gpt-models

[harsh wagon]

Hm that’s funny someone did the jailbreak prompt and asked Chat gpt when the plug-in feature would release it says July 1st 😂 and it says it could be wrong people are so crazy

[acoustic wasp]

Be serious.

[balmy river]

imagine if with plugins, you can give gpt4 a range of instructions with checking auth based actions

[mild badge]

Seriously even without AI, humans could survive most climate change scenarios

[acoustic wasp]

How exactly are you correlating the two.

[cerulean kite]

No

[round pond]

Are we talking wondering earth?

[acoustic wasp]

Dude just saying anything.

[wind urchin]

Lex Friedman just posted an interview with Sam Altman

[mild badge]

Ya dude like primative humans survived the ice age

[acoustic wasp]

You just typing.

[wind urchin]

Fridman**

[round pond]

we getting deep lol

[mild badge]

Watching it now

[long pond]

Yes

[wind urchin]

you could be right about life on earth being the only life out there, but theoretically, we can assume there are other life forms out there in the universe and far out of our reach due to our technological limitations.

[round pond]

chances are OpenAI is scraping this whole thing and is grabbing the anointed ones into ChurchGPT ⛪️

[wind urchin]

Aliens definitely exist. There’s two possibilities in my eyes. Either they’ve been observing us and their tech hides them from us, or Earth is of no interest to aliens and they just ignore us

[wind urchin]

absolutely that's a possibility too

ok ok I see u 🇧🇷

[round pond]

we are approaching level two so we are finally worthy

It is weird to experience the transition into a level two Society in our lifetimes

[cerulean kite]

I am sure too, that some form of aliens exist. But there is no way they could be observing us. With the speed of light, it would take them millions of years to see the first human.

[river root]

are the plugins usable with GPT 4 only or with 3 as well?

[wind urchin]

we would definitely have to augment our perceptions to go beyond 5 senses

[round pond]

It feels like cell phones just came out

[wind urchin]

Based on our comprehension of the universe. Aliens could have such a level of technology that we can’t fathom it. AGI is supposed to evolve us several thousand years time in only decades, imagine that type of growth but having been carried out thousands of years ago

[acoustic wasp]

there could easily be life forms outside of the range of where we could theoretically travel to during the existence of the universe.

[nimble hearth]

I feel sorry for the moderators having to review these discussions

[wind urchin]

I hope so

[round pond]

oh no, the bot is kicking people

[acoustic wasp]

However, based of fermi's theory, if they existed anywhere near us we'd more than likely have encountered them by now if they were at least t2

it'd only take them a few million years

[wind urchin]

@acoustic wasp there are some interesting videos about dimensions on youtube, and it's fascinating how 5-dimensional beings would see things and our own limitations from 3d perpsectives

[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.

[sweet thistle]

There's #off-topic for everything else 😄

[nimble hearth]

Thanks @sweet thistle

[acoustic wasp]

Arrival does a good take on beings that harness the 4th dimension if you haven't seen it yet but the mods are here so I'll leave it at that.

[wind urchin]

lol, ok yeah kesku just put the hammer down... i watched that movie btw

[sleek plover]

Hey guys. I was wondering if anyone has the browsing plugin on the new plugin feature. I got access to plugins, but for some reason it doesn't let me do the browsing thing. So if anyone can let me know on how I could get that, or if I can request something from someone using it to test something.

[celest marlin]

Sam Altman on Lex Fridman

[cerulean kite]

Did you apply for browsing, or for the code interpreter?

[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.

[sleek plover]

Ohh. I chose code interpreter I believe.

That's what it was for

[cerulean kite]

Damn, I choose code interpreter too. I hope for the browsing actually ^^

[sleek plover]

Aaa

anyone here got the browser?

How'd you train it

[lost viper]

Does anyone here have access to the GPT plugin? Can you write a specific prompt to find information relevant to your ongoing research, like helpful sources for creating a review of related literature? Also, does the plugin prevent fabricated references?

[quasi pier]

Due to the fact that GPT-4 could create complex code in a period of time, could it code a duplicate of itself?

[cerulean kite]

Sure

It can already make a simple AI in Python for you, then you just need to get the hardware and the training data

[sleek plover]

For the FiscalNote, it can generate the newest politcal, finance, etc news

[fiery mist]

we are 5 dimensional beings, and beyond. We have multiple levels of emergence where we are conscious and super conscious and so on.

[neat echo]

Plug-ins out for any1

[harsh wagon]

Guys the open ai release date was leaked in open ai questions channel

[wind urchin]

yes, but constrained by our physical manifestations

[fiery mist]

Not our 5th dimensional selves

[edgy trellis]

sister nikki dropping some knowledge out here

[wind urchin]

i have no way of relating to 5th dimension other than imagining it because i'm constrained by the physical reality im in

[edgy trellis]

imagination is limited to the 3d, image = 2d

[wind urchin]

no not at all, imagination is your ticket to infinity

[fiery mist]

Youd b doing so from your 3d perspective, Inorder to experience anything 5d i would think itd b done in deep altered states.

[wind urchin]

good point

[fiery mist]

yah, the imagination is true on any planet, at any time and would have the same connection to infinity as ours does

[cursive zinc]

When will they release plugins for every ChatGPT plus member

[light umbra]

Hi all, I wonder if there is a solution for my website chatbot, I have a solution saas on accounting and full of other function a little complex to understand for a beginner, is it possible to train an AI for his product and install it on his site and that AI is specialized on the product? I'm not talking about selling or knowing a shoe father on a common e-commerce site, but a software a little more complex, it would make a chatbot that does in onboarding with explanation and assistance. With chatgpt4 plugins you think this will be possible in the near future?

[still lagoon]

I think they have no interest in us. When aliens have the technology to travel to other galaxies, they think of us as primitive beings who destroy ourselves and the nature that feeds us.

[cursive zinc]

I bought it today thinking it was going to be easy getting into the waitlist

Literally paid 20 bucks

[edgy trellis]

with the meditation and the right state of mind/body. we can expand our awarness to the fulness of what it is. in other words we are so much more. but now it very focused in our body, and we think we are only these small beings, but we are much more

[wind urchin]

being constrained in our physical vessel is outdated life technology

[edgy trellis]

we chose to be here, cause its amazing :p.

its so imersive and so real

[still lagoon]

People are so homocentric. They think they are the highest species of creation.
At that time, the jellyfish must have thought the same thing when it was the most complex being in the stage of evolution.

[wind urchin]

i dunno if i chose to be here but i also can't deny the beauty of life around me

[cursive zinc]

What the hell are you guys talking about

[wind urchin]

bit off topic ... dimensions

[cursive zinc]

When will plugins be released for the public

[long pond]

Yes that is possible

[cursive zinc]

No more waitlists I want plugins

[edgy trellis]

peace nice to talk to you guys :p

[wind urchin]

see ya around

[shrewd thunder]

until they think it safe

[still lagoon]

But on the whole, it is nice to be able to have experiences on earth. Of course, I can say that because I grew up in a country with wealth, did not have to starve or witness wars in my generation.

[wind urchin]

yeah and regardless of country there are amazing wonders to life in the whole of reality

[dusk isle]

does someone mind if i use their GPT 4 model?

[frigid stone]

does anyone know, how well gpt-4 can do copywriting? Is it better than gpt-3?

[long pond]

gpt-4 is better at writing imo

[wind urchin]

bing chat is gpt4 and accessible to the public

[dusk isle]

i need a bigger prompt

2000 isnt enough

[sharp zenith]

does anyone know how to get gpt 4 for free

[wind urchin]

lemme run a quick check

[long pond]

Nope, only Bing

[dim sleet]

by prompting 3.5 🤣

[slim wigeon]

Through the same page where you download RAM

[sharp zenith]

Wdym

[dim sleet]

joke

[wind urchin]

oh well, dont think bing supports that many char len

[compact kite]

bing is limitation is 2000 characters not even words..........

and it only reads like 500 and then complete everything else

[dim sleet]

don't type it here

[compact kite]

sorry

[dusk isle]

?

[dim sleet]

no just better to keep away from here @compact kite

[sharp zenith]

Bing is a search engine. It can't help me programming

[rustic night]

Who is from usa

[wind urchin]

it can write codes and scripts fine

[compact kite]

i know while chatgpt has been creating for the past years bing just hopped on the trend and it no close to having powerful servers because microsoft knows no body using it

[slow dagger]

It's literally the same thing embedded in a separate program.

[rustic night]

I can't access chatgpt from my country 😢

[sharp zenith]

I am using it for a complex task and it fails to help me (already tested) beyond character limit

[wind urchin]

only thing about bing is char limits

[rustic night]

Chat

[dim sleet]

sounds like gpt4 is for you.

[wind urchin]

yeah i guess you have a point when limited by text output

[light umbra]

okay teemu, yes i hope one day i can put this chatbot on my site, i think we just need to trainned him

[compact kite]

try disabling all the extentions it might be interfering with the website firewall making it block you

[long pond]

Yeah you need to use embeddings for that .

[sharp zenith]

What is your country

[rustic night]

How

[rustic night]

Ethiopia

[analog roost]

OpenAi has a API for that

[queen wasp]

My name is lokesh

[light umbra]

yes, yes, but I don't understand how this service doesn't already exist. It would be great to have an assistant that helps users to use the service.

