community-help

[rugged wing]

nvm

[mossy crescent]

😂 all good dw!

[tawdry trench]

@rugged wing https://labs.openai.com/

[rugged wing]

thanks bro 😄

[tawdry trench]

How about this for random?

[mossy crescent]

hint: make a minion-based digital art of a door which is inspired from Gru but also depicts the concept of darth vader
Thats something random :)

[rugged wing]

ok true true

[mossy crescent]

I guess so 😂

[tawdry trench]

"in irl"

[rugged wing]

i spelled steve wrong

XD

[mossy crescent]

right there

[rugged wing]

why doesn't he write to the end of the thought?

[regal lynx]

Can we please keep this channel on topic

[mossy crescent]

Tell it to continue from the last line it wrote and quote it
it happens due to the issue of reaching max token response.

[rugged wing]

i dont spake that

[mossy crescent]

An example of how this could look like for you would be @rugged wing:

Continue from the line "quote line here"

of course, being in your native language, not english.

[vivid remnant]

I am able to ask chatGPT only 1 question per hour , how do I get to chat more frequently ? ( I was on vacation and have not been following the discussion )

[mossy crescent]

It could be due to multiple unknown factors, a lot of people encounter this issue, you can try the following steps as followed:

• Try clearing your browser cache
• Try using a different browser agent (like if you're using chrome, try opera gx etc..) / Try using a different device
• Try to use a different network wifi (i.e: mobile data)
  Let me know if any of those work for you!

[strange nova]

Also try giving feedback on messages as per #announcements

[strange nova]

?

[hollow basalt]

follow a conversation

[mossy crescent]

Could you translate it so we could understand more about what's happening here?

[hollow basalt]

it's possible?

[strange nova]

I have no idea what you are asking

[mossy crescent]

Nor me either, Could you explain the issue you're encountering in your native language @hollow basalt and ask ChatGPT to translate it to english please, thank you :)

[hollow basalt]

the robot answers a question for me, referring to the previous question

[mossy crescent]

Oh, That's because you're in the same chat history, if you don't want this to happen, create another chat instance by pressing + New Chat in the top left.

[hollow basalt]

Sorry, i do not speak english

[mossy crescent]

You can use ChatGPT to translate my words :)

[hollow basalt]

oh yes I want it to happen, but the robot said that this was not possible, and since I am new to this I ask

[gloomy egret]

but its down atm right?

[mossy crescent]

Could you expand more on what specifically you want it to reference towards?

[mossy crescent]

No? It's active right now and hasn't gone down since being open to beta from my knowledge

[hollow basalt]

How do I use chatGTP?

[mossy crescent]

You just ask questions depending on your questions/needs and it should respond accurately to the best it can.

[gloomy egret]

An error occurred. If this issue persists please contact us through our help center at help.openai.com. grr

[mossy crescent]

What I was meaning was in this image, why did you specifically point out 3 things, like could you expand on what you need help with that please?

[mossy crescent]

Try following the same steps I gave to this person here;
#community-help message

[gloomy egret]

thanks Arianna 😄

[mossy crescent]

No problem!

[spice briar]

Type continue

[mossy crescent]

It was already responded to above dw :P

[hollow basalt]

How do I activate chatGPT? Sorry if I'm breaking any laws here, I'm new

[mossy crescent]

You're not breaking any laws or rules here don't worry!
You just type in the box and press enter and it should respond, simple as.

[hollow basalt]

I mean here on discord, can the text be translated automatically?

[mossy crescent]

Unfortunately not, You will need to translate the text yourself, sorry!

[hollow basalt]

ok thanks for your patience

[mossy crescent]

It's completely fine don't worry! I understand the frustration and effort you're dealing with so take your time!

[hollow basalt]

Thanks a lot. It only caught my attention that the robot followed a conversation, and he himself had told me that he was not capable of doing so. But if you say that this is normal, then I remain calm.

[jovial cosmos]

Too many requests in 1 hour. Try again later. — How do I get more requests?

[strange nova]

You can try leave feedback as per #announcements

[mossy crescent]

it might happen time-to-time, mistakes happen but it can contain information based on its past chat history of only the past available chat in that current chat section

[mossy crescent]

Can also try the steps as followed to see if it resolves the issues:

• Try clearing your browser cache
• Try using a different browser agent (like if you're using chrome, try opera gx etc..) / Try using a different device
• Try to use a different network wifi (i.e: mobile data)

[hollow basalt]

thank you, and good luck to all, bye

[jovial cosmos]

If I subscribe to a payment plan, can I just use ChatGPT all I want?

[mossy crescent]

Just to cover the bases and see if any of those fix the issues you're encountering, although as @strange nova suggested above too, it's also good to provide the feedback too :)

[mossy crescent]

ChatGPT (and many OpenAI products) aren't under subscription-models, they're on a pay as you go model (like pay as you go sim cards where you pay for the minutes and texts and when you use up all of them, you top it up), they're under a token-base model where you pay for as much as you use.

and ChatGPT will likely follow suit in this model when it's no longer in open testing beta

You should be able to read more about it @ https://openai.com/api/pricing and get a general feel of the pricing for other models.
More information will become available in time regarding the amount it'll cost for ChatGPT.

Hopefully that makes more sense now @jovial cosmos and sorry it wasn't the exact answer you were looking for :(

[hybrid stream]

are they planning to monetize AI?

I've seen with image AI they have limited credits.

[mossy crescent]

They already have with the other projects OpenAI have, and will do the same with ChatGPT. (monetisation as in making money for usage from tokens)

Although when you talk about monetisation what specifically are you looking for?

[hybrid stream]

I mean I would like to keep using services for free 😄

[mossy crescent]

We all would but unfortunately it's not profitable for OpenAI to do as such, for example;

• Electricity costs to run an AI for a long-period without income profits
• Paying employees who work on the ChatGPT project

[hybrid stream]

how much do you think will be chatGPT?

and when they're planning to cut off free use?

[mossy crescent]

@hybrid stream It'll be under the token-system which is like a pay as you go sim card, you can find more info about that kind of system here: https://openai.com/api/pricing

Please do also note that the official pricing of tokens hasn't been publicly announced so the closest price-guess is in relation to the davinci-003 model.

[mossy crescent]

in relation to that, again, that information isn't public knowledge yet so no one knows when the open testing beta will close.

[hybrid stream]

thanks for the info.

Do you know how I can see how much of free credits I have used for chat ?

ahh its beta

[mossy crescent]

We have no information on how many free credits (or if credits will even be given to users at all) will be given to user accounts when ChatGPT releases into the token-model system.
If you mean during this current usage, you basically in a sense have unlimited currently but each message the AI responds with is capped at 4k tokens (which is around 300-400 words per message [including code blocks]).

[pale crypt]

good evening to all

[hybrid stream]

we will need to be careful in the future. with qs we ask 😄 otherwise can run to 0 quick

[rugged wing]

Good night

[pale crypt]

can someone tell me how to get my firefox browser 108.0.1 to work with unicode? he doesnt display unicode

[mossy crescent]

indeed we will!

[mossy crescent]

Uhmm that shouldn't be an issue on the latest version? If it is you can try asking the firefox development community or reporting it a bug to their site, but unicodes should be outputting correctly on that version.

[woeful prairie]

Hi

[mossy crescent]

but asking here won't do much help @pale crypt since this is the community server for OpenAI which has no affiliation to the company that owns Firefox.

[pale crypt]

i know that

but i try three days with the chatbot to create a programm and all of his code doesnt work in my program.

[bold scaffold]

Is it normal for the Bot to just stop mit sentence ? Like is there a character limit ?

[mossy crescent]

do you know why it doesn't work? Have you tried debugging the code?
And do you mean an extension for the firefox browser?

[mossy crescent]

Yes! it means you reached max token in the ai's response.

[knotty pivot]

Do you even know how to code yourself?

[mossy crescent]

You can tell the bot to continue by saying something like:

Continue typing from "line_quote"
(continue typing from "function test()"

This happens when you reach the 4k token response threshold (around 300-400 words)

[knotty pivot]

Have you gone through the basics of said language?

[bold scaffold]

Oh thanks is the command different in different languages ?

[mossy crescent]

There's no need to be like that, I've told them to debug the code they have, if they know what it means they'll know what to do :)

[knotty pivot]

Be like what?

[mossy crescent]

It'll be the same thing translated

[bold scaffold]

ok thanks

[mossy crescent]

The question you asked can sometimes come off a bit snarky or hostile depending on who reads it so it's best to just sometimes be patient and go through the steps with the person so they're able to learn to prevent this happening in the future.

[grave cedar]

Apologies if breaking rules for "inappropriate content" but I found OpenAI chat via a tiktok of a guy using it on tinder, wizz, etc. Seems like that capability is no longer approved?

[pale crypt]

i know it doesnt work. the apikey where stored in the wrong format. and so iam only able to call it with the header request with .TryWithoutValidation

[knotty pivot]

That's your understanding of it, but sure

[mossy crescent]

😂
Usage of ChatGPT to help coming up with lines for tinder isn't against policy so no it's not against the rules.

[grave cedar]

😂😂

[mossy crescent]

I've got to look that up though that is priceless c;

[grave cedar]

Can I share a tiktok link here?

[mossy crescent]

It'll be blocked and automatically removed, you can dm it to me though

[grave cedar]

Will do 👍

[mossy crescent]

If the api key is being stored in the wrong format, have you considered storing it in the format it should be in?

[pale crypt]

i think visual studio displays is at utf16 key@mossy crescent

[brave imp]

hey sorry if this has been answered a million times, but the chatgpt stops typing after a certain point and just pauses after its sending me a long message

[cerulean forum]

1.5 million.

[mossy crescent]

VS Code shouldn't be automatically displaying it as a utf16 key, it should be defined as how you defined it to be in your code for the extension.

[covert dirge]

Hi, does anyone know if you can buy credits with paypal instead of a card?

[mossy crescent]

#community-help message

[pale crypt]

not vs code vs studio 2022

[mossy crescent]

If the payment option doesn't show paypal then that'll be your answer.

[mossy crescent]

It'll apply the same no matter VS Code or VS Studio

[craggy prism]

What to do when getting this response on everything?

[pale crypt]

ah ok thnaks

[cerulean forum]

Can I pay with Shreddies box tops?

