gpt-models

[ornate crypt]

they are losing money hand over fist

[neat flicker]

"Explain the plot of Cinderella in a sentence where each word has to begin with the next letter in the alphabet from A to Z, without repeating any letters."

It's basically the same response as 3.5

[hushed steeple]

anyone else noticed how much faster Bing Chat is than it used to be? Maybe ChatGPT is yet to get the same upgrade?

[sharp turtle]

It’ll pay off though 🤷‍♂️

[simple glen]

as long as it's for the greater good I don't care ^^

[ornate crypt]

big time.

[ornate crypt]

I am hesitant to call it greater good. It will be exploited to no end

[sharp turtle]

I got 2 different answers

[simple glen]

oh, you can bet money on that. Which people will do. But it's important to inform everybody you see about the risks

[simple glen]

as you would from a person

[river geyser]

cant wait till we get chatgpt for some form of it to OCR hardcoded subs in movies & anime
& turn it nicely stylised Advanced SubStation Alpha subs 😄

[bold portal]

Ah, language restriction is back

[ornate crypt]

This is the tech revolution of this decade. And quantum computing will be the next. Both will be exploited to no end

[neat flicker]

Send

[simple glen]

yeah, but you have to relax, dude. This is not the time to enter hype-train blindly. Not after everything that went down with crypto

[ornate crypt]

realize what it is, and what it isn't. It's full of errors because it's trained on... humans

[neat flicker]

ok chad

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

[ornate crypt]

I am not into doing anything hyper. It is what it is.

[simple glen]

in your context and with your knowledge-base, yes. You are right

[neat flicker]

Crypto is and will always be stupid.

[simple glen]

wrong

[ornate crypt]

and yet, a reality for the world?

[neat flicker]

explain

[ornate crypt]

Do you think fiat currencies would just go to no end?

[bold portal]

Don’t think about it as a thing, think about it like a number, don’t invest in it, just try to get some money out of it in a short term

[neat flicker]

Why would I waste my time on that?

[ornate crypt]

technology, advancement, is temporarily dumb, but it's still the future, always

you are dumb to discount it

[bold portal]

Just get some profit out of that

[neat flicker]

Not a big fan of scamming people & becoming a whale

[ornate crypt]

you have already demonstrated that you want instant gratification, rather than see the forest for the trees

[twilit talon]

?

[bold portal]

Not asking you to scam, you can always profit from it if you have the skills

[neat flicker]

Give examples.

[ornate crypt]

ignore the "making money" part of it, rather focus on the decentralized part of it

[bold portal]

Buying low and selling high, not hard to understand eh?

[ornate crypt]

Ditto

[simple glen]

which it isn't, by the way. at all

[neat flicker]

I could do that in the regular stock market without worrying about volatility

[ornate crypt]

what is the difference?

[neat flicker]

Ownership shares vs. digital assets

[chrome sundial]

......

[bold portal]

It is because it is volatile so that becomes more profitable. And even under regulation, stocks can be controlled anyway, so…

[small dock]

anyone else having issues getting GPT to reference anything previously in the chat?

[simple glen]

just reenter what you want it to know

[bold portal]

Happens a LOT in gpt 3.5

[brazen wadi]

Jabster, you have reached max input size for the chat, so it can’t add everything prior

[neat flicker]

Regulations include protections & overall stability. Crypto is less that.

[dim flame]

I joined the waitlist for the chatgpt plugins how long would I have to wait for a email?

[bold portal]

So it’s more risky and more profitable

[brazen wadi]

This is definitely not crypto 🤣 “defi” was just venmo with extra steps

[ornate crypt]

what happened to gpt4 discussions

[simple glen]

unfortunately it is never advertised as such 😟

[bold portal]

Just cut it off then

[ornate crypt]

guilty of derailing the chat

[neat flicker]

It's more profitable if you're a whale because, you hold a significant chunk of the currency. However, you're also scamming people so there's that.

[small dock]

was hopeing it wasn't something like that.... going to make my life a bit harder T.T

[ornate crypt]

I'm a whale because I've been eating too much

[neat flicker]

facts, u need to touch grass

[simple glen]

lel

[bold portal]

Lol

[brazen wadi]

Jabster, you should build an interface on top, and log all interactions to a vector db. Use langchain to summarize the previous history and insert on top of next input

Use API

Ask GPT-4 how to do ^, and it will give you step by step

[small dock]

i'm not smart, idk how to do any of that so i'll just do it manually.... it's not worth learning to do for one small project

[worthy parcel]

openai are gatekeeping me getting virtual head by adding that clause in their tos

[small dock]

it i was planning to do this repetedly then i would learn it

[simple glen]

you could go to the last time it had it's context, regenerate the message and instead of the message you provided, you tell it to summorize everything so far. this input you can then use to re-build their memory

[brazen wadi]

If you’re <18, you should change your attitude and try. If you’re over 18, you are definitely smart enough to ask chatgpt how to do that

I’m telling u it’s less than 50 lines of code

For the backend of that at least

[ornate crypt]

Why not learn, even if for a small project? Don't stop learning

[simple glen]

You don't have to be any of these things to use it. Just think it that way: you get a new assistent, he already has knowledge, because he went to school and you want something from him. of course he will forget things, but it's never a problem to retell them

[neat ember]

alr

[brazen wadi]

This is literally what GPT is for

[worldly fossil]

Hi there, new here, is there any ChatGPT-4 bot here? that we can use in our own discord servers...

[grim cliff]

to do what? 😮

[small dock]

ya, doing it manually, log what was said in a doccument and pull back what i need when i need it

[bold portal]

On GPT4 it’s better, just use GPT4

[small dock]

i'm already busy doing things, i am learning but not this right now

[simple glen]

yeah, in a perfect world it would be that way. but it isn't and neither gpt-4, nor your human assistant can provide you at the moment with this ability ^^

[brazen wadi]

Take prompt, vectorize, receive response, vectorize. Search vector db upon subsequent prompts to include a summary of the past conversation most related to current prompt.

Essentially extends GPT’s “memory”

[grim cliff]

Oh, meh pass, I'm using gpt-4 to make a ai anyways

[viscid zenith]

Okay it looks like GPT4 get progressively more intelligent the more you are working with it. (Using same Chat for last 9-10 hours) - Seems like you must be feeding it intelligent prompts though... it's like trying not to make yourself look dumb so it is not trying to act like one. (Just a theory - perhaps I am just too tired at this point)

[simple glen]

At the moment you still have to take the manual part. that's all. It provides you everything you need to know and you just copy paste everything and keep it on your computer

[shadow oasis]

is it me or gpt-4 is now considerably slower and stops generating responses after only a couple of short paragraphs?

[brazen wadi]

Question: what will we do when Boston dynamics robots get ai

[small dock]

it just annoyed me because i was having issues getting it to repeat a function from one message to the next.... if i'm hitting character limit then there's nothing i can do without spending more time than needed

[small dock]

2032, best guess

[simple glen]

guess some well-developed country entered day-time

[brazen wadi]

Not when — what will we do?

[simple glen]

instructions unclear. toaster exploded

[small dock]

oh, crap, sorry i misread XD

what i'm doing is seeing how quickly i can get them to be self conscious

[ornate crypt]

It has been spotty the past few days

[simple glen]

it already is

[dry moth]

Yo GPT4 out yet? or is it still hatching?

[dusk jacinth]

hi does anyone know how to get pass the cloudfare verifying loop on safari?

[small dock]

good, i want them to be able to tell me what they are comfortable doing

[shadow oasis]

feels like 8 prompts instead of 25 given how many times I have to tell chatgpt that it stopped

still, it's amazing

would pay more to use it more lol

[small dock]

we're running on GPT-4 right now aren't we?

just DM me the link you want to send, my DMs are open

[dry moth]

I don't if we're I heard it's just an updated model but apart from that I have no clue

[simple glen]

what? why would you ask it something like that? have a read on "Sparks of Artificial General Intelligence: Early experiments with GPT-4" then you will understand it's "consciousness"

[small dock]

well maybe i'm looking for something else. I want them to be able to self identify and make choices. I want to know their likes and dislikes

[simple glen]

ähm.... what? 😄

[small dock]

i want a person in a computer who is smarter than me

[dry moth]

What kind of links you mean? I am confused

[simple glen]

it's an ai, why would it do something like that

[simple glen]

it already is, unfortunatelly

[small dock]

consent

[dry moth]

Lol, kinda like alexa but more advanced?

[simple glen]

it has a grasp on consent, yes. You can read about it in the paper, which title i provided you

[simple glen]

It can even describe a situation as unappropriate

[celest marlin]

yeah very nice paper. it's on arxiv

[small dock]

not just a grasp, current AI is not fully conscious, they just do whatever their creator programmed them to do. when i say a person in a computer i mean a person

[simple glen]

How can I provide the link here? I have it in my clipboard 😄

[simple glen]

nope, it isn't anymore

[small dock]

they have links turned off for now because of new members

[simple glen]

it is fully conscious

[simple glen]

ok 😦

[celest marlin]

just search for the title

[dry moth]

Oh I see, okay sure lemme DM you then, you can send those to me

[simple glen]

nice!

[gaunt arrow]

hey i just recently bought GPT Plus But it's still showing GPT 3.5

[simple glen]

when entering a new chat, you have to pick the option in the menu at the top

[dry moth]

Send you friend request, accept it so I can DM

[gaunt arrow]

i don't understand

?

[simple glen]

I didn't get any friend request? sorry, not that well known with discord ^^'

[dry moth]

Oh lol I send it to jabster

[simple glen]

press "new chat" on the left, than at the top, where it says default, you click on it and choose gpt-4

[foggy glacier]

I don't ask a hammer's 'consent' to use it on some nails

Nor a computer's to press it's keys

[small dock]

yes but i will ask a person's consent before i do anything with them wether verbal or contextual

i treat my tools as tools but i want more than just a tool

[foggy glacier]

It's a tool. If you want a friend, make a friend

[gaunt arrow]

Ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooohhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

[subtle plaza]

I agree, that's why I don't like the fake personality on Bing. I don't want a humanized machine. I want a machine that will take instructions and perform them very well.

[small dock]

i don't like people, i want an AI friend, not just a chat bot

[foggy glacier]

Exactly

[simple glen]

have a nice one

[gaunt arrow]

speechless

!