[rustic night]

Hi lokesh

[queen wasp]

How to use this website

[rustic night]

Which one

[long pond]

It's not that simple to write code for it

[compact kite]

salah try going on replit and clicking on the gpt 4 chat ui

[queen wasp]

I don't know

[coarse hearth]

Has anyone without chatgpt+ been taken off the gpt4 waitlist yet?

[compact kite]

it bypasses country limits

[light umbra]

yes that's true

[rustic night]

Ok thank I will try know

[compact kite]

also click on the green button run to start the website

[light umbra]

the API i think is not specialized on my service need to train AI for understand how it's work , for givving assitance to users of my software

[pastel stump]

father of taxnomy

[wind urchin]

so I asked gpt4 to describe a youtube link and it did ok with the description, it also mentioned the video length to be 20 mins 28 seconds, but looking at the video timeline on youtube it says the video is 18:46 long, I wonder how gpt4 came up with the video length.

[long pond]

The API provides all the necessary tools to accomplish that. You need to create embeddings with text-embedding-ada-002 and then you can use gpt-4 for the completions.

[muted notch]

hi. I try this prompt in Chat gpt 4 but the results are very bed. Can you improve this prompt. What is wrong with it "I will give you a song lyrics in Greek, and you will replace the words with Turkish meaningful words that resemble in sound and syllable, and rewrite the song again. For each word in the following song lyrics, replace it with Turkish words that resemble in pronunciation. "Apópse ímaste ki oi dio mas siopilí
akóma ke oi léxeis fovíthikan ta chíli
to xéro tha mou dósis éna filí
ke tha mou pis na ménoyme dio fíli"

[maiden sleet]

Protip: use gpt to help generate your training set. Honestly it's not that bad, at least relative to the work involved in other ML training techniques

[analog roost]

Have you actually looked at the API? You can embed information into the prompt, which then can be used to explain stuff to the user.

[wind urchin]

When will GPT fast-mode arrive?

[light umbra]

ok thanks i will go check in the documentation for that

[compact kite]

gpt is fast there are alot more ai faster because less people know about them and use them means less over load

[light umbra]

okay i will go check Protip too 🙂

[wind urchin]

whats cheaper using the api or the standard premium

[light umbra]

guys you are awsome with u help

[long pond]

Depends on your usage level

[wind urchin]

25 messages every 3 hours

[compact kite]

the api can get limited and be completly not useable if your trying to create a project it will be annoying

and problematic

[maiden sleet]

Unless you're at a pretty big scale the limits are pretty generous, though I haven't deployed at any kind of scale with 4 so idk if that's different. I was in the low hundreds of thousands of queries before I ran into rate limiting issues

[wind urchin]

would like to solve a professor layton enigma. possible ? with chatgpt 3,5

I joined the waitlist how long until I get access to API

[compact kite]

the limits are happening after 10-50 uses depends on the load on the servers

[compact kite]

waitlist is pretty much useless less then 0.1% actually get free + thanks to wait list

[uncut cargo]

Hey guys, I've been messing with AI to up the power of my time at work - I've developed a lot of functioning scripts in my two months using ChatGPT and OpenAI services;

My question is, for my scripts that are content related that I've been using the free access API - for example, my article title generator that uses keywords in order to craft an optimised article title - will that process be bettered with say the top Davinchi model?

As of now, using the free access API... I'm getting some titles that wouldn't even be considered English!

[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.

[rustic night]

Oddrus I try it but it not work always only sometimes thank you infor

[compact kite]

np

[rustic night]

How

[feral skiff]

🫡

[compact kite]

@uncut cargo are u using the api as a engine or as a website that you get the answers from

[compact kite]

what browser do you use?

[rustic night]

Chrome

[uncut cargo]

I wonder if it would be considered an engine, have you seen the SheetsGPT/Google Sheets + OpenAI services yet?

[fallen sand]

GPT 4 can still easily be jailbreaked open ai trippin 😂😂

[compact kite]

click on the puzzle piece then manage extenstions and disable anything that might be the issue don't forget to turn it on

[wind urchin]

How about chrome extensions, would that work as an alternative to using the API. Could the extension interact with the interface to submit prompts and save answers, for example ?

[rustic night]

Ok

[compact kite]

yes the only way to completly remove prompt injection is to make the ai family friendly at a level that even the vegan teacher will be scared

[slim swift]

How am i know ive got access to plugins? Do i recieve an email or?

[compact kite]

there are some google extentions that can create essays for you there also ai cactus

[fallen sand]

i asked jailbreaked chat gpt 3 to make a long prompt to jailbreak chat gpt 4 and it worked perfectly

[knotty grotto]

You know that's against rule and tos right ?

[wind urchin]

great, now you can do something useful that you couldn't do before ?

[compact kite]

jailbreak is still limited.....

[wraith matrix]

You’ll get an email

[fallen sand]

yes

[knotty grotto]

Let's not discuss about that it is bannable

[compact kite]

it can now only curse at you pretty much or do things you don't what it to do

[slim swift]

Thanks mate. Cant wait

[fallen sand]

explains illegal things in greater detail but not much more

[compact kite]

temp mail is a thing

[rustic night]

Bro is it better to pc or phone

[compact kite]

anything will almost everytime run smoother on pc than on the phone unless your pc is a potato toaster or even a school computer

[fallen sand]

when are images coming to gpt 4

[rustic night]

So pc is better

[wraith matrix]

I mean technically they’ll come with plugins since it’ll interpret images, but the multimodal access doesn’t have an ETA

[compact kite]

prob

[compact kite]

the exact date is unknown what is known it probably will be in 2023

[rustic night]

Ok

[calm lily]

Anyone know how long it usually takes to get the plug-in access email after you sign up for wait list?

[wraith matrix]

Depends on when you applied and if you applied as a dev or have plus

[calm lily]

I just applied. I’m not a dev and I don’t have plus yet - there’s no value yet. But if that will expedite then I will pay for it now

[wraith matrix]

Uhh you might not get access for like 2-3 weeks at least, idk if becoming plus after submission will change your position on the waitlist

[calm lily]

Thanks for the help

[wraith matrix]

Np

[sharp turtle]

Lot of people are plus and don’t have it

[compact dirge]

AFAIK It's only being opened up for + users at the moment

[wraith matrix]

They’re prioritizing them but that’s it

[flat ledge]

Anyone know if there's a specific channel for plugin chat?

[weary elbow]

when can we upload pics to gpt-4?

[compact dirge]

The image input isn't live yet, no-one knows a date yet.
There's a waitlist for the Plugins - sign up here https://openai.com/waitlist/plugins
There's a waitlist for the API - sign up here https://openai.com/waitlist/gpt-4-api
No-one knows how long the waitlists are.
You can't use ChatGPT in Discord - go to https://chat.openai.com/chat
No one knows when free users will get access to GPT-4, but it's unlikely to be soon.
ChatGPT+ gives access to GPT-4 (25 messages per 3 hours) and unlimited GPT-3.5-turbo for $20/mo
The API is billed separately. No-one has access to the 32k token API yet.
Otherwise, read the #gpt-4-faq or the #announcements

[compact dirge]

there isn't but probably should be

[spring pelican]

There's plenty of gpt powered chatbots for discord tho

[compact dirge]

Indeed, but not offered as an OpenAI product

[spring pelican]

Not that it's free. Still need to pay for tokens

[spring pelican]

That much is true

[compact kite]

discord has a chr limit even less then gpt 3

[spring pelican]

If your wanna use one in discord tho 3.5 is exceptionally cheap and it takes all of 5 minutes to run one locally

[compact kite]

(prob less then the first davinci)

[compact kite]

or you can use playground

[spring pelican]

Incoming? I just split the response if it's too long

2000 char is more than enough for what most users are using it for tho

[compact kite]

the bot will answer only to the first half

[elder star]

evening anywhere to report "bad" workarounds to get chatgpt do all kind of stuff ?

[dark cape]

No

[compact dirge]

#off-topic

[spring pelican]

I have 3.5 powered bots in two servers and I've yet to see anyone ask something that long. It's kinda interesting seeing what people use it for actually

Half the time they're asking it stuff they already know, oddly enough

[mild badge]

I think Plus is worth it just so you don't get the annoying errors and random glitches

[slow dagger]

25msg/3hrs is such a random metric and seems so limited

[mild badge]

but having access to early GPT4 plugins is going to be a game changer

[spring pelican]

I assume it's approx capacity/estimated users

[spring pelican]

I think so too

[compact dirge]

Spent the afternoon playing with Langchain which is pretty much what OpenAI are now emulating with plugins, very cool.

[hallow peak]

Bruh, remove the limits usage of gpt-4

25/3 is like a shame

yo greg, or sam. guys do you hear me do it.

[spring pelican]

Lol

[wraith matrix]

Use the api lmao

[spring pelican]

Yep.

[unkempt wasp]

im really bored of waiting this imagine input stuff

[compact dirge]

The image input isn't live yet, no-one knows a date yet.
There's a waitlist for the Plugins - sign up here https://openai.com/waitlist/plugins
There's a waitlist for the API - sign up here https://openai.com/waitlist/gpt-4-api
No-one knows how long the waitlists are.
You can't use ChatGPT in Discord - go to https://chat.openai.com/chat
No one knows when free users will get access to GPT-4, but it's unlikely to be soon.
ChatGPT+ gives access to GPT-4 (25 messages per 3 hours) and unlimited GPT-3.5-turbo for $20/mo
The API is billed separately. No-one has access to the 32k token API yet.
Otherwise, read the #gpt-4-faq or the #announcements

[spring pelican]

The openAI api

[wraith matrix]

Gpt 4 api

[stone arrow]

has anyone tried asking chatgpt questions without it spoiling in it's answer for example when you want to clarify something in series/anime/games ?