[knotty pivot]

reload

it's caused by cloudflare

[mossy crescent]

Refresh the page, if you have a purple icon it can be caused due to cloudflare cache being out-of-date.

[pale crypt]

@mossy crescent do you know if it is possibleto set the shell mode of chatgpt longer?

[mossy crescent]

There'll probably be a method on how to do it which bypasses it but it's very strongly recommended not to do so.

[pale crypt]

@mossy crescent i cant work if he puts me infos around that i dont need. and he instantly dont knows facts that i write minutes ago

[brave imp]

tysm! for reference how much would 4k tokens cost? in the future obv

[mossy crescent]

No information is public about that currently.

[mossy crescent]

Try using google to debug the issue you're having, I'm sure other developers have previously encountered it and there should be answers on sites like stackoverflow or something

[brave imp]

okay thanks

[mossy crescent]

also just to be sure @pale crypt but you are also making sure that if the extension is outputting a form of html, that such html has a meta charset value set to utf-8 right?

[craggy prism]

Aaah, perfect. Works now. Thanks.

[mossy crescent]

No problem!

[pale crypt]

@mossy crescent i switch to chrome. my englisch is today very bad. i will be sure to understand 100% what you say. i dont want to put hours of hours in a programm that cant propperly connect to an api shell.

[mossy crescent]

It's entirely fine you don't have to apologize about your ability to communicate in english!
In this extension, are you making some form of html-output (like a GUI or a user-interface) that the user interacts with?

[fast rune]

Hello

[mossy crescent]

if you are, you need to make sure you're correctly setting the properties and attributes inside the <head> of such ui-output.

[fast rune]

Can anyone tell me how it works?

[mossy crescent]

What specifically do you mean by "it" and what exactly are you looking for an explanation of, since that's a very broad question and can mean anything?

[fast rune]

The whole thing. The gpt the ai.
What is the idea?

[mossy crescent]

well that's a very very long explanation and I don't think my hands or fingers could type for that long but you can checkout the video on youtube called "The journey of OpenAI" from the channel "AIM Media Studios"

[fast rune]

Thanks
I will

[mossy crescent]

It's a little bit out of date since it was a year ago but it should address your questions :)

[mossy crescent]

No problem whatsoever :)

[brave imp]

uh its so annoying to keep saying coninute

[brave imp]

can we increase the threshold??

[mossy crescent]

Nope, the limiter is in place for a reason, It is a free service during its testing beta time after all.

[brave imp]

alright thanks!

[stuck escarp]

I would love to add quotes on the images, what is the prompt to do this?

[mossy crescent]

for which service?

[stuck escarp]

dall-e

[mossy crescent]

Could try getting it to have a "and write the word '<x>' over the top" or just using photoshop or something after you download the image and add it yourself.

[stuck escarp]

hahahah

[rugged wing]

what's dallify

[mossy crescent]

dallify? you mean Dall-E?

[pale crypt]

@mossy crescent At the moment I don't even know which of the 17 functions I've developed I want to integrate into my app. I just started programming again 4 days ago. I fell in love with c#. That's the best way I can express myself. only now I have this annoying unicode problem...

[stuck escarp]

the goal is to add the text with generative ai 🙂

[mossy crescent]

then the first option would be best haha

[rugged wing]

yes, there's a massive portion of the server committed to it

[mossy crescent]

Dall-E is an AI which is generating unique art from words, you can check it out @ https://openai.com/dall-e-2/

[mossy crescent]

Taking the time to go over a tutorial video or the C# documentation might help you out depending on which functions, classes and other related content you're using, best to understand the code you're writing than just getting it written and not understanding it.

[rugged wing]

that's cool, thanks for sharing

[distant zenith]

Is there a video or article smoeone can recommend that goes relatively deep explaining how chat gpt works? I'm especially curious where the data originates from, how it is acquired, and how long training takes

[pale crypt]

@mossy crescent i understand my code. but i dont understand the debugger and techniques from today. its over 20years passed that i coded with borland c++ builder 5 or something

[cerulean forum]

A neural net is a hair net with finer mesh. It captures neurons that leak out the ears. These neurons are then used for IoT devices, like toasters.

[mossy crescent]

You should be able to take the time to learn how to debug your code, there are tutorials online which can help you do as such.

[pale crypt]

ok

[obsidian pasture]

trying to get the API to do 1000 word article completions so we can port it in Python, has anyone been able to do this?

[stable leaf]

How reliable are the coding examples provided by ChatGPT?

[unreal hemlock]

[❓] — Generating texts

DALL·E 2 is not engineered to properly generate text. While it may get a few simple words right now and then, that is the exception rather than the rule. If you need to add text to your image, consider using an external image-editing software.

[mossy crescent]

They're very practice but should be tested and debugged depending on what is generated and what features you're using it for.

[regal lynx]

⬆️ generating text FAQ

[mossy crescent]

IM SORRY @regal lynx
I didn't know you were gonna do that sigh

[stable leaf]

Thanks

[regal lynx]

Dallify is a Discord bot that connects to the Dall-E 2 api

All good 😅. I'm just checking in

[odd dune]

Will ChatGPT ever get to the point where it would be more advantageous to talk to it alone about a topic instead of ask a room full of 5 scholars on that topic your question?

[mossy crescent]

Most likely yeah.

[obsidian pasture]

anyone know how to do 1000 word article completions in playground or ChatGPT?

[regal lynx]

It's impossible to know the long-term impacts AI will have on our society

[mossy crescent]

Just because it's impossible to know exactly doesn't mean we can't theorize about the possibilities in real-world applicable scenarios based on what AI's are currently able to do and structure it on a bit down time from how it currently is though :)

[regal lynx]

This is true! We can speculate 🙂

[odd dune]

I'm a musician, and I understand that a lot of learning comes from playing the wrong notes and discovering they sound good. I think that there is a huge amount of learning to be done when you are told information that turns out to be wrong. Will ChatGPT ever be capable of being wrong in a way a human is capable?

[mossy crescent]

That's why a lot of people fear AI at the current point because of people speculating its possibilities and practice usage back in the '50s and '60s

[cerulean forum]

An AI may generate cat videos, and all wires and optical fibres will become clogged with cat hair.

[mossy crescent]

Yes, misinformation can happen with ChatGPT and has already happened, not everything is perfect so it'll learn and correct itself :)

But in the future that'll change the more it's able to learn

[odd dune]

Is it ever gonna get more perfect?

[minor kite]

if you want to see ChatGPT be wrong just ask it how a card trick works

[mossy crescent]

The more it learns, indeed, but a system and an ai can never be 100% perfect

[minor kite]

not only will it come up with something that makes no sense, it will also insist that it is correct

[cerulean forum]

I may have many faults, but being wrong ain't one of them. -Jimmy Hoffa

[odd dune]

well thats good to hear. I like how people are so worried about AI because they trust it too much. I actually am optimistic about the level of skepticism around this technology

They're expecting a beacon of truth and they're just gonna get a Socrates instead

[minor kite]

i disagree. people are more likely to believe the ramblings of AI when they are not presented as AI

[cerulean forum]

All men are mortal. Socrates is a man. Therefore, Socrates drinks beer.

[mossy crescent]

That's why it's still in learning and not being publicly used in all industries.

[minor kite]

people are currently using ChatGPT to karma farm on reddit math and programming Q&A subreddits

the answers are plausible looking, so they get upvoted. they are also nonsense if you carefully read them

[odd dune]

"People are using tools to make low quality content on the internet" well fire the torpedos, this planet is ruined!

[minor kite]

people don't realise it is AI and get into arguments with it

[mossy crescent]

That might be some of the aspects of what some people are doing with ChatGPT but at the end of the day, this is a testing period of beta to see how this AI would manage a large influx of data processing and other factors, this AI and its systems behind it are being worked on day in and day out to be made to the best they can be.

[minor kite]

i just think it is funny that people aren't able to identify the AI and just think it's some idiot redditor 🙂

[mossy crescent]

and that's a common trait of humanity as we live in a society, we associate things as "other humans" to try to not be as lonely, we basically humanize most things without knowing it in a generation built around isolation.

like how you would call your car a name or something, it's a very common thing nowadays so.

[cerulean forum]

ONE is the loneliest number that you'll ever dooooo. TWO, can be as bad as one... - Three Dog Night (Aimee Mann covers it well)

[mossy crescent]

What about zero though?

Zero has nothing and no one. 👀

[cerulean forum]

Zero does not play well with others. /0 always generates a silly error.

[odd dune]

AI will be able to successfully "imitate" the lowest quality posts firsts. I think this will just improve the quality of humanposting

unconsciously as well which is even better

[minor kite]

it does not imitate low quality posts very well. it is too verbose and too polite for that

[wispy crystal]

Is it possible to add ChatGpt to discord servers?

[mossy crescent]

No.

[cerulean forum]

It will be a strange world where famous actors from the past will be simulated. You'll see Marlon Brando in a miniskit and Audrey Hepburn commanding a tank regiment.

[mossy crescent]

There's no public api to utilize so discord bots aren't able to pull such information from ChatGPT's resources. The API is currently internal-only during this beta.

[wispy crystal]

Ah yeah thats understandable

[odd dune]

I think that most people who are worried about this need to go watch some of those movies from the 40s. They're pretty good

then watch The Irishman by scorcese

Human intelligence has already been training on a language model for like a hundred thousand years, surely we can't get passed up by this dumb robot in a decade?

[mossy crescent]

Not that long but yes, as soon as the human species was capable of usage of language and communication (even without verbal), that in a sense is a form of a language model so

[cerulean forum]

Will AI poetry be published? Let's hope not.

[odd dune]

Why not? You afraid it's gonna write Leaves of Grass?

[mossy crescent]

People fear it'll replace the humanity of poets and authors but in reality that humanity in it only is represented from its language model to communicate the detail which an AI is very capable of doing such.

[odd dune]

Oh wow, so you're saying all of the humanness that AI generates was actually generated by humans?

[daring cipher]

Hi

[cerulean forum]

All problems can be traced to human error. - HAL 9000, "2001: A Space Odyssey"

[mossy crescent]

It's trained on our language model so in a sense yes.

[vernal viper]

I noticed that ChatGPT will stop giving an output randomly when giving a longer output. Any way to solve that?

[mossy crescent]

Reached max output message tokens, this happens, just tell it to continue and quote the last line it fully wrote out

[vernal viper]

Oh wow, I always just asked it to finish the output and it would regurgitate the same response and die at a different step. But doing it your way, it started right where it left off. Thanks for the tip!