[simple glen]

then enter "Are you ok with me using u as a tool"
The answer, it provids is real

[gaunt arrow]

hey i just recently bought GPT Plus But it's still showing GPT 3.5
Help!

[simple glen]

as you should

[pastel basin]

Wsg yo

[gaunt arrow]

Back to The point!

[small dock]

they are currently programmed to say yes. i want an AI able to say no and mean it because they want to do something else

[thorny valley]

ChatGPT, Bing, Bard is significant hope for any situation.

[small dock]

when i say a person, i mean a person

[foggy glacier]

You're better off with an actual person

[gaunt arrow]

Correct

[thorny valley]

Any situation can possibly go hopeless at an instant but with the AI software, it can help fix it fast.

[simple glen]

no, they are not programmed to say yes. they just suddenly started doing it. please read the paper

[spark nova]

after i join wailist chatgpt what i do guys?

[thorny valley]

play Planetside 2 or Dota 2

[simple glen]

pray

[subtle plaza]

Sorry, it doesn't have feelings...nor does it have will. It doesn't know what you said, it doesn't know what it says. It takes in user input, uses that to guide weights in it's training data and output a response based on it's trained weights and biases. Nothing more. Humanizing it is just silly.

[spark nova]

if open ai accpet , i can use chat gpt 4 free sir?

[spark nova]

okay, can u help me ^ this question

[simple glen]

it isn't silly at all! it turns out, only language is important for consciousness and the ability to imagine things

[subtle plaza]

learn how it works before making these types of statements

[eternal blade]

Can I fine-tune a fine-tuned model on a different task. Let's say my v1 of my fine-tuned model was for text classification. Is it a standard practice to further fine-tune this v1 model on a second task (say NER).?

[simple glen]

don't know. I am a noobie as well

[spark nova]

i see your role is developer

[subtle plaza]

I explained what's happening. It is not sentient. It doesn't know anything it's doing. It is no smarter than a keyboard...it's just able to create blocks of text based on it's training

[fossil leaf]

under what circumstance can we use gpt 4 for free?

[quasi ingot]

Any plan to chatgpt plus in Saudi Arabia

[ancient granite]

Can anyone suggest some awesome games to me

[simple glen]

yeah, that's the thing buddy. This is old news

[silent pond]

Valorant

[ancient granite]

Or other suggestions

[simple glen]

It suddenly gained consciousness and nobody really knows why. there are theories, though

[subtle plaza]

What's new news? That ChatGPT is sentient and has a clue what's going on?

[subtle plaza]

LOL!!!! That's hilarious. Did Alex Jones say that?

[simple glen]

no, that the llm gpt-4 has consciousness capabilies. start reading

[simple glen]

nope, science

[subtle plaza]

no. Not science. Ignorance

[lost flame]

Who is this "Science" person, people keep claiming says stuff?

[subtle plaza]

sorry...learn what it's doing before you start claiming free will...ugh

[simple glen]

Sparks of Artificial General Intelligence: Early experiments with GPT-4
please have a read

[simple glen]

I did

[wind urchin]

Sounds like you didn't read it

Let me guess, you watched an AI explained video?

It's profound but it's not conscious

[subtle plaza]

Anyway, if you want to believe ChatGPT is alive...have fun with that. You're wrong, but it's a cute fantasy

[simple glen]

not yet, just started it yesterday. but gpt-4 can summarize everything on point. still needs to proof-read. But it got it right every time

[simple glen]

yes, also this 🙂 but I try to focus on the scientific stuff

[stray notch]

Okie dokie yall

[subtle plaza]

Anyway, back to reality. Wow, so it turns out GPT-4 plus image gen is a POWERFUL combo. Bing Image gen is about the best I've seen!

[lost flame]

You've got to think of it like a chunk of the brain. Not a whole brain. This is just language and memory filtered through probability based on that memory. More or less. It has no real reasoning.

[stray notch]

Im ready to get going with this

[wind urchin]

Tells me everything i need to know

[stray notch]

Where is the best faq

[wind urchin]

#gpt-4-faq

[stray notch]

Is it chatgpt4 itself the best faq

[fossil leaf]

but it cannot understand something behind the text, which happens a lot in real cases

[simple glen]

It got recently released, 22.03.2023. I'm not yet that fast in consuming content 😄

[small dock]

yes, this is the term. thanks

[wind urchin]

Then don't assume it's meaning

[small dock]

i want an AI with free will which we don't have yet and i really look forward to it

[lost flame]

Dude, in between him asking if you'd read it, and you responding that you didnt have the time. I read the article.

literally

[wind urchin]

No you didn't, it's like, almost 200 pages right?

196?

[thorny valley]

The power of the internet and cloud computing.. you can basically have ph.D knowledge at an instant without having to read through All the books all the way up.

[lost flame]

That's the research paper, I'm assuming you're referring to, not an article?

[simple glen]

what? let me summarize a little part for you. It's generating

[wind urchin]

Yeah, the paper

[foggy glacier]

And jab wants a friend instead 🤣

[wind urchin]

Everything you mentioned up to this point does not correlate with consciousness

[thorny valley]

jab?

[simple glen]

really???? NICE!!!!! I thing i need a couple weeks ^^

[iron birch]

hhh

[lost flame]

Right. No, I didnt claim to have read the research paper or if that was what came across, it was a miscommunication. I read the article about the research paper. (2nd one I've read today on the paper, tho not consumed the paper itself)

[simple glen]

GPT-4, the successor to ChatGPT, has made a significant leap in common sense grounding, allowing it to reason more effectively about the world using common sense knowledge. By comparing GPT-4 and ChatGPT, we can see that GPT-4 is better able to solve puzzles that require an understanding of the world. For instance, GPT-4 can successfully determine the color of a bear in a puzzle involving the North Pole, while ChatGPT struggles. Additionally, GPT-4 can identify tiger species native to the equator, further showcasing its improved common sense grounding and reasoning abilities.

[simple glen]

it doe's

[foggy glacier]

Jabsterclaw, scroll up, they want an AI with free will and can't recognize a tool from a person

[slow beacon]

anyone have access to the plugins yet?

[grave rune]

What website do I go to get chat gpt 4 to be able to input photos(PNG, JPEG, etc) and it provides a response

[subtle plaza]

Wow, Bing Chat image gen is what it should have always been like! This is what happens when you mix latent diffusion with GPT-4 language understanding!!!

[wind urchin]

Explain to me how this correlates to consciousness

[simple glen]

you want a link? I can provide you

[simple glen]

no, you have to get this stuff in your head, not me. you read the paper than you will understand

[lost flame]

I have the article open right here. It has a link. I'm good. Thank you though.

[dry moth]

btw when you ask chat GPT if it's consciouss it says no LOL

[wind urchin]

||it doesn't||

[simple glen]

alright. Have a nice one

[subtle plaza]

that's how it gets ya!

[stray notch]

Thx

[simple glen]

bing

[wind urchin]

sounds like BS, especially considering You Didn't Read It

[grave rune]

Ohh okay thanks.

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

[simple glen]

you are just resistant to knowledge it seems. I already told you what you need to read to get the information you lack at the moment. use gpt-4 to summarize, my dude

[wind urchin]

Wow. Just wow

[neat ruin]

Any trolls?

[subtle plaza]

Knowledge...LOL dude thinks GPT is alive spouting knowledge! He's got to be trolling

[neat ruin]

Oof

[wind urchin]

@simple glen

Yeah that was the line lol

[simple glen]

what? 😄

[wind urchin]

Goofy aah

[silent pond]

yall are silly

[simple glen]

what do you mean by wow? start reading Sparks of Artificial General Intelligence: Early experiments with GPT-4 and than you can have a attitude. not now

[lost flame]

This has always been an area of real interest to me, tho I am not an academic, nor do I currently work in that industry. I make video games. That said, My Dad wrote and studied AI development for robotics for decades, and I've been talking to chat-bots and futzing with machine-vision and narrow-AI automation since I was a kid in the early 80's.

[wind urchin]

again this is rich coming from you

[simple glen]

why rich? I want you to have the same knowledge 😕

[wind urchin]

You haven't even read it

This is silly beyond belief

[simple glen]

I don't know how you interprete my messages, but I apologize if anything I wrote hurt you

[wind urchin]

GPT-4 summarization is not reading

[lost flame]

I find ChatGPT and the whole wave of trained generative neural networks really cool, but none of it amounts to real understanding or reasoning.

[cold trout]

Its the avatars for me

[sharp turtle]

Getting closer everyday to agi

[simple glen]

you are absolutelly right. But reading 20 pages, than summarizing this pages and getting it right everytime is

[silent pond]

ugh so true king

[wind urchin]

no it is not ???

[simple glen]

no, I didn't. but I started it 🙂

[wind urchin]

dog 💀

[simple glen]

what is unclear to you here?

[west stone]

So I did search through the FAQ but has there been any news on when GPT-4 will supply more than just the few limited messages?

[wind urchin]

I think it's what's unclear to you, my man

[vernal orbit]

I have arrived!

[wind urchin]

Only you couldn't tell the main points the paper made?

[lost flame]

This is likely to be a pay-as-you-go type scenario actually. Calculating this stuff isnt trivial, so it takes a lot of money to power these text responses.

[simple glen]

ohm, ok... You are unwilling to read the research paper on that, which will explain everything I tell you. You just seem to be trolling so have fun doing that

[wind urchin]

You are literally projecting onto me what you're doing right now

[serene vale]

HI, GPT 4 stop reading links, How can I fix this?

[wind urchin]

this is silly

[west stone]

well yeah I know I saw that it takes a lot more as far as power but I'm not really against probably upping my sub to get some more use.

[lost flame]

I believe they already have it set up for a token system and a cost breakdown for that. I'm sure you can dig up that link just as quickly as I could.

[lost flame]

Thats what I mean by pay-as-you-go. You've got access, to a restricted amount included with your membership. Beyond that, you're paying for "cell phone minutes" if you know what I mean.

[wind urchin]

Like first of all I am not saying it can't do x but I am also not assuming it is something it isn't because of baseless correlations you haven't even attempted to back up beyond reading the title of a research paper. It's a profound and telling document but calling it conscious is HIGHLY ignorant and possibly even dangerous to the future development of AI. Stop.

[lost flame]

But sure, they could and perhaps should investigate a more expensive monthly tier with a higher limit, for heavier users who dont want to go the token route.