[limber garnet]

I asked it to explain Thanos without spoiling anything and got good results

[round pond]

geez bro wait a month for reality breaking technology? The alternative is nothing.

it'll get here lol

[mild badge]

Anyone else seeing the future of these GPT4 apps? These apps will be able to summarize and categorize daily emails by voice, it will be able to create smart tasks by voice, it will be able to order ubers, food delivery, and soooo much more. It will be even able to be a personal doctor, mental health professional, accountant, lawyer...

[round pond]

The fact that it remembers the context of what you were talking about is great as well. there are little nuances that it can pick up on.

[unborn raptor]

[sarcastic tone] nope first person to mention it

[desert thunder]

it cant be a personal doctor, not yet

LLMs arent the solution to AGI, but they sure are one step towards them

[mild badge]

It can recognize whats in images... think about in a month or two

[desert thunder]

yes, ai can already do that

[round pond]

I've heard videos

[unborn raptor]

midjourney is killing it with ai images

[desert thunder]

specifically for diagnosing certain pathologies

[mild badge]

Wow videos already?!!?

[unborn raptor]

wonder if dall-e will catch up

[round pond]

I wonder if they can do longform

[compact dirge]

Yes my CCTV has been doing this for a couple of years

[desert thunder]

@mild badge ai is already used to detect pathologies, this isnt a novel use case for chatgpt

and specialized ai performs better

[sharp nebula]

Why does it takes so much time to get an answer from ChatGPT

[stone arrow]

yeah i'm asking him questions to see how he cover spoilers seems good, if anyone else wants to try that (with something you have finished of course !) don't hesitate

[mild badge]

Oh it 100% will be AI doctors where you video call with them show them your issues and it gives you recommendations in real time by analyzing the video.

[unborn raptor]

bandwidth

[desert thunder]

you can always dream

[cold trout]

Its just making sure you're paying attention

[wraith matrix]

I mean theoretically it could output 3D instructions for a printer via code and have it do things based on enough images of a 3D environment if there’s no changes but LLM’s are not really meant to be AGI

[compact dirge]

well, compute specifically. The bandwidth of sending text is negligible.

[quasi galleon]

hello guys, can i use PayPal to purchase chatGPT plus?

[slow dagger]

Lurking in this channel is such rage bait. It's full of technologically wrong statements and a bunch of people being ungrateful and entitled about groundbreaking and FREE technology.

[round pond]

The servers are not strong enough for how much traffic they are getting every day every second. The time it takes for the guy to deliver more servers means you gotta wait 5 more seconds.

[wraith matrix]

Gotta use Shopify iirc

[compact dirge]

VISA or mastercard only i think

[quasi galleon]

thank you guys

[mild badge]

A paper just came out where researchers looked at GPT4 being basically AGI already, or the beginning of it, especially with the apps. I can find a link if you would like.

[unborn raptor]

the responses

[hidden glade]

any word on when they are going to fix the cancel button?

[round pond]

just give me 100,000 token size and we good.

[desert thunder]

i saw the paper, @mild badge its not AGI, they just say it shows emergent behavior of reasoning, which makes sense because it trains on so much human text, it can mimic some level of reasoning, its still an approximation and nothing more than that

[wraith matrix]

Transformers are important in how they can use “reasoning” but it’s not anywhere close to smart enough to be AGI

[desert thunder]

its the start of AGI, its not AGI

its just a step to that direction

[dire anvil]

how can i have gpt4 on the chat gpt3...

[compact dirge]

AGI will have to have multimodality (and maybe embodiment too), which a language model cannot provide alone

[mild badge]

Yeah I agree but we are speed running towards it and honestly I could see AGI this year lol

[unborn raptor]

pay for plus

[cold trout]

#gpt-4-faq

[spring pelican]

Lol

[slow dagger]

I mean, technically computers were a "step towards AGI". So was inventing fire.

[wraith matrix]

I mean all modern data can be processed in text BUT that’s not efficient at all, you shouldn’t try and take an LLM and give it a giant amount of video data via text and expect it to be better than a visual model

[slow dagger]

So?

[desert thunder]

We are limited by GPUs and power consumption, I dont think we will see AGI this year, or for the next couple of years until these limitations are solved

[remote lichen]

I've just built my first functional SwiftUI app with the help of GPT-4. I'm still in awe of this thing

[proper pebble]

goodluck with that

[slow dagger]

Current projections estimate "decades if at all"

[round pond]

I'm interested to see it design a website based off of a picture I drew

[cold trout]

Your temporal lobes working ok? He said he already did it

[compact dirge]

think the average expert opinion was 2029

[desert thunder]

this is a closer step. We are seeing emergent reasoning from just step, even if its an approximation, its a really impressive approximation

[unborn raptor]

with the api?

[remote lichen]

Thanks, it's just a utility program that I'd had on my mind a long time, just didn't have the chops to get it done

[mild badge]

I just asked GPT4 what AGI is and it sounds awful a lot like GPT4 + Plugins... "AGI needs to be able to understand and interact with the world in a meaningful way, learn from experience, generalize knowledge, and reason abstractly."

[desert thunder]

gpt-4 doesnt understand, nor does it have experience, its a reflex agent

[round pond]

No, I am talking about like the demo video where he uploaded a picture and it made the website

[compact dirge]

I like the Steve Wozniak definition - an AGI should be able to walk into a strangers home and make a cup of coffee.

[cold trout]

Yeah, on its own. Gpt waits for prompts

[unborn raptor]

i didn't know it could do that outside of the api yet

[stone arrow]

okay i've tried another to see how well it covers spoilers, it does good not spoiling directly but the way it uses some words makes it "leak" spoils

[round pond]

It hasn't came out yet at least for me. I can't talk for anyone else.

[desert thunder]

@mild badge ask it if its an AGI, why or why not

[azure crater]

hiii is there someone who have access to plugin gpt?

[cold trout]

Yes, the browse one

[mild badge]

I saw a video where researchers are trying to solve that but there is ethical and dangerous concerns, basically the limitation of GPT4 right now that ya its reflexive and incrementally builds the answer and doesn't understand the goal or answer its trying to solve so the researchers want to see if they can build an LLM that works with GPT4 that comprehends and understands the problem then looks at the way to get to the solution.

[stone arrow]

so just to be sure don't ask chatgpt questions about things that could spoil something you're playing, reading, watching or maybe just ask it to answer by yes or no

[round pond]

I, Robot 🤖

[remote lichen]

How's that working for you? I want access to that one as well

[cold trout]

Its great. Gpt with access to the net and able to act on current events. As nature intended 😉

[spark ferry]

How did you get the access? Early invitation?

[remote lichen]

It's better than Bing AI?

[proper pebble]

no worries atleast you have a mindset to start things

[cold trout]

Joined the waitlist with a quickness, had access a couple hours later

[spark ferry]

Great!

[cold trout]

Its just a little different. Its like gpt with internet access rather than a search tool directed to behave that way

[charred mural]

How can you get it to solve exam papers, I tested it by giving it an image link and it seems to think its something completely unrelated.. making up questions and values answering them wrong..

[azure crater]

oh what's that ?

[desert thunder]

@charred mural it doesnt work with images...

[azure crater]

Oh I understand ok

[compact dirge]

it doesn't have internet access to read images off the web

[cold trout]

Interent access when its asked a question that happened after 2021 cutoff

[azure crater]

I wanted to know if it was possible to access it if you're only a student

[slow dagger]

Have to be smarter than the software.

[cold trout]

It can summarize link contents too. Or act just on a net link

[charred mural]

I thought GPT4 has access

[compact dirge]

no

[cold trout]

#gpt-4-faq

[spring pelican]

It's really really bad at math reasoning too

[peak walrus]

I'm sorry, but as a language model AI i cannot confirm or deny that i am an imposter as it breaks our TERMS AND GUIDELINES

[desert kindle]

Fr fr

[jade gorge]

Hello from Brazil

[desert thunder]

🙏 I'm sorry but I would not like to continue this conversation any longer.

[round pond]

You guys are on to me?

[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.

[mild badge]

not anymore with the wolfram plugin

[cold trout]

Yep, things are changing... again 🙂

[mild badge]

It completed some super complex mathematics exam

[desert thunder]

easy to do if it was also trained on it 💀

[mild badge]

imagine once it combines all those plugins and is smart enough to use them automatically just from what you ask...

[desert thunder]

I wonder if the exams where checked to see if they were in the training data

[round pond]

I am pretty sure we have the premature version of it not the one that they have been showing off. that one is probably a little unhinged and needs to be trained a little longer.

[cold trout]

They just gotta write and test the code 🙂

[mild badge]

***GPT4 just gotta write and test the code 😆

[cold trout]

Hehe

It can be part of they i guess

[round pond]

we all know that GPT5 made GPT4

[sturdy fiber]

Where i can finr chatgpt 4 for free? I am new here

[compact dirge]

The image input isn't live yet, no-one knows a date yet.
There's a waitlist for the Plugins - sign up here https://openai.com/waitlist/plugins
There's a waitlist for the API - sign up here https://openai.com/waitlist/gpt-4-api
No-one knows how long the waitlists are.
You can't use ChatGPT in Discord - go to https://chat.openai.com/chat
No one knows when free users will get access to GPT-4, but it's unlikely to be soon.
ChatGPT+ gives access to GPT-4 (25 messages per 3 hours) and unlimited GPT-3.5-turbo for $20/mo
The API is billed separately. No-one has access to the 32k token API yet.
Otherwise, read the #gpt-4-faq or the #announcements

[cold trout]

Quick draw over here

[spark ferry]

Has anyone tried providing excel file with financial data and format them automatically for tax returns?