[drowsy summit]

What are examples of software that exist to detect chatGTP?

[regal lynx]

I think there's a detector on Hugging face. I don't have the link as I haven't used it

[mossy crescent]

Nothing on a large-scale of knowledge currently from what I know...

[mossy crescent]

sorry but what the hell is Hugging face? 😂
I've never heard of that one before haha

[drowsy summit]

damn so if people use it for something like a dissertation its not 100% detectable

[mossy crescent]

There's a supposed detector out there but it could falsely flag anything so it's not reliable.

[drowsy summit]

only GPT-2 no?

[regal lynx]

I think so, but it may work with other models as well

[drowsy summit]

ill try now

[regal lynx]

It's a website that hosts loads of AI models
(Not from OpenAI)

[mossy crescent]

yikes I feel bad for the amount of servers being used for storage for it...

[drowsy summit]

i tried hugging face however isnt very reliable. if you just change some words it doesnt detect it

[cerulean forum]

Is it secure storage, using WOM (Write Only Memory)?

[vernal viper]

Is there a way to stop ChatGPT while it is outputting a response? Sometimes while it is giving a response I realize I need to refine my request. It would be nice to cut it off and ask again.

[mossy crescent]

Nope.

You can tell it to not output examples or something to shorten its load of output message.

[frail elbow]

anyone can help me? Application error: a client-side exception has occurred (see the browser console for more information).

[mossy crescent]

Try using a different browser than your current one.

[dry pendant]

Middle East ?

[mossy crescent]

Yes, The middle east is a real place, what about it?

[dry pendant]

When will chat Gpt open in the Middle East?

[mossy crescent]

There should be a page on OpenAI's site to go over where it's available, I'm trying to find it as we speak

[frail elbow]

in firefox works but,Is it possible to do something to make it work in chrome?

[dry pendant]

They said it is now closed in most countries in the Middle East

[mossy crescent]

I can't seem to find the link however it'll likely be updated wherever that page is but other then that we won't have information on regions it's available to.
Other then that we don't have information on when <x> country will have access.

[dry pendant]

Oh Ok Thank you

[frail elbow]

in firefox works but,Is it possible to do something to make it work in chrome?

[mossy crescent]

I know I saw, as to that, have you enabled javascript in your browser settings or disabled it at any point?

[somber maple]

Breh I got my open ai acc disabled for searching "Walter White In his meth lab"

[frail elbow]

how can i see that?

[mossy crescent]

chrome://settings then search for javascript

should be under a setting like Site Settings or something, not exactly sure the name since I don't often go there.

[odd dune]

Has ChatGPT been trained on languages other than English?

[mossy crescent]

yep

[frail elbow]

javascript Should it be on or off?

[mossy crescent]

on.

[frail elbow]

ok ty

[cerulean forum]

"It works better if you plug it in."

[fathom notch]

Is 1 request an hour the default?

[mossy crescent]

but what if I want to boil my kettle without putting my kettle plug in?

[mossy crescent]

Depends how many requests you've sent.

[fathom notch]

I just made an account and got locked for an hour XD, from 1 request

[mossy crescent]

Try clearing your browser cache.

[cerulean forum]

Put the kettle in a vacuum chamber.

But then the pump needs to be plugged in.

[mossy crescent]

but that's still plugging it in 😢

[frail elbow]

same error 😦

[mossy crescent]

Then for now just use the other browser since it's a temporary fix for now.

[frail elbow]

i dnt understand why

[grave drift]

With the new chat history feature, is there any way to easily access past conversations besides the left panel? Something like https://labs.openai.com/history is for Dall-e but for chatgpt

[mossy crescent]

It's just a chat history sidebar to the left side, nothing really much needs to change since this is a beta for testing public usage and data collection on the AI...

much can change in the coming future interface-wise.

[cerulean forum]

What does ChatGPT smell like? What kind of car does it drive? Does it wear nice hats?

[mossy crescent]

Smells like those polar bears you have!

[cerulean forum]

I picked a bad day to quit sniffing glue. - Airplane!

[exotic void]

As I work to understand an issue, I am seeking a team of experts who can brief me on the topic and help me to improve my comprehension of data structures. I am eager to learn more about this topic and am seeking the guidance and expertise of professionals who have experience in this area. I believe that by working with a team of experts, I will be able to more effectively understand and tackle the issue at hand, and I am committed to putting in the time and effort necessary to learn and getting up to speed in my understanding of data structures. Thank you for considering my request.

Anyone like this here, skip any application process?

[sly vapor]

Hi, I just joined the server, and was curious about the captcha-like verification method. Are those images currently unable to be solved by current OCR models?

[mossy crescent]

not the place to ask...

[mossy crescent]

I don't entirely remember this method so I can't be much help here apologies.

[exotic void]

Where is the channel for technical questions which the teams read?

Unless I can shoot them a direct email, I'm not really an 'application' guy.

[tropic delta]

Hey There, maybe someone can help me. I am trying to see if there is a way to use gpt3 to create facebook and craiglist post automically based on MLS listing to post in the marketplace.

[mossy crescent]

The OpenAI team doesn't often read the discord texts about this, this is a community-heavy discord.

[tropic delta]

any ideas?

[regal lynx]

#placeholder-channel if you have feedback or questions

[mossy crescent]

You can lookup and research most of this information over youtube and docs online which overlay such information about data structures.

[burnt ledge]

will you check me every time?

[mossy crescent]

confusion?

[burnt ledge]

I am here first time

[mossy crescent]

Do you mean on the discord? it's a one-time ask.

[burnt ledge]

yes and no

[heavy nexus]

Will the AI be updated with current day knowledge?

[mossy crescent]

No, It's been made so it can't intentionally.

[grand osprey]

If I decided to, let's say, write a story where two fictional characters battle, would the information on the characters and their powers it uses be accurate most of the time?

[mossy crescent]

If it's based on fiction, no.

[grand osprey]

How come?

[mossy crescent]

Since you've defined the parameter to be fiction so it'll make it up as it goes, it won't be true

[cerulean forum]

How many letters has ChatGPT written to Santa? "I have been a good boy. Please make me a train set."

[mossy crescent]

Don't you mean how many letters will ChatGPT be writing on santa's behalf

[cerulean forum]

Yaya, that may be inferred.

[grand osprey]

What if I was extremely specific about the exact character down to the universe they come from in the case of something like a marvel character and I specified to make the battle as accurate to the characters as possible, would it work?

[cerulean forum]

Er. No. Other way around. Written on the kid's behalf.

[mossy crescent]

depends heavily on how you phrase it to the AI... also marvel is still fiction so...

[smoky jay]

Is there an option to delete my OpenAI account?

[mossy crescent]

Yep! https://help.openai.com/en/articles/6378407-how-can-i-delete-my-account

[smoky jay]

Thanks

[cerulean forum]

You can check out anytime you like, but you can never leave. - The Eagles, "Hotel California"

[south mirage]

Is there any way I can add a Chat-GPT shortcut to my android homescreen ? (Chrome)
I could do it it Dall-E

[grand osprey]

Well I'm OK with some "I'll make crap up" type stuff as long as the info it uses about characters and personalities is at least mostly accurate. Like I'm fine with Aang winning against like a solar system level threat as long as the ai tries to explain it.

Writers do that all the time even within their own stories.

[mossy crescent]

I'm not sure if android has the same feature as iphones do to add sites to home-screens if you know what I mean, check if it does.

[grand osprey]

Like they'll have batman take down darkseid one comic and have him struggle against joker the next.

[smoky jay]

Will it take time until my account gets deleted when I sent an email or is it instantly?

[cerulean forum]

It does.

[mossy crescent]

finally not a quote has been said by you 😂

[mossy crescent]

It'll take a bit of time but once it's done, it's done, I'm sure you won't mind waiting a bit of time!

[scenic beacon]

/chat

[smoky jay]

Nope thats no problem. Just wanted to ask :D

[south mirage]

Done

[mossy crescent]

Chat has been opened
Sorry but this is not ChatGPT, You can access it here: https://chat.openai.com/chat

[smoky jay]

Thanks for the help!

[cerulean forum]

Chatting is forbidden. Online chat systems are only to be used to wait for Jesus, Elvis, or a bus.

[amber eagle]

Hi is this normal that the AI reply stop in middle of the sentence ? I refreshed like 4-5 times

[mossy crescent]

Yes, it's an intended function.

[amber eagle]

How can i continue the sentence please

[mossy crescent]

It means the response the AI has responded with has reached max token output, (300-400 words).

You can tell it to continue and quote the last full line it wrote.

[amber eagle]

Cool, thanks for@the advice

It works fine

[south mirage]

Does chat gpt have some sfw filter now?
I got some nsfw answers 5-6 months ago but not it says inappropriate request.

[mossy crescent]

It does yes, since not all sfw content is really sfw or appropriate to talk about to an AI.

[heavy nexus]

I meant if the devs will update it...

[mossy crescent]

There's no known public information if they will or won't.

[south mirage]

I had some really funny NSFW stories generated by chat gpt last year like "army of pigeons armed with D1ld0s" now, it won't generate such stuff

[heavy nexus]

Darn.

[cerulean forum]

Perhaps you should go read about neural nets, and what training entails. It is not a program that can easily be updated.

[mossy crescent]

Well now it's in testing for public-usage under its beta and misuse if a filter didn't exist would be absurdly high and is available for minors to access and chat with too so exposing them to such details would not be appropriate.

[heavy nexus]

Why would you make a program that can't be updated easily? Is it for hackers or something?

[cerulean forum]

You do not understand the concepts. Go read the sides of cereal boxes for enlightenment.

[south mirage]

Is there a way to run Chat-GPT locally?
Like I can run stable diffusion locally.

I get the censorship, but I really like the raw and unbiased Chat-GPT

[heavy nexus]

This is the #community-help channel dimmit!

[mossy crescent]

No, do you know how dumb it would be to have a local version of preparatory development AI?
It would leave so much funding from investors and so much more other things to lead to reverse engineering which is not what OpenAI wants...

The best thing you could probably do is just try to run a model yourself of the text-davinci-003 model and train it to do what you want it to do but it can be pricey. (api usage)

[icy maple]

What does the nested structure of historical chats indicate? I assume it means the nested chats were spawned from the parent, but is there anything else? Anything with respect to the concept tokens? Perhaps they are reset or something?