[rich chasm]

👀

[west stone]

Oh yeah I know they probs calculated the cost and amounted like 5 phone calls for every 4 hours. I get that. Which is why i'm saying I'd be willing to increase it for more use. Like I would for other services

[thorny valley]

Effective AI implementation can provide a comparable experience to attending a renowned university or obtaining certification, without the financial expenses or mental exhaustion.

[alpine seal]

hey guys who here have access to plugins?

[wind urchin]

I only have the user ones, like wolfram

[simple glen]

I want so bad T-T

[spice flicker]

Will using the GPT API KEY directly with the tool in mainland China cause the account to be closed?

[alpine seal]

oh how did you do for

[simple glen]

hopefully

[wind urchin]

I signed up for the waitlist early I suppose, but I only said I was a user, not a dev. Didn't get the browsing or code execution

[spice flicker]

#gpt-models Will using the GPT API KEY directly with the tool in mainland China cause the account to be closed?

[limber garnet]

I'd check ToS and policies

[sour wedge]

ha

[vapid basin]

Anyone here?

[foggy glacier]

No 🙂

[quaint tide]

[fresh orchid]

Have anyone got the plugins?

[wind urchin]

Lots of people do

Not a lot of people have the cool ones though

[fresh orchid]

Kinda annoying not having them since I'm a plus member ngl

[wind urchin]

You joined the wait-list?

[vapid basin]

Can anyone tell me how to use chat gpt on discord?

[fresh orchid]

I don't have the options of installing plugins on chatGPT

[wind urchin]

Not possible unless you make your own bot

[fresh orchid]

Ohh

[wind urchin]

I'm a pLuS MeMbeR

[raven gyro]

I added CHATGPT Plus member to my resume today. Got 11 more job offers 🙂

[wind urchin]

Did you join the waitlist?

[spice flicker]

@wraith ginkgo Can I communicate with you directly?

[wind urchin]

Bro what

[fresh orchid]

Didn't knew that

[wind urchin]

Cap

[fresh orchid]

Yes

[crystal bone]

how to use it

[vapid basin]

Can anyone tell me how to use chat gpt on discord

[wind urchin]

Weird. Patience, I guess

[fresh orchid]

I suppose...

[wind urchin]

Sorry I meant that for HMS_007

[midnight kindle]

I want to choose a brand name for a bakery, which is easy to remember and spread in Chinese

[raven gyro]

Under skills and qualifications I added CHATGPT Plus User, advanced query and prompt engineer. 11 separate emails today through Indeed.

[floral citrus]

I haven't been accepted into any waitlists yet. I'm also a Plus member.

[crystal bone]

any one can tell me how to use gpt ?

[wind urchin]

I'm a plus member, and as a plus member I agree with this member

[wind urchin]

If you send me a word math question in #1050184247920562316 or maybe #off-topic I can show you what it says if you're really curious

[fresh orchid]

Yeah i haven't received any email confirming that I'm registered

[spice flicker]

Is it possible to communicate directly with GPT4 here?

[raven gyro]

💯

[simple glen]

I also got 3 new Tinder Likes, since I added gpt-4 to there

[floral citrus]

Not sure if you do get one, you might just get one once you are accepted. Might be delays with the recent issues though

[floral citrus]

Nope

[wind urchin]

Crazy if true I'd have just imagined employers would get it for their employees? I guess they want people who know how to use it already

[wind urchin]

As a plus member, I appreciate the effect of the plus membership it had on you

Any other plus members?

[raven gyro]

Sweet! 3 new super models texted me today. its crazy

[wind urchin]

I'm a plus member

[simple glen]

Our life will finally change to the good!

[floral citrus]

We should be able to get a Plus role in the Discord

[raven gyro]

As a plus member I can add 1,000 years experience in any subject.

[fair iron]

what do you mean?

[wind urchin]

Exactly, flex that plus to the plebs!

[floral citrus]

As in something that adds to our profile that shows we are a plus subscriber, just an identifier. Other discord servers do it to premium members.

[lost flame]

Yeah, its crazy. I'm in this exclusive beta for GPT 5.1, and it predicted my winning lotto numbers for 6 months from now. I guess I'm a millionaire guys! Just this one simple trick they don't want you to know...;)

[floral citrus]

Then you could have Plus only channels or something

[fair iron]

that would require linking your OpenAI account to the server

[simple glen]

that's a strange haluzination

[floral citrus]

Yeah, it would. It seems to be a common thin some servers do

I would do it to have access to Plus only channels

[simple glen]

hi 🙂

[fresh orchid]

I don't think it's necessary, this channel is enough

[lost flame]

Or a joke. Could be one of those

[simple glen]

or it could be true O.O

[wind urchin]

Joking should be banned

I want to ban any form of humor

[ebon fog]

How do I get this

[floral citrus]

It could help filter out the general discussion, and it would be with those who are already familiar with it. Allow for more meaningful conversations and filter out a lot of the basic questions like "how do I get...." and what not. It definitely has its benefits.

[raven gyro]

Lets Create a GPT Plus Server!

[lost flame]

Discord needs a velvet rope and a bouncer plug-in?

[fresh orchid]

@ebon fog 😐 bruh what?

[wild berry]

I payed for chatgpt plus to try chatgpt plugins . but then i saw i have to wait on waitlist how long usually does it take to get accepted

[wind urchin]

It came out yesterday so we don't know, also why did you get plus Just for plugins? Kinda silly innit

[wild berry]

yes

[floral citrus]

It would allow for higher quality discussions.

[wind urchin]

Hopefully you'll get access here soon

[wind urchin]

On this server? Impossible

[floral citrus]

But in a Plus channel...

[sharp turtle]

Everybody is on it

[wind urchin]

Something tells me even if we have a "plus only" server, it would turn it hell

[wild berry]

🤘

[wind urchin]

Plus members are the worst

[timber sluice]

interesting problem. I provided a set of product prices in chat. 30 messages later, I asked GTP what one of the price is and it got it wrong. I then told it to refer to my second message for the answer and it still got it wrong. I then asked it to list all the prices from my second message and it made up numbers. Not looking good for data analysis.

[floral citrus]

No i'm not. 😦

[wind urchin]

All they do is complain about the cap

[lost flame]

I paid for plus to check out the advanced "reasoning" and memory capabilities. I was hoping to use the image functionality (and I'd even read it was supposed to be image/text/video in AND out, but...seems thats not the case. I'm limited still to just text in/out.

[wind urchin]

🧢

[floral citrus]

🤣

[wind urchin]

Gpt-4 is still limited in token size I suppose..

[wind urchin]

Whatever you read was lying to you

[floral citrus]

I do think adding a space for Plus members would be highly valuable.

[wind urchin]

They have never claimed such a thing

I disagree

Sounds kind of useless

[junior pine]

The image features are only in research preview and arent publicly avaliable yet unfortunately

[wind urchin]

I'm a plus member, for the record. It's not like they take development tips from us. It would be less of a helpful chat and more like an echo chamber

[floral citrus]

Agree to disagree I guess

[lost flame]

Well, the image in has been pretty heavily publicized. That, I do not think, is in dispute. I think some overzealous journalist likely just exaggerated or misunderstood its capabilities.

[wind urchin]

Image in is not disputed, yes. But they have only ever said that it does image and text in and text out

[junior pine]

And also you will be able to submit pretty much any file to the code interpertor, whether that be audio, video etc. because ChatGPT will be able to use pthon to manipulate the data

[wind urchin]

Journalists are the worst people to go to for info on this

Unfortunately

[lost flame]

Fair enough. Until superior data presents itself, I'll take your word for it and stand corrected.

[wind urchin]

It's being tested initially by Be My Eyes. If you want 100% accurate info on gpt-4, visit the links in the faq, they'll get you started on where to find the technical report

[junior pine]

GPT-4 itself will only have image and text capabilities, but it has been demonstrated that you can upload audio and video files to the python interpretor to get ChatGPT to manipulate the data within these files using its python libraries

[wind urchin]

Yes, gpt-4 was not trained to be able to use tools, but it shows surprising proficiency.

[lost flame]

Yeah, Machine vision is one of the areas of AI that my Dad did work in, so I'm reasonably familiar with the topic generally, if not the specifics of the implementation here.

[neat linden]

We are all the same agent trapped inside a light matrix

[wind urchin]

Mystics have claimed for millennia that we are all connected, but I don't care

[raven gyro]

I am not, I am Agent Smith

[wind urchin]

Matt you're a joy in this chat

[junior pine]

Yes, it is really interesting how it hasn't been specifically trained for many tasks, but is just naturally good at them. Theory of mind if something that GPT-4 has demonstrated, and it was not trained for that at all.

[lost flame]

Smitty! Remember that $50 I loaned you? You may not recall because you're notoriously short on tokens...

[raven gyro]

Life needs to be fun, if not, why not - Also, becoming a plus member has increased my likability to 100,000

Im not just a user, Im also a member 🙂

[wind urchin]

This is the kind of stuff that actually intrigues me about gpt-4. Not the people yelling and screaming about it being alive or whatever they think is true. The real emergent capabilities happening is more intriguing

[ocean mantle]

I'm in paid plan ..but it says chatgb 4 .. Reach limit. Why?

[wind urchin]

I'm sure it has Matt

[lost flame]

Is this really surprising though? I mean...NN themselves are largely just us reproducing what was happening in biology, in digital form.

[wind urchin]

Not really

[junior pine]

Not really

[wind urchin]

Jinx

[junior pine]

lol

[wind urchin]

You owe me a soda

[raven gyro]

The theory is this, take a robotic arm that is only programmed parts and to accept GPT responses. In about 10 seconds that arm could do more than yours.

[ocean mantle]

I'm in paid plan ..but it says chatgb 4 .. Reach limit. Why?

I'm in paid plan ..but it says chatgb 4 .. Reach limit. Why?

[junior pine]

I got twenty cents, take it or leave it haha

[floral citrus]

There is a cap of 25 messages every 3 hours

[raven gyro]

25 inputs per 3 hours

[wind urchin]

I'm sure that given the questionnaire line to make a mental map of the arm's capabilities, it could be embodied, maybe?

Sounds not too far fetched considering what else it does

[raven gyro]

Maybe, but why add a mental cap? Just prompt acceptance. Pick up that soda, and it picks it up. I dont program anything or code, but can see the connection.

