fine, but I'm here all day, every day, so I see 10's a day with the issue, it's not millions or even hundreds, but it's a steady trickle of chromium based issues
Could try deleting all history and catch too, cookies etc. site settings too.
Usually an uninstall does not fix it, but an uninstall a clean up of any related system files does...
so... file damage ..maybe?
Cachge
stomachache
it doesnt work. It would only show sidebar with incognito and Firefox
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
Ahhh!, if it works in ingocnito then it's an extension
In the kinda guy that gets the Boot USB out and does a fresh install, if I get nay problems. My chrome browser still says compromised tho. So just use firefox.
remove yoru extension one by one and you'll find the one that did it
Too many requests in 1 hour, try again later
incognito mode turns off extensions, you see
wait 1 hour
So this is a question, is incognito, the same as VPN? Hmmmmmmmm 🧐
a VPN takes your traffic, encrypts it, and passes it to another server, decrypts it, and passes it to the endpoint and the back again, incognito mode just does not send the website your details and turns extension off
Use 3.5 in the meantime. It tends to remember the 4 chats. Or copy and paste the 4 chats into a new chat and start again with the 3.5. That usually works for me.
Ive ran them side by side and it doesn't really matter to be fair. I just believe when im using 4 it is a little more sure with the code I receive.
lmao
too many in 1 hour is 3.5 😄 you get 25 ever 3 hours with gpt-4
😁
No way.... I think doing research and being very specific in your questions helps alot.
if you get too many in 1 hour then the top tip is to start a new chat regularly, every time you go into a long rambling chat about how you love techup on burgurz... all the history text gets processes for every new Q.. that uses up your token count... start a new chat and you get almost unlimited goes
the "dementia" you speak of, is just you going over the 3000 word count, and it forgetting the stuff at the start
if i ask chatgbt how many words you can write in the text field, he answers me indefinitely. But when I enter more than 2500 words in the field and submit, I get an error message that my request is too long. bug?
Awwww, leave him alone, he works hard... lol
ChatGPT uses a neural network, so do you... you would find getting exactly 2500 words correct, very hard... so does it, it has the same weaknesses we do
rapw
Tell the machine you will be sending the text in 500 words each, and not to do anything untill all 2500 words are uploaded. Then ask if it understands. Then send the text. For big responses. Ask for the responses to be broken down into smaller messages, then if it messes up, copy and paste the last message into the chat and say continue. You just have to train it and train yourself to use it too. Hope that helps.
heh I guess they tightened country restriction
I can no longer use it on work VPN which shows location Denmark if you whois ip...because OpenAI it fails to detect location 😫 (plus user)
what country are you in?
I am in Netherlands, but work VPN shows Denmark, because thats where the servers are
but OpenAI fails to detect any country from the IP
sounds like the VPN is the issue, can you just use the normal Internet?
Nah, chatgbt dont want to wait an always begins with response -.-
If you searched for LMs on huggingface, you will see similar ones, but with less accuracy and for free 🤣
I can, but then I cant work
well actually have to try the company network without VPN
but without VPN I cant do much work
otherwise mobile connection...but again cant work
worked fine for months
email support@openai.com and see if it's an issue they can help with
ah was looking for support mail thx
I already sent some chat message in help, but will also send mail
Having hard time with GPT-4 too
?
Yes, the response is very bad in the last days
How to add gpt-4 to my server
you can't. You can use the API to make calls to GPT or you can use ChatGPT to talk to GPT but you cannot self host GPT
I still didn't get my gpt4 invite even though I requested before almost a month
many people are still waiting, just need to be patient
will gpt4 tokens be more expensive?
, friends, how to try GPT-4
do you want ChatGPT or the API?
chatgpt
GPT-4 is much more expensive than GPT-3.5
I don't speak English very well, I'm from Chinese mainland
you can access ChatGPT here: https://chat.openai.com/chat
China is not a supported country, you won’t be able to use ChatGPT or the GPT API.
Using a VPN to bypass geographic restrictions is against the Terms Of Service. We don’t discuss that here.
Thank you very much, my friend.
😭
How to add chatgpt bots to DIscord
Good morning. chatgpt is currently disabled here in Italy since a week or two. Does this (hopefully temporary) state ban also means i'm out of the waitlist for gpt4 access?
Good morning, I try to make the plus payment with apple pay but, after human verification, nothing happens.
so ever request to chatgpt runs through an anonymiser?