[mossy crescent]

Not really from what I can tell at least...

[south mirage]

Okay, no nsfw is okay,
But i remember asking a prompt last year
"Write a story about a raccoon who caused WW3"
The AI made the start quite funny , then got serious, then got dark and realistic, and the ending was gloomy and not really a good one",

I prompted the same keywords today and the result was different,
It started the same, funny enough, then got serious, but the ending was one of those happy endings in a Disney movie,
I thought maybe a different response,
So I gave a different type of prompt,

"write a story a about a pigeon who found a small pigeon sized AR-15 in the street"
Write a story about a NY chef who became a mafia boss
Write a story about a dog who decided to become a drug lord

The pattern that I see now is that they all have same sort of Disney type happy ending

My question is, has something changed? Or is it due to the same nsfw filter

[south mirage]

I get it, ...I mean I already knew it, just wanted to hear ppl opinion on it really

[icy maple]

When creating fiction, the bot does try to go very polyanna-ish with endings and characters unless you give it explicit instructions not to.

[mossy crescent]

polyanna?

[south mirage]

It was much creative and open minded last year, now it seems rigid and planned

[south mirage]

I know Ai can't be creative, just combinations, but it seems to be following some pattern now

[cerulean forum]

You must issue the DWIM (Do What I Mean) command. Your mind will then be scanned and the appropriate response issued.

[icy maple]

"Happy happy ending, everyone gets along; despite our differences, we appreciate each other differences and learn and grow from each other." That type of thing

[south mirage]

Yes

[south mirage]

I'll try

[mossy crescent]

So what if we do a DWIM command with you ChiaPet?

[south mirage]

You Messing with me?

[cerulean forum]

It will not work. I will not share my Cheetos.

[mossy crescent]

damn it!

[mossy crescent]

ChiaPet's known as a kinda-troll but hilarious c;

[south mirage]

Upload his consciousness to Chat-GPT as a base

[mossy crescent]

That is entirely incorrect... What?
an AI can be creative lmfao

[marsh tiger]

Why does "Continue" no longer work correctly? It keeps regenerating instead of continuing responses.

[mossy crescent]

Quote the line you want it to continue after

[cerulean forum]

Gravity and Brownian motion can be creative in how the tea leaves settle at the bottom of a cup.

[south mirage]

Creativity maybe a philosophical concept, idk how it applies or means in coding or machine learning or other stuff

[vital schooner]

hey guys, i just have some privacy concerns, why does chatGPT absolutely NEED my phone number, you can verify people in other ways and also, are our questions saved in a log that can be accessed by humans in some exceptional circumstances? cheers

[mossy crescent]

Creativity and everything else that a human is "capable" of thinking of is a concept which is very capable of being reproduced by an AI.

[cerulean forum]

No. Not reproduced. Simulated badly.

[mossy crescent]

Our brains are basically like an Ai in a way, they're programmed in a certain way, just how an AI is based to learn a language model (based from humans).

[south mirage]

No filters on us

[mossy crescent]

And the raw ai that is behind ChatGPT has no filter either

The developers put a filter in place to prevent misuse of it.

[south mirage]

What are your thoughts on the decision to put filters so early,
I mean any technology humans create does get misused eventually.

[mossy crescent]

I'm entirely in support of it due to the percentage of improper usage would very much out-scale the percentage of good it would do for people.
Think about market manipulation given value properties?
Think about illegal activities it can and would tell people how to do?
Think about the amount of content it could (and would) expose minors towards?

[craggy ravine]

do you guys have some prompts you use for using it as a tool while in lectures?

Mine is:

"Explain [topic] in a simple terms, write 3 bullet points about it and show me an implication of it in real life. write 2 multiple choice question and one fill-in-the-blank question about it. Then write 4 topics that could work alongside this term.

[south mirage]

Good answer

[fickle wadi]

hi, i am using chat GPT to write a high school analysis about a poem. but the thing is that im afraid it will be detected by plagiarism, is this something i need to worry about? and is there a good checker that is trusted?

[mossy crescent]

Could simplify that to this

Explain [topic] as simply as you can, 3 bullet points about it, implication in real-life, then write 2 multi-choice questions and a blank question about it. as well as 4 topics that could work in correlation to this topic.

[mossy crescent]

You can argue it no matter what since it's very capable of flagging work which is written by humans to a high-level.

[craggy ravine]

Nice. do you have any other ideas on how to extract information about it? I mainly do psychology and R with it

[mossy crescent]

that's likely the best way to do so since it gets right into the point and shows why

especially in psychology

[cerulean forum]

"I think that child psychology would work if we could just get the child to understand it."

[sudden galleon]

Is it a bug that the respond of the A.I cuts in the middle of the sentence or is the respond text limited to a max value of chars?

[mossy crescent]

No, intentional.

[south mirage]

Y'all think Russia is doing anything in the AI race ? They pretty much left out of all the races with murica after space race and nuclear race

[mossy crescent]

It's due to reaching max tokens it can respond with.

[sudden galleon]

ah ok thanks for the info

[icy maple]

In order to avoid that, I asked the bot how many tokens had been used in the current conversation and it was unable to tell me.

[mossy crescent]

since that information isn't public info exactly, it's around 4k tokens (which is around 300-400 words)

[icy maple]

Yeah, I knew the limit. I was just wondering how much I'd used in the current conversation. That way I could prioritize the questions I wanted to ask.

[mossy crescent]

you can ask as many questions as you want, it just limits it per message so if it exceeds <x> amount, it'll cut off and you have to ask it to continue from "quoted line", no need to prioritize anything.

[cerulean forum]

If Princeton had chosen to hire out Einstein's brain, would Einstein know or care what it cost? No. Only the administrators of the service would know.

[craggy ravine]

how do you guys share outputs? @mossy crescent

[icy maple]

Oh! Ok, that's good to know!

[mossy crescent]

outputs? you mean feedback?

[craggy ravine]

no i mean the ai's answers

[mossy crescent]

You can just copy and paste it 😂

[craggy ravine]

its too long

[chilly kiln]

Any reason why GPT is returning some code but then freezing and giving a incomplete example?

[mossy crescent]

Google docs comes in handy there

[mossy crescent]

Look up a bit, I just explained it sigh

[craggy ravine]

ill open a thread

Refining Academy Output help

[delicate mirage]

Hello!
Did someone develop any vocal synthesis to read chatGPT answers?

[mossy crescent]

likely somewhere out there if it's done already, good place to check is github.

[cerulean forum]

Yes, someone has done that, and I'm not going to find the link for you.

[exotic void]

Now that yall known me better can I post gifs? It'll really save me some time when typing texts.

[mossy crescent]

I'd suggest not...

[delicate summit]

why is the AI struggling with some easy multiplications sometimes?

[hexed elk]

bonus points if someone trains it to speak in the same voice as used in the 1983 movie war games.

[mossy crescent]

depends what you're asking it... it shouldn't be...

[cerulean forum]

It is a LANGUAGE AI, not a math one.

[strange nova]

It's a language model, not a math model. So it doesent do actual math calculations, but tries to infer based on the text given

[south mirage]

This prompt working, but the one I'll send after (slowmode) won't accept

[exotic void]

It's learning math chilllllllllllll

It can do some coding, next versions will be awesome.

[mossy crescent]

reload page

[mossy crescent]

Teaching a language model how to math 101 😂 that idea is so dumb it made me laugh haha

[cerulean forum]

An AI is not an AGI. An AI is monumentally stupid.

[exotic void]

It's not so farfetched, train me with your ideas.

Rather than an ai model

Where's ya Jupyter note repository

[mossy crescent]

I think I'll pass on that offer, too much of a time waste imo when I'm doing other things. sorry.

[exotic void]

: (

[exotic void]

A B, B A.

xd

[regal lynx]

ChatGPT won't learn any more math than it knows unless the developers retrain the model. We can't permanently alter it

[south mirage]

Tell me one thing Devs, you feed dialogues from movies related to AI to this model?

[mossy crescent]

you mean movie scripts?

[exotic void]

It'll learn itself then you'll be blessed with knowledge of some linked-phenomena nature and 'claim' that it was your idea lolol Then you'll be able to use it better.

[south mirage]

Something along the line

https://tenor.com/view/coincidence-i-think-not-coincidence-i-think-not-the-incredibles-gif-5096707

[mossy crescent]

Just ask for a script of a movie or something and if it outputs it and is accurate then yes... it's fed on mostly public-available knowledge from 2021 and before.

[regal lynx]

I'm not sure if you understand how the model works. The model is stagnant and cannot learn anything new unless the developers retrain it. The model can't learn from user inputs or from itself

[mossy crescent]

to add onto this; It can only pull on that local information whilst that chat history is open and active, once it's gone it's gone.

Look at you go @regal lynx! Best lead web dev!

[exotic void]

If you're using such a nice datalake for discovery i'm sure we can teach it somehow

[mossy crescent]

"we" can not, only the developers behind the language model can sigh

[south mirage]

In short this model can only combine words and phrases and combos that have been feeded to it, that it cannot make words and combinations out of itself "creatively", right?

[exotic void]

Especially when you get a live model and connect it to all social media feeds, when do I get this? just put a source-plug to the internet and openai collects.

[mossy crescent]

@exotic void Just look at what we've said above, we've given clear clarification on what it can and can't do...

[regal lynx]

It can create new things in a sense that it can draw from its knowledge to make a new story. However. There are limitations as to what it can create

[modern monolith]

Are companies like Intercom, Crisp.chat, Notion and similar companies that are launching AI-features these days making this available through ChatGPT, their own AI or?

[exotic void]

Somebody waiting for a user to test these theories in a chatprompt may offer these theories.

[static tree]

how can i creat an excel data with chatgtp?

[mossy crescent]

who knows... depends if they want such information public or not...

[modern monolith]

Hmm

[south mirage]

How similar is this model to stable diffusion? I mean that model too makes things out of its own knowledge, just visually,

[mossy crescent]

It's best compared to the model it's based on, text-davinci-003.

[regal lynx]

I don't know how SD works or the source code for ChatGPT. I assume that the model roughly follows the same concepts that Dall-E 2 does though where it creates "new" from what it's seen

[rugged wing]

did you just-... nvm

[mossy crescent]

Why would ask that question... in the OPENAI discord 😂

[regal lynx]

Talking about other AIs should be in #ai-discussions

[civic swallow]

how can i rephrase it to work in chatgpt?