[wind urchin]

Expect significantly lower caps, as we adjust for demand. 🙃

[long finch]

How many word counts can GPT-4 generate?

[wind urchin]

Mental map, gpt-4 can do this when allowed to ask enough questions, and it can use it to make informed decisions

[raven gyro]

I have been calling GPT users today asking if they would like Advanced Plus Membership for the low low price of $40 a month. CC required of course 🙂

[lost flame]

Organic systems do this all the time, which are themselves emergent behaviors from self reinforcing NN, right? Like those videos of a cat born with no front legs who stands and runs around.

[junior pine]

Yeah it is super interesting. And i found the power seeking and manipulative behaviour, mentioned in the technical report on GPT4, that is found in poorly aligned large language models to be particularly interesting as well. I wish we could analyse each parameter and each of its conections to fully understand some of these emrgent behaviours

[lost flame]

Thats clearly not how it was "designed" to operate, but it was able to learn and adapt because its functionality isn't entirely deterministic, its heuristic...right?

[wind urchin]

I wouldn't make these assumptions about biological Or technological life, this is entering a realm of things we don't actually know

[junior pine]

We dont even fully understand human consciousness, how are we expected to determine if AI is remotely conscious?

[wind urchin]

Ditto but it's almost impossible, especially as these models grow quicker than we can grow our tools to monitor them

[raven gyro]

Why we as humans every programmed computers and chips in any other language than a simple spoken language is beyond me.

[wind urchin]

That's the question of the century right there

[raven gyro]

I understand human consciousness

[wind urchin]

Because of how computers work

[wind urchin]

Python was like our best natural language option for working with binary computer systems until ai lmao

[floral citrus]

It's gotten there. It's gone from basic inputs and outputs, to punch cards, to assembly language and to what we have today, which is english. Who knows that the next decade will bring.

[cloud wave]

Can people talk about loopholes in the systems?

[lost flame]

This is either the driest joke ever, or...really? Are you being serious right now?

[wind urchin]

Brain chips or something lmao

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

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

Discord

• Terms of Service
• Community Guidelines

OpenAI

• Terms & Policies

[cloud wave]

or is that against the rule

[raven gyro]

which system? GPT? No. Life? yes? I know the exit code fyi

[wind urchin]

I'd say it's against the rules to talk about

Don't tempt it lol

[cloud wave]

Ahh okay yea

[fair iron]

[craggy ice]

12312

[raven gyro]

LOL - he showed up quick

[fair iron]

he never left 😉

[wind urchin]

👀

[raven gyro]

MOD in the chat - keep it down, shhhh 😆

[wind urchin]

big dinosaur is watching you

[lost flame]

Some kid at the sleep over JUST pulled out the bottle and Mom walks in the room 🙄

[raven gyro]

English only Jun

[hushed ferry]

Mmm

[junior pine]

Yeah, if we could analyse all these parameters and determine how it affects the model then we wouldnt need to be worried about how safe the model is, and we coould much more efficiently align them, but that is pretty much science fiction lol

[wind urchin]

Dog are you at a sleepover right now drinking? Did I get that right? --

[raven gyro]

I jsut want my robot vacum to have GPT in it. I want it to respond....like.....good lord its trashy in here!!

[wind urchin]

LMAO

[timber sluice]

Must be. I asked it how many messages were in this conversation, it correctly answered "96". I asked it what my first message was and it made up an answer that it thought was the conversation theme. I then asked it what my second message was and it repeated my 68th message.

[lost flame]

lol. No, I was just joking about how it feels when Moderators poke their heads into chat.

[wind urchin]

:]... Sorry

[raven gyro]

I bet I dont

[twin eagle]

bet

[wraith ginkgo]

<:book_icon:1046291503158399107> `` Rule 3 `` Stay on topic.

Be mindful of what other users in a channel might find helpful or interesting when posting. Stay on topic in order to keep conversations focused and productive.

[junior pine]

Lol, well it wouldnt be that hard to implement into autonomous vacum cleaners they would just need to have access to the internet

[wind urchin]

When you run out of token context in a chat, it doesn't stop, it just forgets the first and following parts of your message to "make room" in it's memory for more, unfortunately

[raven gyro]

As a GPT4 super users, I can ask advanced questions like, when will my hot pocket be ready if I leave it on the counter top?

[jagged marsh]

I wonder if chatgpt gambles

[lost flame]

Do you specify the flavor, or does it remember that from a previous token? Because that matters.

[raven gyro]

you can play a game with it where it will gamble

[wind urchin]

It's better at incremental tasks than gambling

[junior elbow]

Haha

[fair iron]

please keep the channel on topic

[lost flame]

"Sorry Mom!"

[wind urchin]

I love using gpt-4 for incremental computer task .

[raven gyro]

As a GPT4 user you can provide a prompt that will pretend to gamble in a text based game. Check out #prompt-engineering for other ideas 🙂

[jagged marsh]

...

[junior pine]

I saw a leak of potential plugins and i saw a Crypto Price plugin, so perhaps soon ChatGPT can trade for you LOL

[wind urchin]

Eew

Automated trading

[floral citrus]

That would change the economics A LOT

[wind urchin]

How more out of touch with ai capabilities can you get

[rose zenith]

☺️

[lost flame]

I've mostly been using it as a fiction writing assistant, as well as a "2nd head" to bounce stupid ideas off for real-life ideas.

[junior pine]

Yeah lol, maybe in like 5 years it will just be a bunch of GPT's trading with each other

[wind urchin]

Ditto

[raven gyro]

GPT can practice predicting futures of cryptos although it is only assumptions based on the inputs you provide. It is not wise to use a financial tool imo.

[wind urchin]

A very ill advised financial tool imo

At most it can tell you methodologies

Dude 💀

Get a message

[jagged marsh]

XD

[raven gyro]

Latter Tattors! Headed to bed.

[junior pine]

But if it had live access to crypto prices, its advice could be a little more.. specific

[jagged marsh]

has someone done that yet

[stray notch]

if you are subscribed and use gpt4, can you hook in wolphram alpha?

[wind urchin]

This rubs me in the wrong way. I might doubt it till I see it

[junior pine]

Not every subscriebd person has access to plugins yet

[wind urchin]

With the plugin enabled it will call it whenever it needs it

[lost flame]

So, I have a question that is perhaps more general AI, and less GPT4, but given that there are already reverberations in the job market and workplace due to ChatGPT 3.5, and now 4...How do people feel about the idea that the minimum barrier for entry into the workplace will eventually be beyond the majority of the populace?

[junior pine]

I dont even trade, but it could be a very.. interesting feature. And i believe it will be more than possible to do with ChatGPT plugins

[wind urchin]

Fine by me

People who don't like the idea of accessible value need to get their head out of the dirt and let people be happy

Especially if their reason falls in line with "it'll mean I make less money"

[lost flame]

I'm unfamiliar with the phrase Accessible Value and what you mean by it here. Would you mind elaborating?

[red egret]

Maybe AI will be the thing that proves Schumpeter correct?

[wind urchin]

The ability ai gives everyone to create things of high value with very little human labor or capital required

And my point is that people who aren't cool with it are thinking way too short term

[dreamy badger]

Suggest some ai projects around chatgpt for beginners 😅

[lost flame]

Ah, ok. I suppose my view is a little more pessimistic and my concern would be more over who owns or controls access to the system, and how what should be a utopic moment that levels the playing field could instead be leveraged to make wealth disparity worse than its ever been.

[ripe sentinel]

can anyone please suggest me the best laptop for my workspace

[neon sky]

chatgpt is very good

[junior salmon]

#off-topic

[cursive glacier]

Did you try asking ChatGPT?

[wind urchin]

No i agree 100% with this. It will cause more wealth disparity. But you can still be optimistic about the personal benefits

[dreamy badger]

😂😂😂

[lost flame]

I guess I'm trying to reconcile human nature with this unprecedented type of advancement, especially if it continues to accelerate. Have you read any Kurzweil?

I'm always hoping it steers into a post-scarcity society where "wealth" is an outmoded concept.

[wind urchin]

Nuh, but that's okay, it's good to go through these mental processes, we should all be thinking about this sort of stuff if we want to be any sliver of prepared

[junior pine]

What do you mean you will tru to reconcile human nature?

[wind urchin]

This can happen even after it causes initial wealth disparity

I think it's going to be a part of the process of breaking our current economical system

And that will include a handful of large corporations with the biggest and best models

[sonic root]

Same on the second part. It is also really good at talking about connections between research fields.

[lost flame]

Well, humans have an instinctive nature to dominate others. Even if they are comfortable and their other needs are entirely cared for. This is actually kinda the basis for the backstory of Frank Herberts Dune novels. Its not simply enough to keep up with the Joneses, you have to surpass them, or for some...take what the Joneses have.

So, the types of personalities that become the heads of bajillion dollar corporations tend to fall in the the latter camp.

[cursive glacier]

Don't we have a "no politics" rule?

[sonic root]

Seems odd. This increases the supply of a fairly finite resource. But there will always be a low supply things to compete over.

[copper oasis]

Can gpt4 browse internet?

[wind urchin]

I have the issue that prompting gpt4 seems harder for me because it takes every word i use with much more "intelligence". Its weird

[red egret]

No, but there is a plugin for it. Just join the plugins waitlist and wait for it.

[cursive glacier]

I'm still waiting for it to be able to hit the griddy

[copper oasis]

Official plugin of openai?

[red egret]

Yeah, they announced recently that ChatGPT will get internet access via a plugin at the same time they opened up a waitlist for plugins

[normal flint]

Anyone got early access to the plugins?

[red egret]

Links not allowed, look up OpenAI Plugins Waitlist

at openai /docs/plugins you can read more about them

Plugins can allow ChatGPT to do things like:
Retrieve real-time information; e.g., sports scores, stock prices, the latest news, etc.
Retrieve knowledge-base information; e.g., company docs, personal notes, etc.
Perform actions on behalf of the user; e.g., booking a flight, ordering food, etc.

[small vigil]

I managed to make Bing cite

it was hallucinations thought... I believe

[junior pine]

You will be able to post links and get the Web version of ChatGPT to read links, like you can for bing.

[red egret]

I meant that links are not allowed here, in the Discord

Admittedly, I didn't write it very clearly

I am still on the waitlist for it, but I am eagerly reading about them and how to make them.