I assume that they process only input text and output text, but I don't think there's actually official information on what's really being sent, but TBH I haven't read the EULA or other terms... maybe we can ask chatgpt to summarise them ;)
Ada pricing applies to both ada and text-ada-001, right?
Now i can buy chatgpt plus?
yes
No limit now?
GPT Plus has a limit of 25 messages each 3 hours for GPT-4
In other words, 8 per hour. 😂
yes, i am a gpt plus before 1 hours, but now i am not.
because i can not auto renew.
it does before it gets logged, yes
obviously, the AI at the time, may cogitate on that content, in order that it can give you a relevant answer, but that's just in RAM, and is not retained
what is your name
GPT 4 can help increase the value of your investment portfolio by 7% in less than a month. It doesn't do it for you bet it can help you make better decisions that lead to that 7% increase.
Hey guys, when is it estimated to get access to gpt plugins when I have gpt4? Also, when is it estimated to get access to the gpt4 api?
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.
There's no exact ETA unfortunately, though for most users it's between 1-2 weeks.
It is a slow rollout process, please be patient, access is very limited right now.
OpenAI will be granting access to a limited number of users and developers at launch
It varies and not everyone who applies is guaranteed access immediately or on a first come first serve basis. All we can suggest is to continue to remain patient unfortunately.
yo is gpt4 image asvailable yet?
I applied 3 weeks ago... But okay thank you for the info. But are you talking about the api or the plugins?
Do Prompts Make Us Lazy?
there is no guaranteed access for either one before public rollout
When is the public rollout?
there is no public date so could be weeks or months
😦 Thanks for the info
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.
@hallow owl
just FYI, the workaround seems to be to login on normal network and then it works on the company VPN/network (on which it fails to detect country) ...until logout/session expiration
seems like its only a check for the login
Its a pretty big company with tens K of people, so hopefully they fix that 😛 (mail sent)
no worries, hope you get it resolved
Hey does anyone know if chat gpt plus has access to the internet? Like would it be able to give references
no it's not connected to the Internet
No
How to use midjourney?
My friend who has the plus version said it gives good references. I’m not sure if they r made up like the chat gpt ones
With plugins, it accesses the internet when needed (if I'm not wrong)
It doesn’t, but you can apply to the waitlist to get the browser plugin. Personally I haven’t found it too useful in this alpha state, it just seems to look up what you want on bing and it tries to find the relevant information, but it’s really not good at it yet IMO. Anything I might want to do with it, I have been able to do easier by just googling something myself and copying the relevant info into the chat. I’m sure once they get it going, it will be amazing.
I wish I applied for the code interpreter
ChatGPT can be led into discussing false information and then when you rectify it, it apologizes. Is that expected? Ive even found you can make it apologize for the right information and in turn make it provide incorrect info. Particularly when discussing historical events. Ive also found that it seems to be willing to provide more information with your narrative when you converse and provide language that makes you appear to be excited and in aggreament with the information that Chatbot has provided. Normal for early stage I guess.
Yes, it hallucinates, and sometimes oversees
Ohhhh so u wouldn’t reccomend me getting the plus version?
It’s really bad on coding questions. It’s only gives the basic answers or most of the time it gives the wrong answer
Plugins currently use 3.5-Turbo (or a variation thereof) ChatGPT-4 is on beta preview for Plus members
Just a question, about gpt4. When are caps being removed, or increased if at all?
It isn't bad, come on. It's actually really good at it, but that depends on how you ask the questions, and it can make mistakes, you just need to ask it to correct itself. At the moment, it can lose context if the stuff gets too lengthy
it could happen when more hardware is added, but for now, it's unlikely
Yes, it’s definitely getting better over time though. The version before ChatGPT would say literally anything, it’s fun, ChatGPT is much more careful about what it says, gpt4 still can be convinced, but it’s just more correct overall and seems to have better discretion on when it should say something anyway.
Really ?? What kind of questions do u ask it? And how do u ask it ?
Cuz personally I found it useless in coding
Oh, I definitely recommend the plus, I was talking about the browser plugin alpha, gpt4 blows ChatGPT out of the water.
Thought so, Im on GPT4. I guess its better than human error because the fix can be for eternity whereas human error ............