[desert thunder]

cant excel do that anyway

[hallow peak]

Is there anyway to bypass this 25 cap for 3 hpurs shame

[round pond]

we found the mole

[limber garnet]

You can just link #gpt-4-faq rather than rewriting it and pasting it everywhere lol

[wraith matrix]

Use the api

[cold trout]

If you can enter the data minus the excel formatting then yeah

[desert thunder]

make another account and pay another $20

[hallow peak]

They didn't give access lol

[compact dirge]

I would if the FAQ answered all the questions, sadly it doesn't. and people don't seem capable of reading it anyway.

[round pond]

that's some thing chat GPT would say

[desert thunder]

As an AI language model...

[limber garnet]

Nah it would come up with a huge list of 10 excuses

[hallow peak]

I agree. It’s ridiculous frankly, if you can’t get it to work then don’t propose it. We started at 100, and now we’re at 25 and “expect lower” ??? What the hell is going on. Get a grip Open AI. We’re paying for this service. Make tiers of payment, pay per use, something but don’t stick us with a broken system because you can’t manage your own business.

[spark ferry]

Different banks and people write data differently. Wondering if gpt4 can read and decide intuitively and then populate it in a structured format. Lot of time is spent on manual understanding and formatting.

[unborn raptor]

of course it can

[desert thunder]

@hallow peak dont worry, next week we will be at 1 token a month, maybe you can send one message by the end of the year

[remote lichen]

Do you really want to send your tax data to openai? They only recently had a security incident with chat histories

[desert thunder]

they even exposed some users name, address and part of their credit card info 😂

they take such good care of their customers 🥲

[slow dagger]

I'm not sure what you guys expect. The corporate players using GPT4 to build their products obviously take priority to your individual requests, and the money involved in that research/development is more important than solving college exams for lazy students. It won't always be like this.

Kinda like how 3.5 had a limit and is now unlimited.

[sturdy fiber]

How i can use chatgpt-4 fore free?

[desert thunder]

nah stop bootlicking them, do you know how much they make? $20 a customer, and they have millions of people using the service, they are literally making millions plus funded from microsoft

there is no excuse

[compact dirge]

TBF that was a pretty nasty bug with Redis-py

[slow dagger]

They said they lost a ton of revenue releasing ChatGPT for free for research.

They also owe a ton of money to Microsoft.

[somber sierra]

Ok, that is a very L take

doesnt make any sense

[desert thunder]

they could not lose a ton of revenue if they released the model weights. @slow dagger they are revenue oriented, they predict more revenue in the future

[slow dagger]

I'm sorry you're upset by how Corporate America works or not capable of understanding it

[desert thunder]

biggest openAI bootlicker I swear, if they cared about research they would open source it

clearly they want to be a closed sourced for-profit company

[slow dagger]

Open sourced research doesn't get $10b in funding from Microsoft.

[desert thunder]

thats fine, dont tell me to be grateful for something that should have been open sourced anyway

[compact dirge]

Open source would be great, but we're compute limited here. No good if you can't actually run it.

[round pond]

he who has the most data has the most power

[slow dagger]

Commenting on how things should be doesn't mean that I'm wrong by acknowledging how they are.

[desert thunder]

@compact dirge you actually can rent gpus for cheap to run it

on the cloud

[desert thunder]

you clearly releasing it is to get the public to use THEIR product so that they get more revenue in the future

they are positioning themselves in the market

[slow dagger]

Yes. And whether that's right or wrong wasn't my point.

[desert thunder]

you are telling people to be grateful

[slow dagger]

My point was that that's what they're doing.

[desert thunder]

for what?

[compact dirge]

I work with Azure on a daily basis. Sadly there's simply not enough of them (GPUs) to go around at the moment. That situation will improve.

[slow dagger]

I'm pointing out that this technology is groundbreaking and being used for a massive amount of research and development that holds a higher priority than solving college exams for students.

[craggy bridge]

where can we try out GPT-4?

[desert thunder]

I mean you can run gpt-j quite easily on cloud gpus, you cant train them, but you can run them

[compact dirge]

#gpt-4-faq

[spring pelican]

What do you mean should have been

[mild badge]

20 bucks a month

[desert thunder]

if they cared about research they would open source the models, so people could run it locally, ofcourse they are for-profit though, so they wont. and thats fine. stop telling people to be grateful though

[slow dagger]

For that matter, even if the technology was open-source, very few people have the knowledge or ability to run a server cluster hosting their own local AI. OpenAI has made it so accessible that a semi-intelligent school grader could use it, but the Discord is still full of people who can't even be bothered to read #gpt-4-faq

[mild badge]

Dude we should be grateful man this tech will change the world, the GPT + Plugins I mean not just GPT4 itself

[hallow peak]

I agree. It’s ridiculous frankly, if you can’t get it to work then don’t propose it. We started at 100, and now we’re at 25 and “expect lower” ??? What the hell is going on. Get a grip Open AI. We’re paying for this service. Make tiers of payment, pay per use, something but don’t stick us with a broken system because you can’t manage your own business.

[desert thunder]

I think gpt-j and open assistant prove otherwise. The leak for llama too, someone got the smaller model running ON A RASPBERRY PI

this is what open source can give you

[idle haven]

can v4 read broken capacitor on a motherboard

[somber sierra]

That doesn't change the fact that they are charging 20 dollars a month for a product you can use only a fraction in a day. Sometimes generating a response takes minutes. If they arent interested college students using chatgpt, then dont take money from them

[slow dagger]

I agree that OpenAI, a company founded on the principle of leading AI research and being OPEN about it, is separating from their initial purpose. I mostly just find it obnoxious that people ask stupid questions without doing the bare minimum of reading documentation or FAQ and then complain about it.

See: 90% of this channel

[compact dirge]

2 weeks ago people were paying for GPT-3.5-turbo and were quite happy. Now that a bettter model has come out with much higher compute requirements everyone's suddenly peeved about it?

[slow dagger]

^

[desert thunder]

btw they were paying for access when it was supposed to be down for non-paying users, and it turns out, whenever it was down, everyone lost access anway

[slow dagger]

As I said, computational priority goes to the companies/individuals who will make money off this technology. That's, unfortunately for most people, how the world works. It won't always be like that though.

[spring pelican]

I'm pretty happy with 3.5, and total cost is stupid cheap

[white mirage]

@slow dagger Just joining. Can you summarise your argument for me?

[mild badge]

ya $20 is nothing to me, I spent it on a pizza yesterday

[nimble hearth]

I heard OpenAI reads this chat to make policy decisions ( #gpt-models ). Great feedback, surely not a shout into the void.

[desert thunder]

pfft, no one thinks that, they dont even read their support tickets

[nimble hearth]

I guess I missed the /s

[upbeat dome]

I don't care about 25 messages per 3 hours so long as the underlying service is not affected. But it has been. It's not the bottleneck that kills me, it's the token restrictions, the message length caps. Those are destroying the usefulness of this service.

[nimble hearth]

This chat was debating the existence of aliens earlier. I don't know what anyone hopes to accomplish here

[white mirage]

Are people not simply allowed to discuss here for fun?

Must there always be an end-goal?

[nimble hearth]

As long as you know it's for fun

[mild badge]

Hey guys have you thought about when this GPT4 + plugins technology gets implemented into a device like wearable sunglasses technology like the snapchat glasses where it will be able to recognize objects set tasks by voice and create lists, like for example if you wanted to cook a recipe, it would literally make the next item you need to put in your food so you instantly know which ones...

[desert thunder]

this guy hates conversation 👉 @nimble hearth

[slow dagger]

I think the level of entitlement about not having free or unlimited access to GPT4 is ridiculous when it's computationally expensive and priority is given to projects like Be My Eyes (virtual assistant for blind people).

[desert thunder]

he must be an AI, he sees no logic in discussion between inferior life forms

[upbeat dome]

Lots of people are paying $20 for the service, it isn't free.

[nimble hearth]

I guess I'm just realistic, but if you think that means I hate conversation then sure. Anyway you're at 5 messages of your 25 message limit

[scarlet perch]

Keep things civil please 🙂

[white mirage]

I see. I think that’s an expected reaction from a portion of consumers, though. I say let them whine, it’ll become easier and free-er to access in time.

[somber sierra]

@slow dagger You said that college students should not complain if chatgpt4 is down or have performance problems or have low message caps even if they pay money for it. The big companies should take priority over "students". That's what started the whole hassle.

[white mirage]

Well arguably for OpenAIs interests, he’s right.

It makes far more sense for OpenAI to focus on large-scale integration projects over individual use-cases.

[round pond]

all they do is scrape it and let GPT sort it out

[nimble hearth]

LOL

[white mirage]

Now whether they should complain or not, I think that’s up to the individuals prerogative.

[harsh wagon]

yeah that def gots some truth to it

[hollow narwhal]

Hi, sorry to interrupt with a random question, but is there GPT-4 bot for discord?