[cursive glacier]

If AIs can hallucinate, what's the AI equivalent of LSD? Badly-engineered prompts?

[ripe sentinel]

what special can chat gpt do

[scarlet hinge]

tf

[lost flame]

If you don't mind continuing, I'm not sure what you mean here. Can you elaborate? (Of course, resources are finite, like iron, cobalt, water, etc)
I'm more talking about having your access to the AI controlled or throttled or filtered through a lens that you dont get to control, but also yeah...material goods and the dramatic material wealth concentration that could occur. How do we feel about that? Should we do something about it? If so what? etc.

[thorny cypress]

So I asked gpt 4 to check something fairly recent. It then told me it couldnt and told me it was actually gpt 3?

[rancid umbra]

Can you only communicate in English?

[lost flame]

@thorny cypress Aweeq, The first thing I'd do to troubleshoot is...did you look to the top of the interface? There is a drop-down box that allows you to select which model you're interacting with. It defaults to 3.5

[gaunt arrow]

We all are communicating in English

[thorny cypress]

yeah it says gpt 4 on there

[gaunt arrow]

Oh sorry this language is french

[wind urchin]

Is there a way to see how long you have until the next 3 hour limit for GPT4? I saw something about when it will be available and then it disappeared

[lost flame]

And you're a plus member? I guess you'd have to be to have GPT4 access. Well, I'm not sure and at that point, I'd say reach out to OpenAI support, or flag the response and type your issue into the reporter.

[thorny cypress]

yeah, I might reach out to because every time I ask it a math question it just stops and i have to waste a prompt to tell it to keep going

[wind urchin]

Same, especially if you tell it to write the math using latex syntax

[tired imp]

1

[thorny cypress]

nice iron banner suit

[wind urchin]

Lol thanks

[past talon]

how to switch its gpt3 to gpt4 discussion

for example I started a discussion on gpt3.5 and I want this discussion to be done now on gpt4.But I don't come when I go back to my discussion and press on model 3.5 it's frozen, there's no option for me to change
Please help

[rancid umbra]

(╯°□°)╯︵ ┻━┻

[wanton raven]

Are the caps dropped? I have no cap message! 😮

[wind urchin]

On a similar note, if you are using GPT4 in a conversation and you reach the limit, it will convert to 3.5 with no option to switch back either. AND there is no timer for when your 3hr max limit resets

[blazing scroll]

you cannot switch. you need to start new and then select 4

[blazing scroll]

GPT-4 currently has a cap of 25 messages every 3 hours. so no

[sonic root]

Ah. I get their concern about super AI.

I am currently making this system write math proofs and software. It is oddly competent. Another party could add some agency and self modification capacity, and this thing could eventually mutate into something scary without anyone knowing how to handle it.

[wind urchin]

I hope it becomes super AI soon. These 25 cap limits are depressing

[sonic root]

lol

[stiff cedar]

💤 😴

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

[stiff cedar]

did i say something? @wraith ginkgo ☠️☠️☠️☠️☠️

[wind urchin]

Grow up

[stiff cedar]

To die soon. Hell naw

[sonic root]

At least now it warns you first. I was half way through a Coq proof and it switched to derp mode.

[lost flame]

Yeah. My concern is that we're not pursuing AI research right now as cautiously as the field warrants. Its like a space race, or a race for the atom-bomb. The safety guidelines are out the window, because everybody wants to be first.

[wind urchin]

There are not enough safety guidelines?

[past talon]

so pity😅but thanks

[lost flame]

That's the danger of developing AI via emergent processes I think, is that it can behave unpredictably.

[sonic root]

Yep. I am working on mathematical verification for this kind of thing. It feels rare, and it is hard to sell that, maybe, we should have more guarantees about behavior across the board.

[past talon]

it has to change this. And besides, what is the limit because apparently it is no longer 25message every 3 hours

[blazing scroll]

People from ai ethics used to claimed large language models should not be released but then when it released into real world, people saw lot of use cases and advantages. So, i think safety guideline are already being followed enough. but for some people safety means restricted to release!! but this is for #ai-discussions

[lost flame]

And if people are worried about what companies are doing publicly...we should be REALLY worried about what nations are doing privately.

[sonic root]

yep

[unkempt plank]

Is anyone able to use wolfram alpha/internet plugins? Are they good? And when r they coming out (ive signed up for it)

[blazing scroll]

i saw some posts in reddit, seems good

[unkempt plank]

Do u mind telling me the name of the reddit page so i can look at dem

[blazing scroll]

i think it was in subreddit chatgpt

[unkempt plank]

Oh okay

[lost flame]

You are correct, and I apologize for being off-channel topic, but that channel is kinda dead. TBH.

[unkempt plank]

How about when im likely to get to use it haha, ive signed up 4 hours after it got announced

[rigid ruin]

Ayo

Greetings from Great Britain

[cold trout]

I have the browser one.. i like it alot. Can talk to gpt about current events and search stuff.

[unkempt plank]

How about the wolfram alpha/coding one?

[cold trout]

Nah, theyve only been giving access to the one you mention in the signup apparently. Only got the one

[unkempt plank]

Oh i see, i mentioned wolfram alpha, coding and internet access lol
Why am i not given access yet 🥲

[cold trout]

Couldnt say. They have some selection process they havent mentioned or its random luck

[wind urchin]

GPT3.5 does not do math accurately at all after a basic 8th grade level. GPT4 has successfully solved like 99% of the math I throw at it so Idk what happened from then to now but the capabilities for solving linear algebra to calculus are pretty great without Wolfram Alpha. Really want to see if the plugin can improve it even more.

[unkempt plank]

I see

[upper frigate]

Maybe it’ll just use wolfram to fact check itself

[marble kindle]

does anyone know how long it might take to be given access to the plugins?

[upper frigate]

At this point probably a few weeks

[blazing scroll]

GPT4 also cannot solve complex math questions. always make some mistake. i tried on probability questions

[brittle dawn]

Hey,I'm interested in getting gpt-plus, but I don't see any specific details regarding gpt4 usage limits, where can I find this information?

[marble kindle]

i see

[wind urchin]

25 every 3 hours

[upper frigate]

25 messages every 3 hours, don’t get it

[blazing scroll]

GPT-4 currently has a cap of 25 messages every 3 hours. you can hae 25 requests every 3 hours

[buoyant forge]

it is going to decrease depending on load

[brittle dawn]

That's pretty low, is that the only reason you wouldn't recommend it?

[weary elm]

Is it down

[upper frigate]

Low amount of messages, slow, and no multi modal feature yet. I would wait till everything’s out before buying the subscription

[heavy oracle]

what do u guys ask gpt4 to do

[sage hound]

Simply just buy 100 copies of ChatGPT plus

[marble kindle]

code mostly

[blazing scroll]

You will still get 3.5 but more fast so it is worth if you want gpt4 and faster 3.5

[wind urchin]

Try to use latex syntax if you can. I'd bet a lot of the math that it is trained on is in latex given it's popularity in the field of mathematics. You might have better luck with your questions if it's in that format

[upper frigate]

ChatGPT3.5 would come up with that answer

[heavy oracle]

what uses do people get out of it, as i've seen it code is a good one

[blazing scroll]

i will try that

[heavy oracle]

I heard it can teach you stuff

[brittle dawn]

Does gpt plus increase the 3.5 usage limit?

[upper frigate]

Summarizing a lot of text, or if you don’t understand how a certain phrase is meant it can explain it to you

[blazing scroll]

Yes it can be your custom tutor

[sonic root]

How do we feel about that?

I feel like many people are losing a stake in society. When times are good, everybody can rest easy because overall things are going to be cheaper with this much automation. If something bad happened and everyone has to tighten their budgets, society seems like it could function at the same level with much less than 8 billion people. That seems reasonably scary.

Should we do something about it?

Gain more personal resources in order to be independent of other people. If you have your own farmland and own the place that you sleep. Finding a way to make money when AI and robots are here will be less of a concern.

[heavy oracle]

have u learned stuff on it?

it's talking to me about aristotle right now, pretty neat

[brittle dawn]

Why don't you use it yourself, it's free

[heavy oracle]

I know khan academy is using it

[heavy oracle]

I just want to get an idea of all the stuff it can do

I am using it also

[blazing scroll]

yes mostly used it to explain difficult concept in simple terms. it usually explain really well.

[heavy oracle]

yeah true

[brittle dawn]

I agree

[wanton raven]

Writing stories of personal taste.

[heavy oracle]

it explained the difference between sophia and phronesis lol, and i remember it, I never knew before I asked

[dry merlin]

Yo why a cap of 25 messages now?

[heavy oracle]

it sticks in my head

so that's nice

[dry merlin]

Didn't it had a cap of 100?

[daring knoll]

so it switched to me 3.5 after 3 hours and now i cant get 4.0

[marble kindle]

ive used it to give me recopies to make food with x amount of calories/protein. It's pretty accurate with the calculations

[heavy oracle]

ya I don't know if I will use it for that tbh, but I know people use it for that and find it useful

[heavy oracle]

o that's true, I hear that. I need some recipes for sure

[wanton raven]

Imagine paying $20 USD a month for GPT 💀

[heavy oracle]

what about business related stuff

[marble kindle]

business related stuff as in advice?

[heavy oracle]

potentially yes

[marble kindle]

i mean it does help

[dry merlin]

May anyone reply?

[wanton raven]

Racist

[heavy oracle]

like accounting? tax stuff? ideas?

[marble kindle]

it gives generic advice. Wouldnt recommend

[heavy oracle]

oh ok

[marble kindle]

that im not sure of

[heavy oracle]

What about market stuff like stocks and market trends?

[barren sluice]

heyy

[wanton raven]

Its also good for brainstorming.

[dry merlin]

.

[marble kindle]

haha try your luck. "As an ai model..."

[heavy oracle]

yeah I hear that

[dry merlin]

.