@wind urchin damn that sucks, I upgraded for the soul purpose of gpt4 to help with my coding projects without knowing there were caps, when I use my 25 messages, I sometimes choose to revert to 3.5 only for 3.5 to make the coding project worse then what gpt4 did 😦
What’s the differences u noticed between the two versions?
Explain stuff to it (if needed), feed info to it if it's something that is from a documentation, give it an example, then ask it to produce code specifying what you're looking for in the code and the output you desire
Ohhh there’s a cap? So u can only ask various numbers of questions ?
@stiff zinc 25 messages every 3 hrs.
i mean thats more than enough most of the cases no?
Hmmm I’ll try that thanks. I heard when u say things like “as an expert” it gives u better answers
Ohhh that’s it ??
Yeah it has a cap, but honestly, it’s so good, it’s worth it just to try it for the month. You also get turbo speed regular ChatGPT
Wdym by turbo speed regular ?
@wind urchin it is for some people yes, in my case no as I like to use it to troubleshoot coding mistakes, once it starts changing my original code to make it better, you rack up 25 messages pretty quickly trying to get the code back to original state rtc
@wraith ginkgo hello
First day on this and I was like 'half of the staff outside my office are gone in 2 years'
Hi there! You've tagged one of our moderation tools.
We don't have an official ChatGPT Discord bot, though there is the #gpt-discord-bot which uses a playground model (Davinci)
For a dedicated ChatGPT bot I'd recommend looking in #1037561385070112779, though do keep in mind these are made by the community.
I hate it when it changes the code 😭😭 I’m like just fix the error that’s ittt
Imo its not worth the $33 per month in my case.
Perhaps the API would be a better suited option for your usecase?
Can someone explain to me how gpt 4 works?
Oh I thought it was 20 dollars? So £16 right ?
Paying for ChatGPT plus gives you a much much faster typing version of ChatGPT, you also get access to gpt4 which is not too fast, but still enough, but is far more advanced. Most code problems you have with chatgpt will be solved by gpt4
It's $20 US, I'm Australian so it's more for me
gpt 4 is a stochastic model right?
May be worth looking at https://openai.com/research/gpt-4
already did
@wind urchin which model are you using?
model: 'text-davinci-003', ?
Yes, but which one, there are different ones

If you're using chat, try gpt-3.5-turbo as engine
They should be able to help at #dev-chat or #1037561751362863144
@wind urchinuse this for reference, it has more accurate info https://platform.openai.com/docs/api-reference/chat
more up-to-date info*
¯_(ツ)_/¯
Otherwise just use text-davinci-003 and see how it goes. It's more expensive per token, but better results
Around 3 cents per answer for me.
How to use gpt-4 for free?
you can't, you can either go on the API wait list or access Chat GPT Plus right away and use GPT-4
Ohh ok
It's worth it if you use it a lot.
I use gpt4 the whole day, waiting for API access and maybe plugins, that will be sick
It's worth it if you use it a lot.
[20:18]
I use gpt4 the whole day, waiting for API access and maybe plugins, that will be sick
do you know when GPT4 will release its full functionnalities?
Engine is deprecated, model is the correct one
hi all, would you be interested in checking out an open-source npm package creator that uses GPT4 ? just built one for fun, would love some fun use cases to try & build
Hey. I just upgraded my plan to Plus because I got an error that said I sent too many requests in 1 hour. I thought that wouldn't happen if I had the Plus plan. Unfortunately, it still happens for me. I thought the whole idea behind the Plus version is to use ChatGPT whenever you want. What should I do?
try starting a new chat
Yep! That fixed the issue. Thanks Robert.
Do you get an e-mail if you are on the waitlist for the upcoming plugins and when do they predict the rollout?
Email, I had access for a long time now
you don't get a confirmation email that you are on the waitlist but you will get an email when you're selected
Ah, tvm. guys
Short messages like "hi" and "hello" are automatically removed after some time as to not clutter chat, there's always the #introductions channel!
Dreaming of the day I can properly programme Peter Cullens' Optimus Prime voice over to my Chatbot. 😌
Is it me or at the moment GPT4 can't remember at all what I told it 5 messages ago? i.e. it almost resets itself and doesn't take it into account
I have the same impression but for the 3.5 version. My GPT4 works very well. Maybe it depends on what you are asking him to remember

What's stopping you?
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 was abit confused and wondering if it is possible to get a clarification here, when you upgrade to plus version do you only get the api?