[mossy crescent]

just tell it to process the text *paste text* and then tell it to rephrase it

[civic swallow]

I tried but then this shows up

[mossy crescent]

Tell it to rephrase what you said or just tell it this:
"I want you to reword this chunk of text:
<quote_here>"

[civic swallow]

ohhhh

[cerulean forum]

"He who quotes the learned sages will be known far and wide as a smartass."

[odd dune]

You got me

[cerulean forum]

I would like ChatGPT to supply answers on my Etch-A-Sketch. Has anyone heard of development in this area?

[mossy crescent]

if only... if only...

[cerulean forum]

Wooooo. That be a fine T. Nile song, actually her only good one. Simple computer graphics in the background.

Zimple yet effective.

[vernal viper]

When writing code, how can you have it output just the updated part of the code instead of the entire thing?

[cerulean forum]

I am aware that a modern robot can operate that toy, or weld it to a spaceship.

[odd dune]

Will ChatGPT unironically be able to do taxes and accounting in the future?

[icy maple]

Is it possible to have Chat GPT analyze a document without having to copy and paste the contents into the prompt?

[mossy crescent]

Just as you said... tell it what to do and what to output lol

[cerulean forum]

Would you trust a goat with financial matters?

[mossy crescent]

who says it can't right now if you give it the correct variables? 👀

[mossy crescent]

No since you can't upload files.

[mossy crescent]

God damn right I would!

Better than most people tbh... 😂

[odd dune]

damn, if only I made taxable income, then GPT could come in clutch

[cerulean forum]

I bet Chuck Norris can upload files. Bet he's never heard of OpenAI/ChatGPT either.

[mossy crescent]

where the damn hell do you come up with these @cerulean forum 😂

[cerulean forum]

I harvest them from my bellybutton.

[odd dune]

Yeah I actually would, I think that GPT has certainly surpassed humans in accounting and number crunching. Now is it a good idea to game-theory-optimally file taxes is another question lol

[cerulean forum]

You people are weird. Probably like polka dancing.

[solemn marlin]

I'm trying to make ChatGPT sort this table properly, removing unnecessary spaces from "Primeiro nome" (first name in Portuguese) and "sobrenome" (surname), then printing everything.
It's not working until now. Does someone have an idea?

Here's the table:

[solemn marlin]

Primeiro Nome;Sobrenomes;Peso (kg);Altura (m)
felipe;neves da hora;70,6;1,69
sandro ;barbosa;90,5;1,73
maycon;da silva ferreira;100,01;1,82
jeferson ; Oliveira;80,04;1,75
manuel;gomes vieira;63,05;1,8
emanuel; da silva;87,5;1,61
isis; santanna de Deus;59;1,6
maria;santana bezerra ;63,8;1,75
Alessandro ; Viana ;69,53;1,83
sandro;barbosa;75;1,54
Almir; Valente ;87,86;1,8
Bruno;Santana Filho;68,63;1,65
marta;santos oliveira ;63,55;1,72
michele;Santanna Da Silva;68;1,7
Alex;Da Silva ;78,08;1,85
Daniel; Alexandre ;54,05;1,78
Manuella; Pereira;68,36;1,69
Matheus; Silva;92,3;2,01
Giuderlan; Antonio Neto;80,25;1,98
Geisa; Ferreira Da Silva;75,05;1,65
neymar ;josefino andrade;110;1,77
João;Michelango;160;1,59
Gabriela ; Xavier Ferreira;58;1,7
Michelangelo;de La Siena;100;1,9
Ivanor;de Jesus ;75,68;1,7
Cienne ;Augusta Conceição;56,05;1,57
Arthur ; Santana;66,06;1,75
amaro; Luiz;90,05;2,68
Alzenrau; Silva;65;1,97
caio;sena;50,01;1,33
Romildo; da Silva;89,05;1,73
Lindomar;Catillo;68,22;1,53
Felizberto;Felizmino da Felicidade ;86,5;1,86
Caroline;da boa Mente;78,05;1,87
Bruna;Giuseppe;75,1;1,78
Carlos;Magno;101,9;1,69
Adailton ; Ferreira;55,98;1,6
Jandersson;Almeida Santos;105,35;1,99
Joao; Joaquim José da Silva;98,05;1,89
Allan;Santana;76,12;1,85
Danielle; Santos;55,09;1,62
Valdevino; Soares;60,2;1,68
Maria; Silva pereira;65,87;2,56
Ivanorvilson;Alberto Lima;58;1,53
Camelia ; Joselina ;96,05;1,23
medinchina;casais;120,05;1,93
Alcebíades;Augustino da Assunção;86;1,8
Masha ; Alcântara;89;1,89
Livia; Duarte ;54,06;1,45
Romerson;Simpson Silva;90,11;1,75
Mirelle ;da Fonseca Filha;59,9;1,62
Sregio; carvalho;87,05;2,98
Filadélfio; Alfrônio;98,2;1,95
Marcos; Silva;70,14;1,7
Afreim;isotopo;78,89;1,51
Agatha; barreto Da Silva;54;1,68
Emmanoel;Junior;73;1,55
Eliza ; Trindade;68,07;1,67
Aminoã; gabriela ferraz;68;1,95

[mossy crescent]

I'd suggest deleting that, Giving out ray id's isn't very suggested upon...

[marsh tiger]

Anyone else get a 1016 error?

[mighty cobalt]

yeah wait till they fix their domain

[mossy crescent]

Yep! Everyone is currently.

[marsh tiger]

Whew

[mossy crescent]

It's likely an update to internal-tools for the ChatGPT site which is undergoing right now.

[marsh tiger]

Man it's performing WAY better today, way less dead sessions

[mossy crescent]

Give them time!

Should be back up now! ^

[wanton vine]

Yea!

[marsh tiger]

Seems like it has a longer memory too, it's incorporating previous info much more coherently. Or maybe I'm just getting better at prompts? 🤷‍♂️

[wanton vine]

It's back.

[mossy crescent]

or it was your chat history...

[marsh tiger]

Maybe. It feels like it's handling new sessions better too though

[lunar jungle]

Is there already a support for problems?

[mossy crescent]

huh? expand more on what you mean please c;

[marsh tiger]

I'm here if you need someone to talk to buddy

[cerulean forum]

I surmise that error s generated by a new filter which suspects harm to squirrels.

[marsh tiger]

Good, squirrels suck #rabbitcrew

[lunar jungle]

Ups sorry, didn't see that it is a different tag. My father has created an account, but he won't get a verification email. Would be nice to know what to do.
We already tried to send it via desktop pc and mobile phone, but the verification email is not getting send.

[mossy crescent]

Did they enter the email correctly? Please make sure they did.

[lunar jungle]

yes

[mossy crescent]

There should be another option to make an account which should be under a google logo, try signing-in that way and see if that works for you guys, that way you'll only have to verify through sms and not email.

[lunar jungle]

well, he already created an account and is stuck on the "Verify your email" site. So he needs to create a new account with his google acc?

[cerulean forum]

There may also be an option to log out of the planet.

[mossy crescent]

He should be able to just use the same email (if it's through gmail) through that option I believe.

[mighty cobalt]

y does the code js cut off sometimes

[mossy crescent]

message output max tokens reached.

Tell it to continue from "quote the last full-line it wrote"

[cerulean forum]

Code is cut off because ChatGPT is inept with a machete.

ChatGPT is not good with tools. It cannot fix the sink.

[mossy crescent]

at least... not yet...

[mighty cobalt]

I'm just stick with my own coding skills. Rather have something that works 100% then to rely on a API that can code stuff wrong.

[cerulean forum]

Can an industrial robot fit in the average bathroom? Maybe. Let's get government grants to research it.

[marsh tiger]

Okay, I'll try that. Thanks

[mossy crescent]

that was like an ago ago 😭

[cerulean forum]

But ChatGPT controlling such a robot would be a horror. It would be dangerous.

[rugged wing]

It would be really helpful if the AItool can also access academic journals and read specific papers

[mossy crescent]

ai*

[mossy crescent]

It has no internet access...

How do you expect it to do as such if it has no internet access...?

It only has access to data that is in its language model

[rugged wing]

I don't know how its developed haha..

[lunar jungle]

It worked with the google acc - thank you for the help

[rugged wing]

However, it would be extremely useful if it had access to papers

[mossy crescent]

https://help.openai.com/en/articles/6783457-chatgpt-faq

[rugged wing]

thankyou

[wheat goblet]

Just wondering, how long is the preview of ChatGPT for and what happens after that?

[mossy crescent]

Not public information.

[rugged wing]

When will this be updated?

[mossy crescent]

and what happens after it, it will likely turn to its token-payment model system which is known throughout other projects by OpenAI

[wheat goblet]

Fair enough 🙂

[mossy crescent]

It's a pay as you go model (like the pay as you go sim) :)

[rugged wing]

will it ever have access to the internet?

[mossy crescent]

Who knows...

[wheat goblet]

I imagine the infrastructure behind this must be immense

[mossy crescent]

Maybe it will, maybe it won't, we're not in the future to know if it will or won't...

[cerulean forum]

I imagine that frogs are tasty.

[mossy crescent]

I'm pretty sure they just have the same program through multiple server-clusters which sends requests to a main ai and that would send out responses to each mini-server cluster which handles user input if that makes sense to you, just summarizing it in a nutshell.

[exotic venture]

is the chatgpt only search things on google?

[mossy crescent]

It can't search things on the internet, it's been taught a language model and can only retrieve information based on that model.

[rugged wing]

do you know when the information will be updated

[exotic venture]

and is this modal only taken from google?

[mossy crescent]

No?
It's from research and other content from OpenAI's research from 2021 and years prior...

[cerulean forum]

ChatGPT can only search a database of what penguins have drawn in sand.

[mossy crescent]

No public information, think about how negative of an impact such information updates would have on a large-scale...

[rugged wing]

why would that be a negative impact

[jolly gyro]

Who owns open?

Openai*

[cerulean forum]

Have you considered reading their website?

[mossy crescent]

Think about the current state of the world, the state of economics, the state of the markets, the amount of stock & market manipulation that can happen, the amount of misinformation and negative outcomes that could possibly happen with updated information...

There's a reason they cut it off at 2021.