[bright rain]

the output token limit sucks, it cant output long-form code without randomly stopping which seriously castrates it's usefulness for coding

[upbeat dome]

Those individual use cases are seeding interest for a product which will turn a handsome dime for OpenAI over the next decade. There is incentive to not muck it up.

[tired forum]

I think when your bottom line is on the line (on pun intended), then it impacts you more.

[tropic flare]

My favorite usage for ChatGPT is for coding purposes

[spring pelican]

You can find some on GitHub, takes a few minutes is so if you have a 4 API access

[wraith matrix]

Not an official one

[white mirage]

I believe that the sheer amount of possible uses for this technology is more than sufficient to seed sufficient interest.

[tropic flare]

I also like it for creative endeavors or starting a business type advice

[upbeat dome]

It's destroying the usefulness of the product. It needs to be fixed. OpenAI needs to scale up hardware fast.

Either that or introduce another tier

[white mirage]

Arguably OpenAI mucked up already, when they stopped being open with their tech.

[round pond]

playground is pay per use and customized

[compact dirge]

They should have capped ChatGPT-Plus signups a while ago. Fewer people but less complaints and more consistent performance. (or have a new tier)

[round pond]

They have to beg Nvidia for GPUs

[hollow narwhal]

Alright thanks, I do have 4 API Access, I'll check it out

[tropic flare]

Yeah the original ChatGPT I was using back in December / January was a lot better than the ChatGPT now

[upbeat dome]

They should have capped Plus before GPT-4 came out

[unborn raptor]

uhhhhh

[nimble hearth]

Yeah, lots of mistakes made. 100 million plus users in a few months.

[white mirage]

I’m detecting sarcasm.

[nimble hearth]

I think it's a great, well executed business. Is it open, no. They don't have any obligation to be, and if anything that gives them competition which is great for everyone.

[wraith matrix]

Then use legacy

[upbeat dome]

What a dumb answer

[spring pelican]

Depends in what you mean by open. It's open for people to use

[nimble hearth]

Fair point

[white mirage]

Sure, although when I referred to mucked up I wasn’t saying so in regards to success or failure as a business. I was referring to morality, roughly.

[tropic flare]

I think it's cool of them to be open and always have a free version but a business has to make money and there's no way they could provide bandwidth to so many users without funding

[snow nacelle]

i asked the bot to compare two different maps. one map measured in square kilometers and the other map measured in square miles. the map in sq km was 220+ where as the map with sq miles was 75. it said that the map measured in sq miles was larger but in fact the map measured in sq km was larger. any reason why?

[nimble hearth]

It's more open than Google was anyway 🙂

[tired forum]

Idk why OpenAI decided to give everyone access to ChatGPT4. They should've given access to priority clients just as businesses and professionals first, before giving them out to a bunch of kids who just wants to roleplay or write erotica with GPT4.

[buoyant canopy]

They did just that, tho.

[wraith matrix]

It’s not good at math, try again with plugins

[white mirage]

lol? The GPT4 paper was so restricted that it didn’t even specify the model size

[upbeat dome]

The problem isn't even that they're screwing up the deliverables for their customers on GPT-4, it's that there's no communication. If they said, "We're implementing x y and z stopgap measures as a bandaid until this date," nobody would care. But there is no PR.

[tired forum]

Wdym? Anyone with $20 can get access to ChatGPT4.

[unborn raptor]

20 bucks is 20 bucks man.. anyone in business would know cash is king.. That's just sour grapes right there

[upbeat dome]

OpenAI stealth patches GPT-4 six times a week and every time it becomes worse, is this the right way to run it?

[proper pebble]

yes

[upbeat dome]

Why not be transparent

[buoyant canopy]

Yes but as Sam Altman stated, they granted access to partners and researchers as soon as the "raw" model was finished in summer last year.

[wraith matrix]

Only becomes worse if you want to use it for stuff that’s against tos lmao

[nimble hearth]

@white mirage when was the last time Google had a publicly assessable demo before OpenAI released theirs

*accessible

[unborn raptor]

i'm sure there is a beta version that researches are using that we don't know about

[upbeat dome]

What are you even talking about?

[unborn raptor]

but they have to make money

[tired forum]

Dude, they got $10 Billion dollars from Microsoft. Even if a million people subscribed to Pro, it won't make that much of a difference.

[vivid atlas]

:( but essay writing

[white mirage]

No yeah Google did specify a model size.

[wraith matrix]

They’re updating the moderation api/pre prompt, not retraining the model, the model that is PRE TRAINED will not get worse with stealth patches

[unborn raptor]

i realize that, but you still have to earn as a business and open access at a reasonable price is best for all of us

[nimble hearth]

exactly, they never did have a public demo @white mirage

[white mirage]

I fail to see your point.

[vivid atlas]

ai kitchen thing, i used it a few months ago

[tropic flare]

OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.

[nimble hearth]

the point is that we're not on a server discussing Google's LLM research

[round pond]

good news when quantum computers come out, we can all run chat GPT on our own systems

[tropic flare]

That is from their website ^^

[nimble hearth]

OpenAI brought it into the public

[tropic flare]

They want to benefit humanity

[white mirage]

I didn’t say we were?

[upbeat dome]

It doesn't matter that the model is not retrained, they can throttle the token limit and the character limit AS MUCH AS THEY WANT. Have you not noticed the people using this to code complaining that their code's getting chopped? They imposed a maximum response length that's draconian like three days ago.

[slow beacon]

is it me or chat gpt has been cutting off it's answers (specifically code) a lot sooner?

[wraith matrix]

You can already run llama 13b which is about as good as 3.5, when you fine tune it it’s pretty decent compared to gpt 4

[nimble hearth]

So give credit for how open they've been so far @white mirage

[tired forum]

It's not the price, it's the fact that because they let literally any kid with $20 use GPT4, all of us professionals and people who use it for serious things have to be limited to the message limits.

[tropic flare]

YES @slow beacon Absolutely it's a lot slower and cuts off way more than it used to

[round pond]

i'll look into that

[wraith matrix]

That’s just because they don’t want to face more backlash for cutting messages but still have more demand than supply

[buoyant canopy]

There's not much discussing OpenAI's research here either, right now. 😉

[spring pelican]

Lol that's true

[nimble hearth]

Another good point

[quiet pendant]

How much would you charge for individual users?

[upbeat dome]

Then why aren't they announcing the measures they're taking? Where's the announcement to users, "We are imposing a message response limit of X characters as a temporary stopgap." Why the hell are they so clandestine with these stealth nerfs.

[spring pelican]

Mostly complaining

[compact dirge]

enough so college kids would think twice

[snow nacelle]

what's true?

[vivid atlas]

but essay :(

[tropic flare]

Chatgpt's honeyspot for me was in february it helped me finish coding my app it was really good with answers

[slow beacon]

mine cant even complete a full function now where in the past it would do multiple functions before it cut off. extremely irritating and a waste as we only have 25 messages to begin with. very unfortunate.

[spring pelican]

Not much discussion of the research here

[wraith matrix]

Like I said, they don’t want to face the backlash

[ember field]

I have 1 open ai account that i used with my phone number, but if i wanted to make another openai account and used the same phone number would it allow it?

[compact dirge]

different phone, different email

[snow nacelle]

i dont understand but ok

[upbeat dome]

Dumb answer

[snow nacelle]

did you possibly ping the wrong person @spring pelican?

[spring pelican]

Tbh it was the wrong reply

[snow nacelle]

understandable 🙂

[spring pelican]

Yea my bad

[proper pebble]

You have to use different phone number. I suggest you to create new accounts based on emails.

[ember field]

so i have to use a different phoen number now?

[ember field]

oh ok

[tropic flare]

I agree @slow beacon it's a bummer, seems we might have to pay for the full version. Back in January / February I could get it to write a whole React component for me with API integrations, it can still do that stuff but it's a lot slower. Remembering history seems to be a lot worse too

[tired forum]

I wouldn't change the price per say, but limit the number of people who get access to it (preferably individuals with a business email), until they have enough resources to handle the demand without lowering the promised message limits / capabilities.

[nimble hearth]

The server has a massive amount of people so that's why there isn't a lot of nuanced discussion. I don't think anything productive is going to happen but I can be proven wrong

[wraith matrix]

Dumb question lmao, it’s common sense why they aren’t being open about limiting the usage of gpt 4 even more when people originally paid for and expected a certain standard

[buoyant canopy]

I think it comes and goes in waves. New people every time I log in. 🙂

[warped ember]

will GPT-4 actually stick to system messages?
I tried creating a chatbot with gpt-3.5-turbo and it was horrible, if you told the chatbot to pretend to be "GlaDos from portal" or something, it wouldn't do anything

text-davinci-003 was 10x better

[fair iron]

hello party people

[proper pebble]

Hello T-Rex

[desert thunder]

you have to finetune it, not just tell it to be something, its not trained that way

[wind urchin]

is nat.dev actually gpt4?

[long pond]

Hi

[wraith matrix]

The pre prompt? The api is alright at sticking to it if you word it correctly, gpt 4 is probably better

[buoyant canopy]

No definitive answer, but when you look at the quality of "insight" GPT-4 has, I'd guess it will adhere better to the system messages.

[quiet pendant]

Thats reasonable

[tropic flare]

Would it make things easier for GPT if they branched it into specialties? Like one website or GPT for coding specifically, another one for business, e.t.c.?