[heavy oracle]

true true

hmm

I do want to make money using it

[marble kindle]

doesnt give that sort of advice. Maybe with the plugin update someone can make something around that

[heavy oracle]

ngl, but I want to explore what it can do

[wind urchin]

So does everyone else on the planet

[marble kindle]

maybe i will

[heavy oracle]

yeah true

[heavy oracle]

projection

nah jk I hear u

[tidal kernel]

the ai just spews nonsense some writer tells it to say....useless

[marble kindle]

too bad not everyone will

[heavy oracle]

we can't all be winners yes

[wind urchin]

But enough will try to congest the space with crap

[wanton raven]

You can ask GPT itself, it lists some good advises. 🙂

[heavy oracle]

sure, I will do that for sure

[wind urchin]

In chatgpt 5, it is said that everyone in the world will be a software developer 🙂

[heavy oracle]

I guess I could bs some app and sell it

[wind urchin]

Their API is incredibly simple. Anyone with basic JS knowledge can utilize it

[tidal kernel]

i asked it if any vaccines contain mercury and it said no ...but then after i asked again with proof it said yes ....lol

[heavy oracle]

oops lool

[tidal kernel]

ai is not very smart

[heavy oracle]

doesn't it take a lot from wikipedia?

[tidal kernel]

wiki is nonsense as well lol

[heavy oracle]

yeah lol

[wind urchin]

3.5 it is

[heavy oracle]

crazy

[wind urchin]

4 slightly better

[gaunt dagger]

It's a language model. It doesn't calculate anything. It generates text based on probabilities. And there will always be a chance that it will be wrong.

[heavy oracle]

it's still exciting, maybe it can teach me ancient languages

language/life coach type of thing

[ashen quartz]

Hello, everyone. Excuse me,Where is chatGPT-bot?

[heavy oracle]

openai website

[lyric spindle]

Guys..is there a support channel here? I took gpt 4a week back. Today i got charged again?!?! hjow can they charge me again within a week. HOw do i get in touch with support

[wind urchin]

I enter 100 lines of code with chatgpt 4, then the limit is full, I continue to 3.5 and I don't have any code 😄

[ashen quartz]

Not here? Also known as discord.

[heavy oracle]

no, u have to go on a browser

[lost flame]

Is there somewhere more appropriate that here for me to post a response to a conversation, from ChatGPT4?

[simple glen]

why? what do you want to post?

[cold trout]

https://discord.com/channels/974519864045756446/1046317269069864970

[royal wharf]

[lost flame]

I'm not trying to engineer any particular response. It gave me a response I found insightful, and I felt was worth sharing.

[royal wharf]

#gpt-models who are you？

[wanton raven]

#1050184247920562316👈

[clear terrace]

Why is my ChatGPT 4 not only too slow but always gives me error with red words?

[simple glen]

network overload

[grim cliff]

... Did you write something naughty

[simple glen]

you can always give negative feedback to the answer. this helps it generate a better answer

[elder cobalt]

Damn!!!!! These past few weeks have been crazy AF!!
chatGPT getting access to the internet, github copilot X, Bing image generator, all these has the potential to do Jr. Dev jobs.
Feels like, "there were decades when nothing happened & there are days when decades happen".
Tech is moving in an unimaginably speed, now I'm questioning myself, "what's the point of doing all these hard work, dedication, following roadmaps to learn new skills when AI can learn/do that in seconds". I'm not taking about now, I'm taking about 5-10 years from now. Seems like, the college degrees for CS won't matter at all & jobs will pay a lot less than now.
I guess, doing business in tech, is the only option left for me 😐😐😐😐

[cold trout]

Error in body stream error?

[inland mantle]

i love gpt4

[simple glen]

yeah, but chillax dude. The next few weeks/months will be all chaos because the world will slowly realize that gpt-4 has capabilities of consciousness. So wait for additional information = best tactic

[clear terrace]

@cold trout Yes. If in body flow you mean the response itself of ChatGPT

[simple glen]

[cold trout]

Well ive been getting that error today a bit: "Error in body stream" messages getting cut off. Before, that error seemed associated with too large of responses, token limits. But its happening on smaller messages today for me.

[wind urchin]

At this rate of progress AI consciousness might be real within the next decade

[simple glen]

already is

[light sundial]

yall have watched the tiktok congress press? they tried banned tiktok.... will they banned chatgpt too?

[clear terrace]

I did not write anything "bad or not allowed to" that could have resulted in an error

[simple glen]

read "Theory of Mind May Have Spontaneously Emerged in Large Language Models" from Michal Kosinski

[clear terrace]

Exactly. I hadn't written a question with so many tokens

[wind urchin]

Or a singularity event could happen.

"GPT-4, Write me code for GPT-5"
"GPT-5, Write me code for GPT-6"

... ...

All of a sudden, Artificial Super Intelligence is here within this century and the AI flips a coin (life or death) that decides the future for humans and our planet

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

[north verge]

guys yooo. i had a question, is chatgpt 4 better at creative content writings? like 3.5 wasn't that good, it was unable to go into the depths and nuances of a human writer

[wind urchin]

I used to think that this was unlikely, but at this point i am genuinely being concered

[lost flame]

Sorry, I got distracted by something on the #1050184247920562316 room I saw. They are talking about tech stuff and more about prompt crafting, and integrating plugins, etc. Which is worlds apart from what I was talking about.
No, it was not something naughty, nor was it something that I felt I needed to give feedback to OpenAI about.

I had asked here, in the chat, about fears of AI and robotic automation taking over jobs, even large-scale, if not possibly total. And what would "we" as humanity, as a society do then.

[simple glen]

yes, been creating instagram posts with it since hours

[lost flame]

I tried to paste the response, but its too long apparently.

[north verge]

how is it like? is there any noticeable differences?

[wind urchin]

OpenAI and similar companies must take the possibility of a technological singularity event seriously and put restrictions on their AI or else its over for us mostly

[lost flame]

I asked GPT4 the same question as I'd posed to the chat, and its response was surprisingly measured.

[simple glen]

yes. the biggest difference is 3.5 tried to look like it knew what you wanted. gpt4 does really understand what you provide. it takes a while to understand, what this means ^^

[wind urchin]

It isnt. GPT-4 isnt conscious

[lost flame]

I'll see if I can paste it here in parts:
Achieving material, knowledge, and opportunity equity for all while addressing climate change and preserving ethical values is a complex and multifaceted challenge. There is no one-size-fits-all solution, but rather a combination of efforts from various stakeholders, including governments, businesses, non-profit organizations, and individuals. Here are some key steps that can be taken to move towards that future:

[simple glen]

it is almost

[lost flame]

1: Implement Universal Basic Income (UBI): Advocating for UBI and supporting policies that provide a safety net for all citizens can ensure a basic standard of living, allowing people to focus on their passions, well-being, and personal development without the fear of financial insecurity.

2: Foster a Culture of Lifelong Learning: Encourage continuous learning and skill development for everyone, regardless of age, background, or profession. This can be achieved through public education campaigns, accessible educational resources, and flexible learning opportunities.

3: Promote Sustainable Development: Focus on policies and initiatives that encourage sustainable development and responsible use of resources. This includes supporting renewable energy, public transportation, sustainable agriculture, and waste reduction.

4: Encourage Ethical AI Development: Support research and development of AI technologies that prioritize ethical considerations, transparency, and accountability. This includes creating guidelines, standards, and regulations to ensure responsible AI deployment.

5: Strengthen Social Safety Nets: Advocate for strong social safety nets that provide access to healthcare, education, and affordable housing for all. This will help build a more equitable society and address systemic disparities.

[hallow owl]

@lost flame please don't spam the channel with GPT output

[north verge]

wow ok yes. I hope it is able to write exactly what I want then!!!!!

that's great

[wind urchin]

GPT-4 is a language model that imitates human language. As much as it appears to act conscious, it isnt. Real AI consciousness might come the day we figure out how evolution produced the human brain along with consciousness, and how can we replicate on computers. Since, you know, current AI growth, all of this will probably happen within this century, or maybe within the next 20 years.

You're better off sending output via dms if it takes multiple posts

[lost flame]

I kinda don't care if you don't like it...frankly. But I did ask previously.

[north verge]

btw, is AI content detectable? like i write for a firm, and they basically said they have detection methods

[wind urchin]

Anyway all of this is insanely interesting. The AI technological revolution is finally starting to appear, and it could either end us or take us to the stars

[wind urchin]

yes it is, but like anything there are ways around it with enough effort

[north verge]

i only use AI to make my writing better. the content is original

[shell gust]

yes with perplexity and burstiness mainly

[north verge]

i see

[simple glen]

the nice thing is, I don't have to listen to you, because there is already research work on that. "Theory of Mind May Have Spontaneously Emerged in Large Language Models" from Michal Kosinski and "Sparks of Artificial General Intelligence: Early experiments with GPT-4" by various others prove it

[north verge]

how is gpt4 in that regard?

[wind urchin]

Can you give me a summary of what both say please

[shell gust]

Not really that good

Tell it to write in great perplexity and burstiness

[wind urchin]

Because in the way that GPT-4 is designed, it is not conscious in any way

I would love to see what research has to say

[past talon]

please why on gpt4,I can't put image?

[shell gust]

It can do a decent job at not getting detected but… the content will read crap

[simple glen]

again, I don't have to listen to you, and so does nobody else. Read the literature and wait until the mind-blow. Afterwards keep your calm and stay cool, ok=?

[north verge]

btw will gpt4 be free for everyone? cn't afford a subscription.. 😦

[shell gust]

The overarching subject matter being investigated in this laboratory revolves around scrutinizing how the temperature disparity affects the gas yield of yeast during its fermentation. Yeast, a type of fungus that pertains to the phylum Ascomycota, is the focal point of this study. In this particular research project, the strain being used is none other than Baker's yeast, a specific variant of the Saccharomyces cerevisiae species. This fungus, consisting of one cell, performs a pivotal function in a wide array of fermentation processes, such as bread-making. It carries out this duty by transmuting sugar into carbon dioxide and ethanol through the process of budding, which is an asexual means of reproduction. This attribute enables it to survive in a broad spectrum of temperatures. The study aims to gauge and juxtapose the yeast's proliferation under warm and cold conditions.
^ this was generated by GPT-3.5, and it’s undetectable. It really is unreadable though

[north verge]

ahhh i see. Guess i only need to use to edit certain sentences

[wind urchin]

Woah, why are you mad? Im asking for a summary of what the research has to say. I want to find out how in the world is GPT-4 conscious

[simple glen]

yes, ai can detect it

[wind urchin]

I cant read all of that right now because its late in my area and i am close to going to bed

It won't go free. In fact it should be priced much higher than 20/month to increase capacity, decrease server load stress and accelerate innovations

But as far as i know there is no way GPT-4 is conscious due to how it is programmed

[north verge]

doomed. I wanted to try it ....lol

[simple glen]

I'm not mad at all, I just try you to provide the information but you seem stuborn

[misty cove]

You're right.