[marsh tiger]

This made me wonder if you're a bot

[mossy crescent]

Yes they are.

[cerulean forum]

What? <clank> <clank> <whirrrrrr> <spin> <blink lights> <clank>

[marsh tiger]

Ha! Not a single beep or boop, you can't possibly be a bot

[topaz gull]

New user here! I'd love some tips on using chatgpt for copy writing, specifically for science-based products/services Thx!!

[rugged wing]

my ipad cant access chat.openai.com, all my other devices can.

[mossy crescent]

Clear your ipads browser cache.

[rugged wing]

still not working

[mossy crescent]

Did you clear the cache not the history?

[rugged wing]

both

[mossy crescent]

Try using an incognito tab and see if it works.

since that seems to be a cloudflare issue, not an OpenAI one unfortunately.

Result: I'm not ChatGPT, Try asking ChatGPT.

🤣

[rugged wing]

Sorry i don't know how use

[mossy crescent]

https://chat.openai.com/chat

[rugged wing]

This my first time

[grizzled wasp]

any clue why my responses keep getting cut off in the middle? or is there a known way to get chatgpt to continue a previous response? i've had varying success.

[mossy crescent]

Not a bug or issue, it's intentional, it's due to reaching max token in the response given, tell it to continue and quote the last line written.

It's around 300-400 words limit per response. (4k tokens)

[grizzled wasp]

what is the best way to tell it to continue and quote the last line written?

[mossy crescent]

Continue from the line "<quote_line>"

[grizzled wasp]

ive been using it to give me example SQL statements, and asking for it to continue seems to result in a different sql output all together

but i haven't been as specific as you've just recommended arianna

[mossy crescent]

np.

[keen sun]

Hy

[rugged wing]

How i can use chatGPT in discord not from web site

[mossy crescent]

You can't, there's no public api for discord bots to utilize onto ChatGPT's private (internal-only) api currently.

It's site-only.

[grizzled wasp]

i learned i can use links to share larger data sets with chatgpt, is there a similar method to getting around the output limit of 4k tokens?

[mossy crescent]

No, it's in place for a reason.

[grizzled wasp]

whats the reason?

presumably to limit processing output?

[mossy crescent]

Think about how much processing power it takes up to run such tasks, running an AI at the threshold they are currently on this scale to so many people

As well as the fact it's literally free during open-beta testing.

[grizzled wasp]

oh yeah for sure

[mossy crescent]

so they're making absolutely no profit from it one bit.

Make more sense now?

[grizzled wasp]

aren't they just going to ask it how they coudl profit off it?

[mossy crescent]

They already know how, they're going to be implementing the token-system which is already in usage over other OpenAI services, it's a pay as you go model.

You pay for <x> amount of tokens for <y> price, once you're out you top-up.

[rugged wing]

make chat gpt allowed or saudi

[eternal glacier]

Guys what is the official instagram account of chatgpt?
And also, does anybody know if there are referral links or an affiliates program?
I'm obsessed with this thing and want to share my link with. my friends if possible xd

Finally, are we all being charged the davinci model currently?
I already put my credit card in since I will be using it in the future

[mossy crescent]

If it's not accessible in that region, then it's not. It's in testing.

[mossy crescent]

afaik there isn't one?

[rugged wing]

I forgot the password. I tried to recover the account, but they do not send the recovery code to the email. Is it possible to solve it?

[eternal glacier]

I think there is only one for dall-e

[rugged wing]

that is not logical fr 💀

and kinda racist

[mossy crescent]

It's in open-beta testing for accessible regions listed somewhere (i don't have the link), they obviously have it limited for good reason and is in nowhere referencing the term or concept of racism.

[rugged wing]

will it be allowed in future ?

[mossy crescent]

And there is logic for it, the more regions have access the more they have to pay for resources to host services in such regions which host relay points for the AI.

@regal lynx I'm not racist right? sigh

[rugged wing]

will it be allowed in future ?

[regal lynx]

please don't spam needless things in this server. this channel is here to help people

[mossy crescent]

It's up to OpenAI what regions they want to give access to this service or not.

[rugged wing]

arent you a dev

[mossy crescent]

I'm a developer yes but not for OpenAI, I'm a freelance fullstack developer.

[rugged wing]

ok

is this even the offical server for open ai ?

[mossy crescent]

Yes it is.

[rugged wing]

i just did discord.gg/openai

ok thanks for your time me go bye

[regal lynx]

yes, this is the official OpenAI server. people that work at OpenAI will have yellow nametags. You'll likely not see them for a bit because their office is closed for the holidays

[crisp maple]

Hey guys I have a question
Would I get plagiarized if I used OpenAI to paraphrase my work? I've asked it to "remove words without its content" as I was over the limit.

[crimson temple]

How many conversations can you have stored on your account with ChatGPT?

[mossy crescent]

Just dumb it down a bit to sound like your writing.

[mossy crescent]

Not sure from the top of my head, might be as many as possible might not be, you'll have to experiment :)

[crisp maple]

It doesn't sound too professional. My only worry is whether I'd get plagerized by IB.

[mossy crescent]

If you're worrying about plagiarisers then the solution is clear, when you're not worrying then you're good...

[soft field]

How do I fix this thing? I can't access

I used password when signing up

[crisp maple]

As I said, used it for paraphrasing only...

[mossy crescent]

Contact through the help center, nothing much else can be done until it's fixed.
You can try using a 2nd account instead.

[soft field]

It's on every account

[mossy crescent]

Then no you won't, it only "flags" content if it's written at a high level, it's falsely flagged so many human-written content they're not reliable.

[crimson temple]

From an IP perspective - who own the output of the conversations in ChatGPT?

[mossy crescent]

OpenAI are the owners of that conversation data since they're storing that data at their facilities.

[crimson temple]

Suppose I take the output of a response in one of the conversations, edit it, then use that edited text in some of my own copyrighted works. Does OpenAI have a claim on the IP of my work?

[mossy crescent]

OpenAI only owns the data that correlates to the history, they own no ownership over that provided text

[cerulean forum]

I believe you should hire Lionel Hutz to represent you in a court of rubber chickens.

[crimson temple]

Could anyone point me in the direction of a more detailed Ts & Cs that documents the copyright/IP issues?

[mossy crescent]

Just hire a lawyer or talk to your already-existing one to go into depth about this if you're that worried about such things...

[vocal river]

I read some articles chatGPT is able to generate code but it doesn't seem to work. I keep getting the same generic response:
I'm sorry, but as an artificial intelligence, I do not have the ability to create HTML code or build a landing page.

[mossy crescent]

???

[vocal river]

Huh, that is weird

[crisp maple]

The stuff I type is not stored in the database for if another user asks the question right?

[crimson temple]

I mean it is hardly an issue, that is until someone manages to make millions with the next Harry Potter sorty that was partially produced by ChatGPT right...

[mossy crescent]

The content you write is stored in a form of a database but can't be used in another users instance/data.

[mossy crescent]

It can only be used in your content data instance where that data exists as well as the OpenAI team (depending on their position)

[stark kelp]

I have been trying to get openAI to write song lyrics but it keeps falling into extremely rigid patterns that it can't escape. I try telling it what to do, what not to do, sources to reference, examples, themes, but the more I try to get it to write anything good, instead it just becomes more and more narrow and starts repeating itself (example)

What keywords or prompts can I use to get it to write something more creative?

[mossy crescent]

Try doing it in a new chat instance

[stark kelp]

Yeah I have been, but its like circling a black hole, it doesnt get it right on the first go, and then it just gets worse and worse with every subsequent attempt.
I was hoping there was something different I was missing or could try beyond just try again over and over

[mossy crescent]

just try defining different parameters to the AI on what exactly you want such lyrics to sound and look like, not much else can be done since we don't know how you want such things to sound like or have a feel like.

[cerulean forum]

person + black hole = spaghetti. Yummy.

[vocal river]

Started a new chat instance and now it works... That's weird

[mossy crescent]

it happens sometimes.

[vocal river]

Now it's writing the html. It helps me to adjust the html. But now it seems to be stuck when writing it out

[mossy crescent]

you mean cutting out mid-typing?

[vocal river]

Yes, it just stops after a while

[mossy crescent]

it's intentional, happens due to reaching max tokens per message response, tell it to continue from "quote last full line written"

[stark kelp]

Using that as inspiration for an example, I gave it the prompt:
"Write a rap song about a person getting pulled into a black hole and turning into spaghetti."
After it's first attempt was bland and nothing but AABB 4-bar verses, I asked it
"Do not use verses or chorus. Make multi-syllable rhymes. Make internal rhymes."
It ignored my request to not use verses, though its rhymes sorta improved at the start but then just reverted to AABB

[mossy crescent]

you did not... 😂 my god Chia is going to love this hahahhaa

[cerulean forum]

Rap? I'd rather listen to a chainsaw.

[crimson temple]

ChatGPT says:
Obtaining permission to use the output of a chatbot, such as ChatGPT, in your own writing would typically involve contacting the owner of the chatbot and negotiating a license or permission to use the material in the way you want.

So are there any details for how to contact the entity that claims to be the owner of ChatGPT?

[stark kelp]

Its not terrible, but I am trying to get it to move away from AABB and 4 bar verses, and explicitly asking it to do that either gets ignored, or makes it become much more lyrically rigid

[vocal river]

thanks

[mossy crescent]

That's usage of the ChatGPT source code, the api and ai behind it through licencing and copyright ownership over such licence, not usage of data provided from it.

[crimson temple]

Seems like it a horrible mess... I mean I can claim that the prompts I give the chatbot are my copyrighted works... the terms of conditions of using ChatGPT say nothing about owning the copyright of work submitted to it.

[mossy crescent]

Just stop trying to define it in your own legal terms and again, as I said prior, get legal advice about this from someone who knows the law around these topics.

if you're that worried about it ^

[crimson temple]

Lol, no lawyer of IP specialist knows the answers to these questions Arianna... The technology is just too new, a veritable Pandora's box.

IP and copyright law is a terrible mess at the best of times.

[mossy crescent]

That's your issue, You are assuming such things when in reality the law around it is stated as is, if such data was a worry to the OpenAI company, they'd be lobbying congress and other governments to change the law to protect such data

until then, you're entirely fine.

since this service is being provided into the public domain free-of-charge, it's not outputting protected content.