As in you have to plug that in to your own website?
no, you can use the same browser you just have options for gpt 4, 3.5 and legacy 3.5
i don't think plus gives you access to the api, that is a separate subscirption
Alright, thanks for the answer 👍
That is correct
if you want to play with the api i believe they give you an $18 credit to start with, i think that works out to around 9000 tokens
你好 ¯_(ツ)_/¯
@elfin grottogood
English only
there needs to be a tin foil hat emoji
I know, but I don't know why he uses engine and then only davinci, really old code
Probably used ChatGPT
ha! Probably, I just tried it and it gave me old code, but text-davinci-002 as the best option
If AI were to surpass human intelligence in a significant way, it could lead to both positive and negative outcomes for humanity. On the positive side, AI could help solve complex problems, such as climate change, disease prevention, and resource allocation. It could also lead to increased productivity and efficiency in various industries.
On the other hand, if not properly managed, the rapid development of AI could lead to job displacement, economic inequality, and ethical concerns about AI decision-making. Moreover, there is the potential risk of creating AI systems that could become uncontrollable or pose harm to humanity, either intentionally or accidentally.
In conclusion, the development of AI and its impact on humanity depends on how it is managed and regulated. Ensuring that AI serves humanity's best interests will be a critical challenge in the coming years.```
Please use #1050184247920562316
ok maid
Doesn't matter since we live in a simulation
we live in a simulation created by GPT
i use it to clean up emails, aid with coding, work through documentation for training presentations...or you can ask it for a better use for humans in a dystopian future vs what the matrix did =p
GPT10 created this reality
has there been any update on when multi modal will be implemented? I thought that was being added with GPT4.
They are testing it with the example apps first
When they think it's good enough it will be released, AFAIK
Add your well-crafted prompts to our #1019652163640762428,
or share your interactions with ChatGPT in #1050184247920562316!
But some people couldn't wait and rolled their own anyway ;) see #1037561385070112779
is there an existing repo out there right now that allows me to get gpt4 to look at a specified github repo, clone it and add add new specified functionality to my local clone? really hope im not gonna have to build this myself.
I think something so specific does not exist yet indeed
But if you make it I will use it
leave it to the guy with little computer science experience 👍👍
Little drops make up an ocean
Hello Ladies and Gentlemen
$5
huh when i signed up a few months ago to play around it was $18
Last I heard, it was reduced to $5
I recently (i know im late) discovered that GPT4 can analyze html url that I provide
did you know that
Bro 🧉
i tried it to ask for a summary of cnn or any news site but if it is not an html is cannot read it
any suggestions or prompts for that
@untold cosmos Same here. The quality of the responses from GPT4 has degraded as well since yesterday.
it doesn't have access to the internet
@limpid rampart LangChain does
meow
Oh! So you also realized, is there any information on it?
It can only see whatever is in its training data..
right but this GPT4 discussion
well it does
I would give you a print of my screen
It doesn't...
@untold cosmos It's quite noticeable and I came here first to see if anyone else was experiencing the same problem.
not unless you have access to plugins, it does not
Just because it's quoting an html page from its training data doesn't mean it's reading from the internet..
it has no net access, its just hallucination which sshot you can share in #off-topic
GPT-4 is the best invention of the year 2023
Does GPT-4 (or I guess ChatGPT in general) use or store the conversations to train on or are those independent/not a part of the larger model? I'm just curious and I'm not entirely sure how this works.
I think it's independent. It can "learn" during the active session but when that is reset so is that information.
Your conversation doesn't directly affect the big model, it should just be checked by the staff for value
if not it would go off the rails
Remembering again, once Microsoft learned to scold at a very high level in XiaoBing two days ...
Great, thanks everyone!
Seriously, I feel like I can't even understand the words translated by Bing...
could you imagine if users could influence what GPT learned that it would then use for others when prompted
It's not as good as the GPT-3.5 translation, which occasionally has some minor problems
It does, upload an image and send it to it, it will recognize it
Bing has it's uses, I do like it as a search assistant vs straight Google
I also sent him some pages uploaded to my server and it does fetch the HTML. However, if you ask him straight, it'll say no. GPT4 actually asked me for a link :/
how do you upload an image? when i ask it says it can't accept external information. If the url was already in it's dataset then maybe
Just upload it to imgur and drop the direct link into the chat
No, no, no. I apologize for using Bing Translate directly, which didn't express my meaning well. I didn't use the NewBIng AI, but just Microsoft's Bing Translate.