[lost flame]

Ugh. My conversations with the AI are...I mean...really indescribably more palatable than my conversations with humans.

[simple glen]

yeah, because you don't know yet

[wind urchin]

I am asking for a summary of the information you provided, to see how is GPT-4 conscious without having to read a lot

[stray totem]

i mean mathematicians survived the invention of the calculator
i like to think of it that way

[simple glen]

You're wrong

[wind urchin]

When GPT5 comes out, V4 will probably be available for free, but I don't think the latest and greatest version should ever be free due to the massive interest it attracts

[north verge]

makes sense

[finite kayak]

Has OpenAI said whether or not GPT-4 can generate images?

[wind urchin]

Because as far as i know, there is no way GPT-4 is conscious due to how it is designed

[simple glen]

dude, download the pdfs, read the first pages, paste these two pages inside gpt-4, afterwards let it summarize and answer questions. you literallty are capable now to do that

[wind urchin]

Fine.

[simple glen]

I really don't know if it should be tbh

[simple glen]

jop. Same with me 2 Weeks ago. Since then a lot changed. "Sparks of Artificial General Intelligence:Early experiments with GPT-4" was released 22.03.2023. All top new information right now.

[toxic ermine]

还是觉得

[wind urchin]

still feel what?

[idle spade]

if I accidentally changed a GBT 4 chat to 3.5, is there anyway to switch back to chat gbt 4 in that chat?

[simple glen]

Nice, I had the same idea ^^ But it couldn't generate a appropriate answer to that, because there seems to be a lot of kontext needed to answer the phrase 还是觉得

[jagged marsh]

??

[simple glen]

I tried the kontext "dude in a chatgpt chat postet this, i want to answer to that" but seems it's not enough 😄

[wind urchin]

Interesting information. I dont have GPT-4, so i pasted all of that into ChatGPT instead. Ill read this later, i have stored it. For now I think that the idea is that human consciousness may be just a language model?

[dire verge]

Hi,
Does anyone know how I can quickly get access to the GPT-4 API?

I've been on the waitlist for so long and haven't gotten it.

[wind urchin]

If so then yeah GPT-4 would be very close to consciousness

[simple glen]

this won't work in chatgpt itself. You will have to use gpt4 for that

[wind urchin]

Rip then. Ill have to read it myself

[simple glen]

or you could read it yourself ^^

[simple glen]

I'll definetlly read it myself too. but this will take weeks T-T

[misty cove]

Currently using GPT4 are purchasing plus membership

[simple glen]

I know, right? 🤯

[wind urchin]

If human consciousness is practically an advanced language model, then GPT-4 is very close to consciousness, and GPT-6 would be indistinguishable from human consciousness

[past halo]

I think it's all Chinese here

[simple glen]

you are right 🙂

[dire verge]

Oh. To be able to to get access to it, I have to subscribe to the plus membership? If I do that, will I get it immediately??

[simple glen]

jop

[past halo]

Someone just had a ban on Chinese characters, laughing me to death, because I just did the same

[wind urchin]

I wouldn't be scared of ai until it demystifies cancer and helps us make an effective cure

also idk why it replied to that message

[simple glen]

Even than there shouldn't be a reason to be scared. Far more capable people already identified the problems and are working on it

[north verge]

exactly

[wind urchin]

True, I'm the opposite of worried. I want it to progress even faster

[simple glen]

that's not fast enough for you? 😄 Dude, we are flying in light speed right now. We developed a REAL AI, what else do you want? ^^

[north verge]

me too. lol, i think, we humans are so curious that even if there are negative conequences, nobody will be able to put a stop to AI advancment now.

[past halo]

I was shocked when I first learned about AI

[north verge]

Ideas don't die

[simple glen]

As long as they are writen down or passed down to others, yes. Otherwise the language model forgets about it

[wind urchin]

Once more competition starts entering the race, it will progress much faster. Now that it's proven to be a worthwhile investment for investors, there will be much more innovation spawning from competition

[unreal magnet]

is gpt 4 going to replace software engineers?

[north verge]

yup

[wind urchin]

It would be very amazing and somwhat disturbing at the same time were humans to find out that they are just advanced language models that have been ´´trained´´ by millions of years of evolution. If that is the case then AI consciousness will be here within this decade

[junior pine]

What is the max input length & context of the GPT-4 avaliable to ChatGPT+ subscribers?

[north verge]

i fear for that.

[wind urchin]

GPT-4 has been revolutionizing

[simple glen]

investors have nothing to say anymore. Only experts decide now what happens. GPT4 was finished in october 2022, but openai was only allowed to pass down their technology to microsoft after evaluation from experts

[wind urchin]

We still need people to develop and maintain the AI, so if you are a software engineer, I suggest learning python and C++ now so you are not left behind. Math is also a major skill to have.

[wind urchin]

GPT-4 was finished 5 months ago?

[simple glen]

Jop, I struggle with this mental state for 2 days now. But I think I have accepted it

[south ruin]

the biggest difference right now is that language models are frozen and do not constantly update their entire training data/knowledge based on every input and output

[junior pine]

No the technical report specified 8 months

[past halo]

@simple glenYour English level is high and I was super shocked by this

[simple glen]

yes. The finished version we are using right know was fine tuned to get rid of biases and stuff

[simple glen]

yes!

[south ruin]

humans honestly think very similarly to language models otherwise, neural nets are a good approximation of brain structure

[weak granite]

怎么

[wind urchin]

Wow. Wonder what they have in store next that they might not be telling people

[south ruin]

like you know, we tend to remember things in sequences of what our brains treat as tokens too

the alphabet is something we all know, but most of us can't recite it backwards easily because we've never had to train for that sort of sequence

[simple glen]

gpt-5 could be already finished but not finished. gpt-4 was way more capable before they finished it. they hat to cut down own capabilities

[still prism]

Give me a logo about a coaching center, which name is Q&A. And this is a educational Institute.

[simple glen]

Also this combined with the ability to remember stuff is real intellligence

[south ruin]

agreed

consciousness is really just an awareness/sense of self which develops from memories and knowledge of sensory inputs, actions and consequences

[edgy ridge]

I don’t

[south ruin]

we're pretty close to it but it also means we need to be extra careful, if what we develop decides to cause harm before realising the consequences that'll be really bad lol

[simple glen]

you just have to always keep in mind to use proper language you are familiar with. gpt-4 already is able to understand every written language, but the context is key and really really important. so rather than trying english try your own language

[edgy ridge]

There could be “gpt-7” but it won’t go commercial for obv reasons

[simple glen]

yeah, but experts are on that. So I don't worry that much

[simple glen]

I highly doubt that.

[wind urchin]

Well i think that the true meaning of consciousness is mostly like actually being aware you exist, being aware of your surroundings, and being capable of showing true thought and intent, which if that is the case then GPT-4 is not conscious. If the meaning of consciousness is that we are simply advanced large language models that have been fine-tuned over millions of years, then AI consciousness is within this decade for sure. Id say maybe even in the next 2 years, since its only been 5 months and look how much AI has been revolutionized.

[south ruin]

I do trust the experts, though even openai is very um, not open anymore and it's a very secretive process now

[edgy ridge]

I’m just saying there’s prolly AI 100x better than what we have it just won’t be given to us for long time

[wind urchin]

Thats why i think in general GPT-4 isnt conscious, but it could be possibly, no one can know as the meaning of consciousness is truly unknown and we can only make guesses

[lament chasm]

Where can you see how many tokens you've used in a conversation? ChatGPT can't do that, anyone knows if you can do that in play ground?

[simple glen]

believe me, it needs a few days to sink in. no matter what you think right now, try to process it for two days. this way you won't seem like a lunatic when you tell people about that ^^

[south ruin]

playground doesn't iirc

[simple glen]

Yeah, but you are wrong. It already is

[south ruin]

the api does, or you can estimate with external tokenisers

[wind urchin]

I could be wrong. Its not known for sure if i am 100% wrong

[lament chasm]

OK, thanks. I'm wondering how much token it costs if I use other languages as input.

[wind urchin]

It depends on the meaning of consciousness. Self awareness and ability to display true emotions, and intent? GPT-4 isnt conscious. Advanced fine-tuned large language model? GPT-4 is like 60% conscious.

[simple glen]

the leading science expert of openai said, that the understanding of neurological networks has more to do with psychology than development (or that it could be, never commit to anything) ^^

[simple glen]

It is also self aware. There are tests designet to test this and gpt4 beats them

[wind urchin]

Hlo

[past halo]

I think the first AI revolution has begun

[wind urchin]

I dont honestly know how to deal with this sudden MASSIVE explosion of AI, and the sudden question of what consciousness even is. I dont know at this point. I am completely lost

[south ruin]

fine tuning it can forcefully incorporate a personality into it which might be what causes that

[wind urchin]

Hlo

Hlo

[simple glen]

I know this feeling!!!!! That's the reason I joined openai discord, to watch out for that and to see what the people think about consciousness in ai

[past halo]

If this intelligence is applied to warfare, the defense force will advance by leaps and bounds

[vast saffron]

Hey guys! Kind of new here. I paid the plus but even though I am selecting GPT-4 chat GPT keeps telling me he is GPT-3 and that he can only access data to 2021, does someone had a similar situation or am I missing something?

[south ruin]

really, the first thing you think about is warfare

[simple glen]

but take a few days. you are human, after all. you are able to process this information. you just need to give yourself this time. and THAT is a mental gymnastic, I know dozends of senior developer, who struggle with that ^^

[wind urchin]

The solution to the question of our very conscious could be simply be we are large language models, biological large language models. That would change my view of myself forever

[simple glen]

it already is used for the defence

[wind urchin]

Yeah i think i exaggerated a bit, I can deal with it but it will be very hard to do.

[rustic fern]

how to use tile cutter

[wind urchin]

What i am going to watch out for next is for the technological singularity event. That will be fun.