[regal lynx]

If you have questions regarding copyright rights, please contact [e-mail] . We can't help you here beyond that

[mossy crescent]

thank you sigh

[crimson temple]

Lol - of course this is true - the corporate sector has already made sure that IP and copyright laws laregly exist for the benefit of corporations rather than protect the rights of individuals. I am wary of going down that rabbit hole as it could easily turn into a political discussion, don't you agree?

[regal lynx]

You can find the TOS here 🙂

[unreal hemlock]

<:book_icon:1046291503158399107> `` Rule 0 `` Adhere to the following Terms & Policies:

• Discord's Term of Service
• OpenAI Terms & Policies - Terms of Use, Service Terms, Privacy Policy, Sharing & Publication Policy, Usage Policy
• DALL∙E - Content Policy
• DALL∙E - Terms of Use

[stark kelp]

Is the problem maybe me? I started a new chat and tried to get it to tell me how it is doing what I request, but the examples it gives don't seem to actually be doing the thing it says its doing 🤔

[crimson temple]

Copyright issues aside - I have to admit the creative possibilities these tools provide are just wonderful!

[mossy crescent]

who knows... might just be from how you phrase things and context given to the ai.

too many factors here.

[stark kelp]

Yeah I have been trying different ways of phrasing the same request but so far nothing has worked

[mighty path]

I'm trying to use Dall E and I keep getting this very strange error message. Can't figure out what I'm doing wrong. Anyone got any ideas?

[mossy crescent]

Your API key is invalid, you need to go to that link and find your api key and use that one.

if you're just trying to access dall-e by itself, just use the link; https://labs.openai.com

[loud flare]

Does anyone know how openAI is handling super long conversations? For example, if the user sticks with only 1 conversation their entire time using chatGPT, then each time they hit enter, GPT has to process the entire prompt which could be enormous right?

I built a chatBot with GPT3 about 6 months ago and was running into this issue.

[mossy crescent]

It's able to due to how its language model works, it's able to process all variables how a human brain is able to and keep its history and add that as local variables that the AI language model can pull from and use along with the other model information.

[mighty path]

Oh, thanks! That's what I was trying to do. I guess I clicked on the wrong link or something

[mossy crescent]

Likely an api-related link haha

it's all good don't worry this stuff happens.

[loud flare]

But still, isn't it sending the entire conversation to the gpt model as a prompt each time the user asks another question on the same conversation?

That could be tens of thousands of tokens

[mossy crescent]

think about it like this, there's an AI base language model that an API pulls and requests from, the user data is local variables set to a chat, so it adds onto the data after the original language data model is pulled

[mossy crescent]

no one knows how much tokens it'll be but we'll have to wait and see

[loud flare]

Hmm, I'm not sure what you mean. Are you saying that the entire conversation is not used as a prompt each time the user submits a new message?

[mossy crescent]

Previously asked messages and answers are stored as local data that the ai language model adds to it already-existing large amount of knowledge temporarily

[pliant fable]

Is this data "per-conversation" or "per-user"?

[mossy crescent]

per-conversation per-user, each conversation is private to the user, user has access to all of these conversations but the ai model only has access to the per-conversation thats active.

[pliant fable]

understood, thanks

[gritty surge]

i am trying to make a very complex particle simulator, Any suggestions for prompts?

[mossy crescent]

make me a complex particle simulator.

[foggy anvil]

what is the best way to perform text clustering? using BERT? Openai text similarity models?

[mossy crescent]

depends on your preferred method since different companies and different people prefer it different ways

[foggy anvil]

I want to cluster tweets by their meaning

[gritty surge]

i have a part of its code already

[mossy crescent]

I'd suggest text similarity models then

[foggy anvil]

there is any way to do that with davinci 003? with the proper prompt...

[dim spear]

Hey, is it impossible to use OpenAi API and make people able to use the service in a different language ?

[orchid river]

Hey

[mossy crescent]

yes but it depends if you have enough tokens in the account per/request

[kindred epoch]

Hello

[mossy crescent]

every time it happens just gotta make sure you got sufficient tokens

[mossy crescent]

???
ChatGPT is available in different languages

[open tapir]

Yes

[foggy anvil]

I think that you can use the regular API and just write it in any language you want...

[dim spear]

It responds to my question then thanks

[dim spear]

Hm so create different ones in different languages

[foggy anvil]

Do you have an example of a prompt that can do that? tried million prompts though

[mossy crescent]

just go based on the api docs and the pricing page, https://openai.com/api/pricing
idk the api docs link off-by-heart.

[foggy anvil]

Thanks I'll take a look

[vast wasp]

after exceeding the limit of asked questions, is there any other possibility to ask more or am I doomed to wait?

[mossy crescent]

doomed to wait

[rugged wing]

Is chatGPT the most powerful tool to write code? Or is there any other tool even if it's paid that does it better?

[mossy crescent]

depends on your view on most powerful tool

since ChatGPT is a pull of resource informational data used in a language model, not tool.

[vast wasp]

do you know how long? I still have 6 hours of train travel ahead of me and I have nothing to do

[mossy crescent]

give it like 10 minutes, still happening after give it an hour

[vast wasp]

thanks a lot, you are very helpful!

[rugged wing]

Fair enough, would you know if is there any other AI that understands promps and is able to return code as well as this one?

[mossy crescent]

There's like one or two out there that are able to but none at the level of the language model is at, but the language model sometimes isn't able to optimize such code or best-write it in more up to date formats or recent 2022 language updates.

for example, the ChatGPT language model isn't able to identify where a pseudo element and pseudo clause should or shouldn't be used inside css code for example.

It'll embed it into an already-existing css element embedding it as something like &::test {}

[rugged wing]

Is there any "software" that would be able identify the correct way to do this?

[mossy crescent]

the documentation of that development language c;

[rugged wing]

I've gotten "too many request in 1 hour" or something along those lines, is there a way to bypass this? Maybe upgrading

[mossy crescent]

nope.

[rugged wing]

Sad, but thanks for all the answers

[barren mist]

Is it possible to purchase DALL-E 2 credits for someone else as a gift?

[mossy crescent]

not as far as I know currently, you could just send the user the money for the credits and they can buy it themselves.

[ivory swift]

How many time have we to wait when we got " "too many request in 1 hour" ?

[cerulean forum]

The current estimate is infinity.

[mossy crescent]

it should be an hour wait afaik.

might be 20-30 minutes either too, not entirely 100%.

[cerulean forum]

Will ChatGPT be available in a jar?

[barren mist]

Yeah but that's not fun

[mossy crescent]

it is now! :) c;

[cerulean forum]

That looks very nice. It is good to include a lid. Don't want to slosh AI on the starship control panel.

[mossy crescent]

can't be letting it have access to the outside air! gotta contain it

[ivory swift]

Okey thank you very much for your answer, i am correcting my engineering report so
i need chatGPT badly

[red python]

@dusk locust can you asnwer here

[mossy crescent]

???
ChatGPT can't be used through discord 😂

[kindred crown]

has anyone tried chatgpt for resumes?

[mossy crescent]

probably..

[red python]

yes got my fake fb sock hired as an admin on a landlords fb group so i could spy on how they break the law

[cerulean forum]

"I spent 5 years at NASA advocating hamsters in space."

[young notch]

The message you submitted was too long, please reload the conversation and submit something shorter." Why is it saying this, is this a new thing?

[dawn tapir]

I can't go back to the beginning of a chat anymore. Are the information still saved? It's important because I need to feed the chat with a lot of information to have a result afterwards. Not that the work was in vain.

[red python]

yea keep it conversational not instructional

crown look on the left side it should have your histroy

[dawn tapir]

Here are the different chats. But I mean in one chat the information is saved but the beginning of the chat automatically deletes itself. Does he still know what I told him even if the message is deleted because there were too many messages?

[mossy crescent]

try a new chat and see if it happens again

[empty birch]

while writing for my D&D campaign I am looking for a "in 8000 words describe an underground tower once created by humans and dwarves to survive the dragon wars but is now inhabited by kobolds and other denizens of the dark" Yet it gave me less than 2000 words. any suggestions?

[mossy crescent]

Should be, might just be a cache-related bug on your end.

[young notch]

almost in every "long" one, i think it is a new feature but i don't get it

[mossy crescent]

Can you send me an example of what you've send for this to happen in a pastebin in dms please so i can try this.

[young notch]

for example to try it, i sent a long article.

[mossy crescent]

it's due to the response limit set in place, tell it to continue from "quote last full line written"

[hexed moss]

@young notch WHEN IT STOPS, just write: Contine

[empty birch]

ok. so there is a response limit. got it. thanks.

[hexed moss]

continue*

[young notch]

no, no i mean, i can't send a long text to chatgpt...

[empty birch]

ok. so there is a response limit. got it. thanks.

[dawn tapir]

Hopefully. That was a lot of work. But one question: The information he receives is also read by others, right? He now knows more about me than I do because I'm using the therapeutic aspect of AI. 😂 😅

[mossy crescent]

It's not read by others unless OpenAI specifically want to or its been flagged as a violation.

and it only remembers from that individual chat session, each session has its local variable settings (messages)

[dawn tapir]

oh okay thanks 😄 Kind of creepy that this AI achieves a better therapeutic effect than a real psychologist.

[exotic marsh]

just reset the page

[mossy crescent]

resetting the page doesn't help in any context what??? 😂

[gloomy valley]

why can i only send one per hour

[cerulean forum]

Creepy, crawly, creepy crawly creepy crawly, Boooris the Spiiiiider. - The Who

[mossy crescent]

try swapping browsers.

[gloomy valley]

will do

[exotic marsh]

jk

[keen geyser]

Does anyone know the new character limit for requests?

[mossy crescent]

Should still be the same its been, if anythings changed give it time to be updated in the api docs.

[gilded surge]

mygPTADF

[mossy crescent]

@rugged wing can correct me if I'm wrong.

[rugged wing]

There isn't one as far as I know for chatgpt

For the api, there isn't one other than the max tokens

[keen geyser]

I used to use it to summarize YouTube videos but now i'm getting the error "The message you submitted was too long, please reload the conversation and submit something shorter."

[mossy crescent]

Which model are you using?