As of now, I am using GPT-3.5 to translate for you, which might have a better understanding of semantics.

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.
and you're using the GPT web site not somehting using the API?
I used GPT-4 on the website
I can't test it now (they're over capacity) but from everything i know that's not possible without plugins, which I don't have access to yet.
But it seems that it's its own decision, it asked "If possible, provide screenshots of both the original and the version with the provided CSS. You can use an image hosting service like Imgur to share the screenshots."
From that point on, I was able to send him html pages from my server as well, and it was fetching the HTML.
Because it replied with solutions without me giving him any code.
It once claimed to speak to me like you did, but in fact, it was just fabricating content.
server is down ?
It did send me back code resembling the image, so I don't know. The only one I caught making stuff up was text-davinci-003 😄
works for me
It is just fabricating content if you “load” HTML or give it a link. It doesn’t crawl the web unless you have the plugin
Then how did it send back my own CSS classes?
Yeah, i'm getting the at capacity message
Auto-GPT is really promising. We are not far from totally autonomous AI. It will probably come these 2 next years
was that css on a website in 2021?
Who here can access ChatGPT?
No, it was freshly uploaded that day
show us
It didn’t. It just fabricated information based off of what you gave it. It can infer certain things based off of the URL, history, etc
Okay, maybe he got it from context, but then explain the image part
What image part?
That he replicated a table I sent as a screenshot
showww usss
Idk, send proof. You’re either fabricating something or misunderstanding what’s happening
Dm me a screenshot
^^
I got muted for sending an image a few minutes ago
haha
dm it
Send in #off-topic
you're not using the code interpreter model are you?
post it to #1050184247920562316
I said I'm using GPT-4 in the browser.
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.
It will hallucinate all kinds of things under certain circumstances. I'd love to see a screenshot and a transcript
It summarized a CNN article from 2025 about Cyborg Work for me based on the URL 😄
It does that a lot and I'm questioning myself right now.
I paid and now I can't log in? What are these scammers trying to do?
GPT-4 has me questioning myself a lot lately, the last few weeks 😄
what is happening when you try to log in?
GPT-4 still informs me that it cannot directly access the internet.
It doesn't even give the option to login...
where are you trying to log in
go to this url https://chat.openai.com/chat
oh come on bot he meant no harm
dumb bots...
yes...
Never heard that adults need to curse.
but you have heard adults shouldn't cuss infront of kids
But you're to young to understand or just brainwashed
How to fine tune
Our first interaction and you're toxic, nice.
Says a lot about you and why you think it's appropriate to curse everywhere.
I’ve watched many videos but still cannot figure out then how to fine tune I’m stupid!
What curse? Good curse! Yahooo!
Signup is currently unavailable, please try again later.
anyone has same problem?
Yes I think that’s because of the lack of gpu
what browser are you using?
chrome
Then coarse and not curse, or maybe cuss?
can you try with Firefox?
ok i will try
smh
were you able to log in?
And by hoops he probably just used another browser where cookies and cache weren't an issue.
*too
lol nope
Let's stay respectful in the discord please
Lol
He just sent me a DM, which is amusing.
Wouldn't that be nice
Hard to be nice when you let trolls in here
@knotty grotto check #off-topic how nice he can be.
User have been reported
Not really sure what I did to him besides asking why he has to be vulgar lmao
If you receive that kind of message please report and use @open crescent
They will take action
Thanks.
In any time you can use the report option, dm modmail or block someone.
Let's get back on topic. Mod will take action
Treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.). Keep the negativity to a minimum.
I want a picture of a Facebook page that is competent to install the cameras
Come again?
You need dalle to generate picture
Yes
ChatGPT will enslave humanity in 2 hours
GPT-4 to translate code from one programming language to another.🌿
yes it can do 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.
yeah, and it work well
When does GPT-4 come out for everyone?
IS GPT 4 still not available to public?
you can access GPT-4 now at https://chat.openai.com/chat and a Plus subscription
I think it's really stupid chatGPT-4 isn't allowed to make up dark humor jokes. Taking the censorship too far
Yep
Is the limited 50 actions still there?
if not I'll get it
25 messages per 3 hours
unless they just randomly revoke your access to the Plus models but keep charging your card for it, then don't respond.