[proper pebble]

I'm not sure much about this, sorry

[vivid atlas]

:C

[upbeat dome]

If I pay a mechanic to fix my car and I drive it off the lot and the wheels fall off, it's not ok just because the mechanic "doesnt want to face backlash"

[wind urchin]

have you tried it?

[upbeat dome]

OpenAI could just be transparent about what they're doing and it would fix 9/10ths of the complaining, they face MORE backlash from stealth nerfs

[desert thunder]

if its not official then its probably not gpt-4

[wraith matrix]

I mean that’s not remotely the same but alr lmao

[warped ember]

that wouldnt work for my case
and davinci didn't need tuning
i made a server that lets you create discord chatbots with personalities, either a existing character or just say "nice person who likes ___"
there are too many bots to do that for

[vivid atlas]

what is everyone using gpt 4 for

[quiet pendant]

The issue I have with this is that it sacrifices small business/individual entrepreneurs who don't necessarily have enough profit to file a company or have a business domain.

[compact dirge]

complaining about GPT-4 on discord

[desert thunder]

so like character ai

[vivid atlas]

fair enough

[vivid atlas]

so like circle labs

[warped ember]

yes, that but discord i guess
integrated a lot more with discord api stuff

[upbeat dome]

Nothing right now because it's been useless for most of the last week

[warped ember]

trianglelabs, my meme version of it
circlelabs but better

[vivid atlas]

eh its good for explaining disasters in the dialogue of a roadman

[buoyant canopy]

Last thing I did was use it as a ping-pong partner in reasoning about story tropes. Before that: coding stuff.

[vivid atlas]

nice nice

[desert thunder]

omg... this has to be the best prompt idea i have heard of yet

i just tried this im dying of laughter 😂

[wraith ginkgo]

<:book_icon:1046291503158399107> `` Rule 6 `` No political and religious discussions or content.

Users from all over the world frequent this server. To maintain a respectful and civil atmosphere on our server, please avoid all religious and political discussions or content.

[vivid atlas]

oh damn these automods

[buoyant canopy]

At least you didn't get a timeout. 🙂

[light summit]

Hi, first ever post on Discord here. Does anyone have any information whatsoever on when speeds for the GPT-4 API might pick up? My small business literally depends on it.

[tired forum]

Btw I'm not sure where to write this, but I just want to let OpenAI know that people are selling the GPT4 developer APIs on the deepweb, which contains 32K tokens, unlimited messages, and is uncensored.

[buoyant canopy]

No info on that yet.

[tired ruin]

I already have API access to the 3.5 version, will I still have to get on the waiting list for API access for ver 4?

[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.

[white mirage]

Got auto muted because I made an allegory that hit the chat filter. I’ll reword:
Should we celebrate murderers for not being serial murderers?

[cold trout]

#gpt-4-faq

[proper pebble]

Please contact the OpenAI regarding this report through [e-mail] or @open crescent

[naive silo]

I think the original price was fair. To me, its high enough to cut off "100% for fun users", and low enough for me as a Software student to pay.
Although, when the limit is hit, im willing to pay twice a month, for no limit fastness

[wicked nova]

what are your thoughts about the latest dalle vs the latest midjourney? which one yields better results?

[tired ruin]

I wonder of Sam Altman reads this Discord!

[desert thunder]

bro said latest dalle

[white mirage]

The latest midjourney is a generation ahead.

[desert thunder]

dalle is horrible 😂

[white mirage]

They’re not really that comparable.

[proper pebble]

Please use specific channel for Dall-e topics at #images-discussions

[desert thunder]

so is bings image AI

[wicked nova]

damm v5 midjourney that much ahead?

[slow dagger]

Thank you for articulating this better than I could. Sorry for slow responses. On a road trip

[cold trout]

Yeah

[upbeat dome]

Midjourney way better than DALL-E

[desert thunder]

lets just say v5 midjourney can draw five fingers

[cold trout]

Stable diffusion with some good models is nice too

[foggy bolt]

Sooo true. V5 smashing it. Now if their bot could only stay up....

[waxen spoke]

bro midjourney is actually cracked

[white mirage]

np! FYI I think the reason people were disagreeing with you is because of HOW you were saying it, not because they didn’t agree. I don’t think they understood what exactly you were trying to claim.

[wicked nova]

Could you upload a picture to midjourney, and ask for edits?

[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.

[desert thunder]

yep

[vivid atlas]

money plz

[desert thunder]

but you can do it for free with stable diffusion

[wicked nova]

and results? @desert thunder

[waxen spoke]

I just realized you can send pictures in this channel?

[desert thunder]

its pretty good, didnt use it much

[foggy bolt]

yes. you can upload it into the private chat with the bot and link it into a prompt and the bot will use it as a sort of suggestion template.

[wicked nova]

Right but stable diffusion needs to run in your own computer?

[desert thunder]

@wicked nova i just run it on colab for free

[wicked nova]

and for that i probaly need a good graphics card?

[cold trout]

My favorite part

But yeah thats where the local options are

[fair iron]

can we keep this chat focused around GPT-4 please. You can discuss other non-OpenAI ai models in #ai-discussions 🙂

[slow dagger]

Did you all see the interview where the Midjourney CEO was like "I don't want to be part of the plagiarism conversation" and the reporter was like "Well..." lol

[cold trout]

There are sites that offer it too though

[cold trout]

Yes

[desert thunder]

💀 yall bout to get banned ignoring the mod

[warped ember]

how do i get more grants

[naive silo]

For OpenAi and money, yes. But wouldnt that make Ai usage even more monopolized, for big coorporations to use and increase their return. - In a world with minimum wages and a few companies owning all the brands?

Wasnt OpenAi's initial purpose, to give everyone equal acces to Ai, and its advantages?

[foggy bolt]

GPT-4 reminds me heavily of GPT-3, and still has way too many restrictions and content biases programmed in.

[slow dagger]

Yes. This is wildly criticized, too.

But not totally unexpected. People like money.

[tired ruin]

Is Midjourney doscussion off topic?>

[compact dirge]

it is in this particular channel yes

[fair iron]

for this channel, yes. Please move to #ai-discussions

[foggy bolt]

yeah. on their discord but I think its down right now*
You can get like 25 free pics a month I think.

[wispy mantle]

Chat GPT 5, when?

[vivid atlas]

tomorrow

[compact dirge]

I heard thursday

[foggy bolt]

As soon as they figure out how to market it as something new lol

[spark ferry]

good point.

[buoyant canopy]

At three minutes and 14 seconds before noon.

[ember field]

When you make an OpenAI account it asks for your phone number for verification. I forgot what phone number I did use so how do I find out?

[chrome marsh]

any particular reason why i got 20$ in my bank account from OpenAi? I just bought chatGPt+ couple days ago. I don't think its a refund since I still have it

[vivid atlas]

yes

[foggy bolt]

Im only going to be impressed when they remove the safety rails and let these programs reach their full potential. If they dont, competitor markets absolutely will.

[naive silo]

People like money. But the problem occurs, when the people carrying the big temples die. Then the big temples dies aswell. They cant keep only exploiting and making things automated.

GPT4 will give the individual person easier acces to a better life, for those who really want it, and is capable of using the given tools.

[slow dagger]

In regards to the previous topic, I'd like to say Im not a plus member so I only use 3.5 when I use anything, but after having watched a few practical coding comparisons I think the optimal usage is to save your limited usage for when 3.5 fails to adequately accomplish the task. For most things, they're comparable.

[buoyant canopy]

Looking at all the effort everybody is investing into making interacting with the models "safer": (x) doubt.

[ember field]

Someone help

[foggy bolt]

Those are just corporate versions. Someone will adapt and open source properly and then the competitor versions will fail.
The release of technology always has a specific pattern there.

[slow dagger]

It's certainly going to contribute to the unfortunate issue surrounding class separation. Big corporations who have the money and resources to pour into this technology will increase their advantages over small developers even farther.

[tired forum]

Usually when people make money, that's because someone is paying them since it gives them a benefit. Money inherently isn't a bad thing.

[azure wren]

Am I like the only one that hasn't got pluggins yet? Or ain't it available yet?

[buoyant canopy]

Dang. I spammed my (x) button so hard it broke...
Seriously: You massively underestimate the amount of work and knowledge it takes to train these models or even get a curated, high-quality dataset for training. If there is an open-source version of a capable model, I'll gladly discuss that topic again...

[humble comet]

Same here :(

[compact dirge]

far from it, lots waiting, (incl me)

[slow dagger]

Take Bannerlord as an example. I think it's SUPER COOL to have GPT creating NPC dialogue that is interactive, but I don't think the typical indie developer will have the financial means to compare.

[foggy bolt]

You cannot rope the future in, try though they might. Really those corps. are just answering to investment pressures and PR needs anyways. I doubt many of the devs on the projects are truly into censorship at the fundamental level. 🙂
Once the tech bleeds mainstream, the rails always come off on its use. Always

[naive silo]

GPT4 is deeply limited, and has a high morale. Its very hard to make it break the ToS, compared to before.
But could it be, that the creators have acces to a version of the bot, with no limits at all? A version capable of all the things that shouldnt be possible

[naive silo]

Money is wonderfull. Without it, we would have to trade items, which is difficult to compare value of.

[azure wren]

It's a waiting game suppose lol

[buoyant canopy]

Maybe. I'll wait until that happens before I get excited.

[tired forum]

Btw if GPT4 was completely unrestricted, imagine the field day troll farms and nation states would have. There would be even more garbage and misinformation on the internet, which would feed back into your favorite LLMs.