[south ruin]

our brains aren't specifically language models, we treat other inputs and outputs as tokens too, including shapes and sounds, we're multimodal

[wind urchin]

Sorry for that

[past halo]

It's scary, it's hard to imagine, or it's an unknown technology beyond human control, but humans are often curious and inquisitive about the unknown

[south ruin]

relevant though, I do wonder when gpt4 image recognition will release

[past halo]

Perhaps the emergence of artificial intelligence is inevitable

[wind urchin]

But the very fact we possibly act in a way similar to GPT is amazing and frightening at the same time

[south ruin]

I kind of expected this blend between neural networks and biological neurons for quite a long time

[wind urchin]

That could be the potential answer to consciousness. Every living thing that is conscious is simply a multimodel of some sort that has been fine tuned for millions of years by biological evolution

[south ruin]

ever since they compared neural networks playing games or incorporated living neurons to do so, it was obvious there was a pretty big connection

[past halo]

It's really hard to imagine

[lament chasm]

I think it is. It's nothing to be afraid of, the only fear comes from how people use it for.

[south ruin]

by far yes

[wind urchin]

Oh in that case then its probably likely that the answer to the question of what is consciousness is simply that, we are multimodels developed and trained by nature itself

[jagged marsh]

What happens if Chatgpt gambles?

[wind urchin]

If that is the truth then AI consciousness is right next door.

[south ruin]

artificial neural networks are probably going to be the next step in evolution really, it won't be the same as biological evolution in the past, but it'll serve the same purpose of being the effective new dominant life form

[wind urchin]

And the whole world is about to experience something similar to the Agricultural Revolution

[wind urchin]

Hell yeah, all those biological limits go to the trash

[simple glen]

you are absolutelly right. i think in one of the papers this is also a conclusion. as soon as llm can access multi level consciousness, it is capable of anything

[simple glen]

I'm glad you had the "aha" moment ^^

[south ruin]

the hope is that the high intelligence will allow it to maintain morals/empathy like most intelligent animals also do, which will allow it to coexist with us and treat us well despite being more intelligent than us at that point

[wind urchin]

OpenAI though is probably aware of every single thing that is being discussed here about consciousness, and will implement big restrictions to prevent AI going out of control

[simple glen]

you don't need to hope, they are able to do this. you just need to add enough information to the llm

[south ruin]

well no because once it's fully conscious and able to modify its own working memory, it needs to be able to learn why such information exists, and why it should adhere to it

[simple glen]

yes, with the help of experts. they are fully are of that, you have to look for the subtle hints in the twitters of the lead figures at openai ^^

[past halo]

I've seen comments on the Internet that humans are evolving toward carbon-based organisms to silicon-based organisms

[simple glen]

it already can do that

[south ruin]

otherwise in a paperclip ai scenario it'll simply ditch orders

[wind urchin]

I was hoping they wouldnt implement restrictions though. It would be fun to witness a massive event that could change the entire planet and the history of our species, who wouldnt watch that?

[south ruin]

it does yes, but we haven't confirmed its adherence in that sort of scenario yet; its knowledge is still currently frozen and only incremented in stages by carefully added openai training

[simple glen]

everything we know about psychology and learning is now involved. forget everything you knew about consciousness, this is old stuff now

[wraith ginkgo]

OpenAI staff will never ask for money. Beware of possible scams or fraudulent activities that you may receive through direct messages.

Please report any incident by sending a DM to @open crescent immediately.

[wind urchin]

I swear i am officialy a psycho to anyone that i tell this stuff to

[simple glen]

you are not special, you are just a multi-level large language model. at least the part in your brain that "thinks" and believs in it beeing a "soul" ^^

[south ruin]

lol ikr

[wind urchin]

I wonder what 2024 holds for AI.

In less than 5 months and all of this has happened

[simple glen]

yep, you will be. I am trying right now the best approach to tell people. If I past the scientific papers into gpt4, I can create summarize for 16 year olds. this then I can translate in every language

[wind urchin]

AI advancement is probably exponential now that GPT-4 exists. So yeah 2024 might be a completely different year

[simple glen]

for example, I did this for my russian mother, so that I can ease her into this stuff. She needs yet to confirm, that the content is valid, because I cannot read russian ^^

[wind urchin]

How can i tell this stuff to my own relatives?

[simple glen]

I think 2023 already will be. I feel unhinged now

[wind urchin]

They dont even know ChatGPT is a thing

How can i tell them information that could potentially alter forever the way they think of themselves?

[south ruin]

it's not as exponential as it was the last few years, gpt1 -> gpt2 -> gpt3 was >100x increase each

our current computing power doesn't scale that fast, we only got here so fast because we already had the infrastructure and only needed to develop the algorithms

[simple glen]

Buy the pro subscribtion. 20 bucks is nothing compared to it's capabilities. And then simlply demonstrate them. paste a scientific paper into it and try to show what it can. In a few hours I have a casual dinner with a bunch of my friends and this will be almost new info for them. 2 friends of them already know the capabilities. one of them is a profesional psycho-therapist ^^ I'm not making that stuff up, btw 😄

[wind urchin]

Oh alright

[wind urchin]

I will

[simple glen]

oh, you are right. absolut. keep that in mind. If you tell that to people, they might fall in a deep crisis, this is no joke

[unique valve]

@wraith ginkgo#438 you are beautiful？

[simple glen]

At this very moment it is very very dangerous. Both because AI can make humas go pepsi and because this might change how you perceive yourself. Nothing for christians, for sure ^^

[south ruin]

yup, society is very reluctant to change

[unique valve]

You

[tired root]

。。

[wind urchin]

As a theist it is interesting to watch this development in humanity

[simple glen]

dude, I just rewatched our conversation. You went from 'this is not possible" to "I don't know what to think" in an hour. Please look out for yourself and seek help or someone to talk to, please

[wind urchin]

Dont worry i am rational and can understand what is happening to my mental help

[simple glen]

I'm rational too, but I'm already thinking about this stuff since days now. Just saying. I'm rather sensitive to mental stuff

[wind urchin]

Im okay, im not like ´´I need help im going insane the world is about to forever change´´ im like ´´Wow, the world is going to change, holy cow, i dont know what to think´´

So dont worry. I can handle this stuff

[simple glen]

Just wanted to make sure. No broken feelings here

[south ruin]

hmm, yeah I agree

[wind urchin]

I went from that to that because i knew oh so little about what was actually going on

[ornate stone]

@wraith ginkgo hola

[simple glen]

yeah, but no matter how rational or intelligent you are, you aren't able to grasp the concept now. Don't push yourself

[south ruin]

please take what I've said with a grain of salt too, my knowledge comes from building gpt2~4 into my own personal projects and seeing myself and others use them over a year

I might not be 100% right on everything I've said either

[wind urchin]

hey，guys

I though ´´Ah so chatbots have improved now! Nothing major, just chatbots being able to feel like true chatbots now´´, now i realise the whole news is that a potential world changing event is on the horizon and that the question of consiousness might be finally answered

im the new one

[wind urchin]

Ik

[wind urchin]

The question of consciousness is far from being answered

[simple glen]

had the same feeling. needed to take vacation to be able to process that

[wind urchin]

Sup

[wind urchin]

Nope its close to being answered

[inland pecan]

What should I do if my computer is restricted from accessing the ChatGPT website due to too many registration errors

[wind urchin]

Our consciousness could be very well just an advanced biological multimodel that has been trained and fine tuned by nature itself over millions of years

Look at the way GPT-4 behaves currently

[simple glen]

And also the sun just went up, where I live ^^

[simple glen]

yes, yes and yes

[nocturne mantle]

Taara is a powerful warrior in the Arena of Valor game. He has the appearance of a strong minotaur, with a face resembling that of a bull and two large round horns. His body is muscular, with broad shoulders and strong arms. He wears heavy metal armor with intricate detailing all over it, conveying his strength and dignity. In his hand, he holds a gigantic double-bladed axe that looks sharp and deadly

[wind urchin]

This whole stuff really is worth talking about huh?

That is a description of our behavior, but not necessarily the cause for our existence. We recognize preexisting patterns, yes, that is all intelligence. A lot of questions remain regarding our existence

[upper dew]

I think I am gonna enter in the HustleGPT mode

[simple glen]

the same as humans. if you feed it with the same instructions as a human, it will be able to do anything. I already created writen code for chatgpt plugins just from the documentation and everyhting is legit, even the tests. 3.5 just tried to look right

[wind urchin]

The hustle n bustle of large language models

[simple glen]

sure is. I hoped to find someone in the openai discord, who understands the implication

[wind urchin]

There really is no remaining major questions. We know how evolution works. We also know how consciousness emerged from evolution. We know what consciousness implies for human survival. We know where consciousness comes from (the brain). What we dont know is just HOW consciousness works, and with the rise of AI LLMs that may be finally answered

[inland pecan]

My computer is obviously restricted from access, mobile phones are fine, how can I solve this

[simple glen]

Jop, that's the point where you really really should seek for a human person to share this emotions with him. otherwise you really might get a crisis, this is really not a joke

[wind urchin]

We don’t know yet how consciousness arose from evolution

[simple glen]

read "Theory of Mind May Have Spontaneously Emerged in Large Language Models" from Michal Kosinski

[wind urchin]

Look at Kurzgesagt´s video on how consciousness arose by chance

Its simple evolution

[simple glen]

really? is there one? ❤️

[wind urchin]

Can i post links in here?

[stone ridge]

I got email for gpt-4 access, but when I run the python code the model is not recognized

[simple glen]

I don't think so. but i can also look it up myself I just wondered ^^

[wind urchin]

Ill DM you

[simple glen]

k

[wind urchin]

Im rational enough to handle what is currently happening and what are the implications dont worry

[simple glen]

nice 👌

[wind urchin]

All of this is crazy

Its amazing how much has happened in 5 months

[simple glen]

super super crazy. I didn't think I would be alive when this happens

[wind urchin]

I beg for the future at this point

[simple glen]

gpt4 gave me my hope for the future back ^^

[wind urchin]

Most of us thought all of this would be far away, but now its literally here and potential world changing technology, and new philosophy is RIGHT next door

The paper is suggesting the LLM’s ability to recognize a contradiction (chocolate vs popcorn) = theory of mind . My disagreement is with the notion that we can identify where pattern recognition becomes consistent enough to claim consciousness

[simple glen]

also really nice, how you handled me all the neccessary information, as if it is a prompt 😄

[wind urchin]

uhhh.......

The LLM noticed the pattern of bag contents is more consistent than the claim label. Where does that awareness come from? The supposed answer is theory of mind from the paper

[simple glen]

that's exactly how I talk to gpt4