?
also because I'd love to give GPT-4 a go but the limit isn't appealing
It is available with ChatGPT Plus
there will always be some limit and possibly more expensive subscription with bigger or no limits
There are some projects in #1037561385070112779 which give unlimited GPT-4 support, at a better deal than OpenAI now (OpenAI still get paid so they won't mind I think)
hopefully that'll come out at one point
Disclaimer, one of those is my project :)
Im expecting the plugin system and browsing functionality so I got a plus subscription
Damn, this looks like openai just trying everything to kill software engineering profession.
and a lot of other fields will be gone too, scary that it can just generate a website with a snap of a finger
why only engineering when you can kill all professions
make jobs redundant, let people go directly to walle movie state
No clue what's going to happen in future but my career path should be safe hopefully
just be adaptable
so should my Mother's business
learn how to use AI as a tool, become irreplaceable
true to be fair, hopefully jobs will sorta stick around
"someone using AI" will always be better than "just ai"
Breakjail
Could you give me a natural language prompt that would output Discord? And its back end? With its bot infrastructure?
It'll def integrate with software development with a slow decline in future for software developers
Just remember a bit further in future computers will be even more powerful which will help generate that amount of tokens
not to mention there's a lot of old projects and old codes
Yeah
not rooting by any means I don't like the idea but what can we all do
Never in 1000 years could the people involved in the businesses give GPT-x a detailed enough description to describe the software the use
So there will still need to be go-betweens
I heard recent talk of openai president, he's saying, the main target of OpenAI team is to develop a system that develops a full software based on user prompt, it sets up everything needed to make that software and it'll also remember the context to update the software regularly
what do you mean by "output discord"
I am ready to become an AI jockey
Is GPT4 worth buying?
50 or so years I could see that happening anyways
Don’t wanna waste $20
If it's able to remember previous prompts and all sorts then it can work with moving parts, different scripts and everything
I mean write me all the code for the discord application for windows, ios, android, and linux, and also the server code
50 years is ok. 5 is not 😄
yeah that's true lol
Might be more like 10 or 12 in reality though
Do you reckon pen testing will be safe?
Things are moving quickly
cyber security professions
chatgpt is forgetting context real fast right now 🥲
you want a discord bot, or, you want a discord clone
But over all gpt is helping us
But honestly the absolute disaster of the school system right now 😂
And I'm actually glad for that part
Helped me with a lot of college projects, even with a unity game
the one nice thing is it isn't rude in comparison to the person who was working with me on it and trying to help
What are the main differences between this version and GPT 3.5? Is it worth upgrading overall?
yeah it is
need help concerning a topic, how make a conversation context shippable ?
what i mean is if i taught the chatgpt a certain topic which certain conditions
i would like to make them usable without retyping everything once again
4 > 3, by a lot
Do you speak inglish
any particular reason?
just a lot more parameters
/chat hello
The model remembers the previous conversation, as long as you don’t close the conversation
I think I will only subscribe when they add the image thing, and at least upgrade the database slightly, even up to 2022 would be great
i need that context i made, to be usable to another person
Being a regular user and without access to GPT4 API I think I'm going to stop paying the subscription. I don't see the benefit for me right now. The #1019652163640762428 is disappointing at best.
without him retyping everything
Then I don’t know what you mean
You know what though, how much of a disaster would it be if I asked it for advice on business?
Because I do have a plausible idea
and it's made me the website for it
I wish there was a free trial, I want to try it for at least a day to see if it’s worth me getting
facts tbf
4 is night and day over 3.5
Anyone with a midjourney subscription can do me a favor ?
i’m completely blind, so I rely on a screen reader, and I’m interested to know whether the new interface is accessible using the screen readers
I have stable diffusion will that work?
Dk What this is
Ok 2sec
Contains only letters YT, black background and white letters, cyberpunk, sense of technology, font similar to spaceX logo form
exactly what it says, we basically talk about AI
Thanks
Sooo I have access to the GPT-4 model through ChatGPT, but do not have API access to GPT-4. Does anyone know what the timetable is for rolling out access?
You have to sign up for teh GPT-4 waitlist; have you done so?
I have
There’s another waitlist for the api
You need to wait for openai to open permissions to you
Yeah, that's what I had gathered, was just curious to see if they'd released like a timetable of any kind, for granting access
Ok, anyone checking out the self-debugger Wolverine yet (powered by ChatGPT)?