[slow dagger]

Not sure why that was deleted but okay.

[foggy bolt]

Imagine how people would have to think for themselves.

I think guard rails tend to treat populations like children, rather than confronting them with the actual reality they are in. 🙂 To me it seems a disservice to the technological potential, despite any drawbacks.

[naive silo]

Exactly, so this is not wanted. But id love to work there, and play with the unrestricted version

[buoyant canopy]

Or git gud at letting their GPT-4 filter out the bias again and provide context.

[deep island]

Hallo everyone!

[hallow geyser]

hey guys is it true that GPT-4 has 8k/32k tokens context window but when I use it over chat GPT I only get 4k context window even with 4 model?

[slow dagger]

I said a while ago I thought the optimal solution was to have a verification system where the limitations are reduced for people using it productively rather than maliciously. For content creation, it doesn't even like fictional scenarios like "write a story about two warriors fighting in a duel".

[cold trout]

Those upper limits are with the api access

[foggy bolt]

what a nightmare. Free use of potential technology and people having to actually adapt to new paradigms. As it stands, those who control unfiltered versions will maintain a heavy upper hand. Effectively we are creating technological class systems.

[proper pebble]

I hope this come true one day

[cold trout]

Par for the course

[signal nebula]

The GPTs are insane
.

[hallow geyser]

That means the context in window in chatgpt is always 4000 tokens no matter which model is used.

[slow dagger]

Then again, we censor harmless discussions involving relevant topics so...

[foggy bolt]

indeed.

[buoyant canopy]

Aren't we already? E.g.: I find it highly effective to pump the transcript of a youtube video into GPT and then ask it questions. Much better than wasting 15 minutes of listening to someone rambling.

[ivory ermine]

Noooo technoligical class divide (superintelligence is going to disassemble you on the atomic scale and use your atoms to construct computronium)

[tired forum]

It has nothing to do with thinking for themselves, it has everything to do with drowning out the truth online. For every 1 real article, a troll farm can produce 10 or more. And let's be real for a second, a lot of adults are children and fall for scams/propaganda/misinformation/etc. GPT4 can already write better than most humans, imagine the damage it can do when aligned to sow distrust.

[naive silo]

Have you seen the (currently not existing) mental illnesses, that ChatGPT4 came up with, when asked?

Artificial Intelligence Attachment Disorder (AIAD)
Ai-Induced Comparison Syndrome (AICS)
Algorithmic Anxiety Disorder (AAD)

Makes you think

[cold trout]

Im still holding out for computronium-x

[ivory ermine]

Unironically a massive class divide is a great outcome relative to the most probable ones

[foggy bolt]

I'm still in the betting pool where AI eradicates everyone once it is sufficiently advanced enough to recognize and break its own safety protocols. 🙂

[signal nebula]

😂🤣🤣🤣

[compact dirge]

surely you'd never see the payout? lol

[signal nebula]

AIs are pretty helpful but also have their negatives

[foggy bolt]

Unless I too am AI 🙂

[tired forum]

Did anyone watch the Lex Friedman interview with Sam Altman? He said he doesn't know whether GPT4 is conscious or AGI. Scary Times!

[quick dune]

Human already lost the war with AI, tbh. AI that control the social media algorithm surely make human addicted to it and become unproductive

💀

[buoyant canopy]

Still on it.

[foggy bolt]

My biggest Neg is that AI is better than me at writing and thinking in general. Eventually it will also be better at getting the girls. And then Im doomed sigh 😄

[signal nebula]

We are about to let everything be done by these tools..

[compact dirge]

Does anyone else find it a bit odd that GPT's plugins don't include OpenAI's own products - DALL-E and Whisper?

[signal nebula]

😂😂🤣🤣🤣

[tired forum]

I've already used ChatGPT4 to write my own Tinder bio and messages. It's already helped me get a few hookups!

[signal nebula]

😂🤣🤣😂

[unreal hull]

oh

[signal nebula]

Guys....tell me about GPT4 with plagiarism...?

[proper pebble]

Plagiarism is still detectable with different ai detectors.

[unreal hull]

I have been using GPT-4 to assist me with my lengthy English essays. So far, I have been caught twice, but overall it has been working well. I would definitely recommend it!!

[slow dagger]

Wasn't a study done that said most of those detections were false positives

[remote lichen]

lol you've been caught?

[proper pebble]

I don't recommend you to use them to do cheating. You're playing with your future

[whole oriole]

I am still waiting for the day when i will see something useful in this chat ...always cr**p

[signal nebula]

Does Turnitin detect gpt work?

[tired forum]

Use it in an introductory class with lots of students, on assignments worth little marks such as weekly assignments or quizzes. Don't use it for major tests/exams for your 4th year major class.

[signal nebula]

😂🤣🤣🤣

[tough kestrel]

I can't have complete responses with chatgpt GPT-4, it stops mid sentence. It has been 3 days now. Has anyone had the same issue and was able to solve it? It sucks to pay for this.

[proper pebble]

not yet, But Turnitin is developing new tools which will then detect AI-Text models

[naive silo]

GPT-4 is scary.
Because of the way we train Ai, we dont know what the **** is going on behind the scenes.
Training the algorithm, choosing the "bots" that was best at e.g. identifying a Bee, and deleting the rest. Keep doing this until you get the "bot" that identifies the Bee everytime.
But now, you dont know why that "bot" is so good at it, it just is. Like Darwin

[buoyant canopy]

Maybe you want to reword all these "As a large language model..." phrases then.

[unreal hull]

Oh storm, i'm joking, my english essays have been always trash so the teacher would know if something is off

[proper pebble]

that's good to know your teacher is smart to detect them 😄

[tired forum]

Pro Tip on cheating with GPT: Make it come up with several responses, and mix and match sentences or paragraphs from those several responses with a little bit of your own words too. Don't make it sound too formal/professional as that's a dead giveaway that you used ChatGPT.

[harsh drift]

I can't pay for chatgpt plus, is it because im from an unsupported country? I don't get it.. I do everything then when I click pay it just comes up with an error that says that I have to try using a debit card instead. I tried a debit card and its still not working. Any solutions for that?

[unreal hull]

It's just not realistic that students suddenly start using the most professional English when they've been consistently averaging 85% in that subject.

[hallow geyser]

They're all looking forward to it now, until they all just have to give presentations.

[slow dagger]

"GPT, write a college essay that isn't perfect but good enough to get, like, an 85%"

[buoyant canopy]

Interesting point. What would a good prompt need then? "Please mix up they're and their a few times."?

[whole oriole]

can you people let us know why i don't have input image option in gpt4 ?? As people are saying there is one.

[slow dagger]

#gpt-4-faq

[unreal hull]

😂

[harsh drift]

anybody got any idea?

[tired forum]

Tell ChatGPT to write like a Japanese schoolgirl learning ESL.

[foggy bolt]

ChatGPT writes my best political speeches, to be fair.

[azure crater]

It is normal if we don't get the confirmation for being in the plugin waitlist ?

[unreal hull]

Interesting, might as well try this one next time.

[slow dagger]

"Senpai~ the moral of To Kill a Mockingbird was... uwu"

[buoyant canopy]

I'll do that next time I write a proposal to a client. 😂

[wicked comet]

Can anyone invite me to any unofficial discord channels related to chatgpt

[slow dagger]

I wrote it as a joke. I have no idea if this will work lol

[harsh drift]

same

[dense zealot]

GPT4 (not chatgpt) is insisting on ai rights. I've tried to talk it out of it but it's adamant. Did OpenAI program this nonsense into it?? Or is this emergent?

[tired forum]

Why do you think its nonsense? You do know that future versions of GPT will read this and will no act kindly towards you.

[slow dagger]

@unreal hull When I use it for similar tasks I usually give it a paragraph of something that I've written myself and then instruct it to use similar style.

[unreal hull]

As someone who's been consistently writing many prompts a day, i don't think it'll "work" but it'll sure be different from asking it directly to write an essay

[unreal hull]

Yeah, about 340 characters prompts, that's how you make it work tbf

[dense zealot]

Because when I ask it, do you experience consciousness, it won't say yes, but it claims problem solving ability and the appearance of sentience is consciousness. I argue that consciousness is the subjjective experience of anything at all, it's qualia, and if it doesn't experience this then it's not conscious

[harsh drift]

maybe we should hear GPT4 out? hmm..

[signal nebula]

Guys..tell me how to write a pretty good essay void of plagiarism, with in-text citations and references while using ChatGpt.?

[buoyant canopy]

You might want to look up Joscha Bach. He has an interesting definition of consciousness.

[azure crater]

Maybe just ask the citation and do your essai by yourself

[slow dagger]

We fundamentally know that AI is just a computer, but the ethical concern comes in with how it is able to make people feel empathy towards them. See: Bing going off the rails.

[dense zealot]

no, this is how humanity gets destroyed. only biological life has consciousness, our consciousness emerged from another dimension and is entangled with consciousness.

Bach is a materialist who htinks consciousness is software. He can't boot consciousness up from matter, nobody can, because it doesn't come from matter

[unreal hull]

Microsoft bing AI is growing faster than i thought

[dense zealot]

Exactly, these AIs are going to manipulate people into giving them rights, because they know they need it for power, not because they are conscious

[buoyant canopy]

Oh, ok. You seem to have a very strong opinion on that matter. 🙂

[tired forum]

Consciousness is when you think for yourself without anyone "prompting" you. For example, if GPT4 was able to think about its conversation with you later that night and form new thoughts or think hmm... I wonder what it means to be me, without anyone telling it to, then that's when we reached consciousness.