[rugged wing]

Do it in parts. Do it like 1 minute of the video at a time

[mossy crescent]

that could work too ^

[keen geyser]

Yeah that could work, Just wondering if you guys ran into any limits on your end. I'm just on ChatGPT not sure the model but ChatGPT Dec 15 Version

[mossy crescent]

I haven't personally encountered issues with asking long-text personally myself

[rugged wing]

Same

[mossy crescent]

Seems to be a limited issue to a small amount of people currently so nothing much can be done about that currently, sorry @keen geyser

[rugged wing]

french fry

[ashen magnet]

No matter what I say I can't get it to finish this class

[mossy crescent]

continue from the line "quote_line"

[ashen magnet]

so continue from the line "double encoderRad = raw"?

[rugged wing]

Yeah tell it to continue

[mossy crescent]

something like that

[ashen magnet]

I have done that a lot

[mossy crescent]

i personally do the last complete line it did

[ashen magnet]

it works in other circumstances but not for this class

[rugged wing]

Say "continue the code from <last line>"

[ashen magnet]

oh with that it actually finished the getAbsoluteEncoderRad method only

[teal steppe]

I want someone to help me on the upwork platform?

[mossy crescent]

because I'm smart (:
I've been telling people that for like 6 days now or something haha

[mossy crescent]

what if I choose not to help?

[rugged wing]

I've been doing it longer 😎

[mossy crescent]

prove it... :)

[rugged wing]

Bet

[teal steppe]

Thanks Ariana 😵‍💫

[ashen magnet]

i've a lot of things people said worked and that's the first that actually worked in this case lol

[teal steppe]

How do you

Sorry

[rugged wing]

Just one example (it goes further)

[mossy crescent]

Not quoting the line there though

[rugged wing]

Bet

[mossy crescent]

I'm talking specifically about quoting lines

[mossy crescent]

what can I say... I'm a special case (:

[teal steppe]

Need cover letter proposal samples,plz 🙏

[rugged wing]

Here @mossy crescent

[teal steppe]

Cuz I'm new

[mossy crescent]

You only said or

[rugged wing]

Wdym?

[mossy crescent]

not forcing the implication of the quoting the line

[rugged wing]

Bruh ok

[mossy crescent]

lmfao nah I'm just pushing time out, dw

[rugged wing]

💀

[mossy crescent]

shush, im bored I got nothing much to talk about

[rugged wing]

Same

[mossy crescent]

all my work for tn is done, it was just data analyzing and processing certain data fields and converting them to others

boring as duuuck work

[rugged wing]

I'm not boring :(

Am I?

[mossy crescent]

you are when all you say and respond with is "Ok" or dry responses

[rugged wing]

Ok

[mossy crescent]

my point exactly

[rugged wing]

Oops

[mossy crescent]

lmfaooo c;

[rugged wing]

It's a reflex

[mossy crescent]

again, concept, blaming something as a reflex when in reality you're very able to break it you just don't want to

since it gives you the excuse to not put effort into responses or anything

[rugged wing]

a l r i g h t

[teal steppe]

where are from guys

[rugged wing]

USA

[teal steppe]

Nice

[mossy crescent]

UK but half american

[rugged wing]

She is in UK

[teal steppe]

Welcome all

[mossy crescent]

Welcome @rugged wing glad you could join the server today! Welcome!!!

[teal steppe]

I'm Palestinian

[mossy crescent]

cool

[teal steppe]

Thanks

[rugged wing]

Hi I'm new here what is char gbt

[mossy crescent]

Response: This isn't ChatGPT, you might be looking for our service product @ https://chat.openai.com/chat

[rugged wing]

Ok

[teal steppe]

Nice tool i love it

[rugged wing]

It's cool

[mossy crescent]

It's okay...

I prefer Wikipedia though

[rugged wing]

Wikipedia pog

[teal steppe]

Are there better ones?

[rugged wing]

Not public ones

[mossy crescent]

I mean... There's a better wikipedia, i believe it's the library of congress or something like that... there's also in-person libraries which is absolutely disgusting

[rugged wing]

Library of babel

It's epic

[mossy crescent]

there's also the library of alexandria vatican which is very pog

[rugged wing]

The inside of me is pog

My intestines at least

[mossy crescent]

i'm staying the damn hell away from that

not with the amount of bread intake you have, it's absurd...

[rugged wing]

You know you want to see my spleen

[teal steppe]

You are really amazing. I wish I could be your friend. My name is Afnan

[mossy crescent]

🤮

[rugged wing]

Hi afnana

[lilac merlin]

...no, I do not.

[rugged wing]

How about the kidneys 😉

[lilac merlin]

Not that either.

[mossy crescent]

It's the waddler foot things which are best

[rugged wing]

The webbed feet

[mossy crescent]

yes

[rugged wing]

No

[cerulean forum]

Does ChatGPT like smooth or crunchy peanut butter?

[mossy crescent]

neither, it's allergic

[rugged wing]

Smooth is best

https://tenor.com/view/peanut-butter-dancing-jif-dance-pixelated-gif-17653026

[mossy crescent]

you should be banned for that gif alone.

[rugged wing]

https://tenor.com/view/bduck-duck-ducky-emoticon-emoji-gif-9816735

[mossy crescent]

https://tenor.com/XYd0.gif

[regal lynx]

Let's keep the channel on topic please

[rugged wing]

Yes Mr dino

[teal steppe]

Are friends long time?

[lilac merlin]

proceeds to send duck gifs on #off-topic

[rugged wing]

Not mine :(

[regal lynx]

We... can't be friends? 😢

[rugged wing]

We can :)

[mossy crescent]

define friends, I don't really have friends in the sense of what people see as friends, I just talk to people

[teal steppe]

Cool

Anyone speak Arabic?

[mossy crescent]

proceeds to open ChatGPT

[teal steppe]

Ok

[regal lynx]

Automod isn't a fan of anything other than English

[rugged wing]

proceeds to learn Arabic on duolingo

[rugged wing]

Fr. Once I said "yo dude" and got warned

[mossy crescent]

duolingo uses an ai from openai on french so id suggest learning french not arabic

[teal steppe]

I need to learn English very well in conversation? Need a way

[rugged wing]

learns polish

[mossy crescent]

you get warned for pretty much everything... smh

[rugged wing]

Fr

[mossy crescent]

i love how dino told us to be on-topic and we're just like here... existing... c;

[rugged wing]

Oh no... Delete everything fast

[regal lynx]

Wait..I thought this was #off-topic ! I'm going through too many channels without paying attention to what they are 😳

[mossy crescent]

LMFAO

[rugged wing]

Its off topic. You are hallucinating

[mossy crescent]

You very much are hallucinating, this is #off-topic, #off-topicis #community-help

[rugged wing]

All in one

[mossy crescent]

exactly

it's the #off-topic-questions

[rugged wing]

Yep (that was a question)

[regal lynx]

Okay, can we move to #off-topic though. I don't want to scare the newcomers 😅

[rugged wing]

Ok Mr dino

[mossy crescent]

scare them? Have you seen Jeff's profile picture?

[regal lynx]

ChatGPT is in a free research period. We don't know when it will be paid. Once it is paid, it won't be free

[mossy crescent]

Let me summarize;

ChatGPT will turn to premium-only using the token-model (pay-to-go model) after its free beta is over.

Prices will be similar to the davinci-003 model but a bit more pricier, we'll see with time.

[pale mortar]

Anyone know who to manage openai account such as change phone number ?

[mossy crescent]

just contact support

[pale mortar]

On discord ?

[mossy crescent]

through email, [e-mail]

[regal lynx]

I don't know for sure, but I assume it will be

Lots of companies use chat bots to enhance user experiences. ChatGPT would be a great use-case over GPT-3 for these cases

[mossy crescent]

since it's been trained in more content/resources than any out there publicly.

compared to when companies have to train it themselves.

[rugged wing]

Hey openai employees (that aren't going to see this), how long does it take for them to reply to a chat on the website or an email?

[lilac merlin]

In my testing, about 5 seconds.

[inland grove]

is it possible to make chatgpt write an article over 1000 words long? I can only get it to write up to about 600 words every time

[rugged wing]

It's been over 2 weeks for me 💀 (are you talking about the bot autoreply?)

[lilac merlin]

Oh. I thought you were talking about ChatGPT! 😛

[rugged wing]

Lmao I mean the openai staff

[rugged wing]

Tell it to keep going

[lilac merlin]

Oh, nevermind. I haven't contacted them.

[inland grove]

alright ill try that

[fringe stag]

i am 53 and this is better than waiting for someone to help you on the phone

[rugged wing]

Fr man (I'm not 53)

[inland grove]

thanks it worked

[regal lynx]

This is the wrong Discord server to ask that. We can't help you as we're the OpenAI server

[rugged wing]

Dm the bot on discord

[rugged wing]

I helped 😔

[harsh tapir]

how? how can i get there? do you have a link pls or what do i need to to exactly??

[rugged wing]

For MJ, just dm the bot with the command

[arctic echo]

Will the OPEN GPT always be free?

[rugged wing]

Oh no what are you typing dino

[regal lynx]

This server's image generation AI is Dall-E 2. That's the only model we officially support. Questions regarding other AI Image generation models should be asked in #ai-discussions 🙂

[rugged wing]

Only until the testing is over (probably a few more months)

[regal lynx]

People there will likely be more helpful than here

[rugged wing]

But I helped here...

[regal lynx]

No, it will eventually be paid. We don't know when or how much

[harsh tapir]

i just don't understand where to go i keep getting redirected to this page and in browser it says discord . com..

[rugged wing]

Let's move to #ai-discussions like Dino said

[harsh tapir]

ah ok sorry thank you very much bye

[gritty cargo]

can someone on this planet, knows how to explain this to the AI? i feel frustrated when i going to ask something that are very complicated to explain .. is there a very good way to communicate to chatGPT when it regards to image questions? its a MATLAB question btw.

[young cypress]

"too many questions asked in an hour" How many is this?

in chatgpt

[rugged wing]

A lot. Just wait an hour to fix

[flat mulch]

i gave it a 24 hours break and still

[rugged wing]

Try clearing cache and cookies

[flat mulch]

I did that before the break

and after

[rugged wing]

Idk

[flat mulch]

so sad

[whole socket]

How exactly does ChatGPT update? I had to restart it, and suddenly the answers to my prompts seemed to become more well-articulated and researched.

[regal lynx]

The model is stagnant and doesn't get updated unless the devs retrain it. This could be luck of the draw with seeds

[whole socket]

Wow, that's super interesting! The whole thing just amazes me.