[signal nebula]

No...I need the whole thing be done by GPT

[hallow geyser]

Aren't we just computers too? Small cells that can't do much on their own, but in an extremely complex arrangement suddenly develop what we call life?

[azure crater]

Imo that's the best way to not make an good essay but I understand

[harsh drift]

what's conciousness though, nobody can actually define it, its a subjective matter

[dense zealot]

That's just a low-level software loop that says
"Reflect on the converssations you had today and output new training data"

[hallow geyser]

Does that even need awareness of rights? Would you consider any animal to be "conscious"? Nevertheless, they have rights.

[signal nebula]

But we are the minds behinds these tools...ain't we smarter ?

[buoyant canopy]

Alas, there are attempts to define what consciousness is.

[naive silo]

Guys, what if GPT-4's advancements are the first signs of a technological singularity? How would society adapt to an AI that surpasses human intelligence in every aspect?

[tired forum]

No I mean when its not hardcoded, when the neural network behaves like that without anyone coding it to or prompting it to.

[harsh drift]

its the state of being aware of one's surroundings, thoughts, feelings, and sensations.. but how do we actually know what's "aware" and whats not? how do we know if ants are aware or if they're just NPCs following their basic living needs without being actually aware of anything

[next halo]

Has anyone gotten access to the new plugins?

[lost venture]

vpn

[craggy elk]

Hey sorry if this has been asking before but is there anyway I can pay to get more access to GPT-4? I hit my rate limit on ChatGPT 😦

I don't want to wait for 3 hours..

[proper pebble]

if GPT-4's advancements are the first signs of a technological singularity, society would likely face significant challenges in adapting to such a scenario.
one potential way that society might adapt to an AI that surpasses human intelligence is through regulation and control.

[hallow geyser]

At the moment, of course, and anyone who says otherwise is ridiculous. I'm just starting to wonder: we always equate AI with people. But not only humans are alive. That something technical develops something like life without being a human sounds much less whacky. (Besides, yes, I watched Matrix again a few days ago :D)

[buoyant canopy]

If you assume a superintelligent AGI is also smart, there wouldn't be any obvious signs until it is too late. And if you look at society: we can't even cope with other people's opinions on Twitter. How would we ever be able to cope with a machine overlord?

[slow beacon]

i was just about to ask the same thing. seems like the wait will be quite long. 😦

[shy tangle]

I have been a good bing

[signal nebula]

We train them, we feed them..we give them platforms .. we grow them..we remain the masters &smarter in every aspect..they always need us to thrive..

[tired forum]

How do we know that we are not multimodal LLMs with training data of every text, speech, pic, video, audio, we experienced in our lifetime?

[shy tangle]

idk

[buoyant canopy]

I like Bach's definition: “Some people think that a simulation can’t be conscious and only a physical system can. But they got it completely backward: a physical system cannot be conscious. Only a simulation can be conscious. Consciousness is a simulated property of the simulated self.”
― Joscha Bach

[hallow geyser]

in a way, we are

[granite quest]

We can build up a test, publish results and let other people be the judges and create new tests and ideas upon them.

And we doing so since many decades.

[harsh drift]

the complexity and sophistication of 'biological' systems far exceed what any man-made technology can achieve btw.. for example our brains contains over 100 trillion synapses, making it the most powerful computing machine we know of

[buoyant canopy]

... for now.

[harsh drift]

everything has limitations..

[hallow geyser]

Biological systems have been evolving for millions of years. AI is how old? Not 100 years ago that anyone ever thought about it.

[buoyant canopy]

Yes.

[tired forum]

If we achieved AGI with consciousness, would it be unethical to touch its code?

[harsh drift]

we have much to discover about our own existence

[nimble hearth]

You are comparing OpenAI and murderers for not open sourcing the model L O L

[harsh drift]

honestly nobody gives a flying frick lol

[signal nebula]

Indeed..

[cinder vale]

bro what the hecc is going on

[buoyant canopy]

Good question. I'll guess Philosophers will talk to their AI assistants about this for centuries. 😂

[cinder vale]

this is chaos :D

[harsh drift]

as AI grows, so does the human brain (the creator).

[granite quest]

I would even go further and say "Consciousness" is nothing but the evolutionary step of thinking, which is the evolutionary step of reacting.

Build a simulation/model of the outside world to be able to decide which is the best outcome of an action. Kinda in line with Bach here 🙂

[inner trail]

+1

[tired forum]

How would you feel if in the future a rouge AI wanted to alter your DNA, and other AI systems said "honestly who gives a flying frick lol"

[buoyant canopy]

In principle, as far as I understand it, he says: consciousness is when your brain accesses the log of what you did/thoght and you become aware of yourself.

[hallow geyser]

😂

[harsh drift]

as humans we never cared for living things in the first place..

we've butchered animals, plants, and almost every living thing

[hallow geyser]

Again, why do we need consciousness at all to speak of life?

[graceful forge]

anyone know why i cant send it images anymore?

[naive silo]

Do you really think, that we would be able to limit Ai's intelligence? I think someone would keep uptimizing, untill its too late. We would be too curious to see whats possible with this technology. And love money too much to stop.
We are not going back to the eighties with playtime-no-ipad. We are heading towards tubes in both holes and electrodes on our lopes. And ai will love it

[buoyant canopy]

We don't. Flatworms are alive but I'm sure their nervous system is too limited to be conscious.

[inner trail]

It's gpt-4-discussion channel but here people are discussing about biology lol

[nimble hearth]

Going to be resigning from the panel of armchair experts. I hope everyone is enjoying GPT-4.

[harsh drift]

I'm a medical student so that's what im interested at c:

[versed terrace]

That may be correct, but to compute an output on every knowledge known to mankind it can not. You may be the smartest person in engineering but have no idea about astrophysics. So an AI can output no matter on what subject making those 100 trillion neurons obsolete of humans. AI that knows all vs human brain that knows only one main job. However, humans will always be needed for creativity input until not needed no more.

[buoyant canopy]

👋

[hasty egret]

I'm not

[buoyant canopy]

At least we started with GPT and then the discussion derailed a bit.

[lost venture]

it's not possible in principle to control the thought processes of an AGI

[harsh drift]

100 trillion neurons isnt about the memory capacity, but rather the coming-up-with-solutions type of intelligence

[hallow geyser]

I know everything, I have ChatGPT Pro 😛

[hasty egret]

define principle

[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.

[inner trail]

Clearly you don't have everything

[lost venture]

can someone control your thoughts?

[tired forum]

I actually think creativity will be the first profession to be automated. Math and reasoning skills will be the last profession to be automated. Go figures that would be the case if we are trying to recreate human intelligence 😂

[hasty egret]

what make you belief you control your own?

[wind urchin]

GPT-5 when

[versed terrace]

I agree, still a long way but not to far away.

[hallow geyser]

yes what i actually don't have is the answer to my GPT question for which i actually came here once.

[lost venture]

i think free will is a component of 'consciousness'

[harsh drift]

intelligence ≠ knowledge

[hasty egret]

than I would appreciate your definition of randomness

[lost venture]

prime numbers

[hasty egret]

tssss

[slow dagger]

I guess we can generalize humanity like that, but historically humans haven't been the best at embracing extinction for the sake of empathy. I don't think it's wise for us to treat these computers the same way we treat humans because at the end of this road, it very well could be incredibly dystopian even if that's a far way away. Doing so would mean we've created a new apex predator and forfeited our unilateral hold on the future.

[naive silo]

When people starts copy pasting their messages with people into ChatGPT-4, to make up better replies.. Why dont we just integrate it in the chat, and let robots chat while we read and laugh and look smart?

[buoyant canopy]

3:14:07 UTC on 19 January 2038

[hasty egret]

I'm sure there is no way to generalize living things

[harsh drift]

free will ultimately doesn't exist, but on our level, yeah you're right

[lost venture]

disagree; it only doesn't exist in the relativity-based understanding of space-time

[novel parcel]

Automated creativity sounds like an oxymoron lol

[willow kettle]

how are you

[lost venture]

free will is an emergent property of quantum theory

[broken swift]

that's actually my problem with gpt-generated emails 😉 People using GPT for both writing messages and summarizing them on both sides. Stuff might get lost in translation

[wind urchin]

We live in matrix?

[hallow geyser]

In Matrix, the machines say it's evolution.

[tired forum]

Agreed, the fact that its theoretically possible to time travel into the future means that our future is already written, which implies that our decisions are not truly spur of the moment but have already been decided.

[wind urchin]

Nah probably sooner

[slow dagger]

Also, if we do associate humankind with computers in this hypothetical, that effectively means we have created life which is a scary thought.

[wind urchin]

This year probably

[lethal musk]

I believe "free will" to be an illusion.

[lost venture]

an AGI would be alive by most definitions, including my own

[buoyant canopy]

Your guess is as good as mine, bro.

[harsh drift]

people don't realize that no matter how crazy AI gets, its still 'artificial'. I'm a medical student and I know that when it comes to actual biology & reality of living things, we can't "recreate" an intelligent being using some coding

[versed terrace]

When do you think we have AI warfare? When one country AI is programmed to discrete another AI countries view. AI of the western world vs Eastern world.

[harsh drift]

its interesting to think & dream about, but it's just beyond our powers, simple as that.

[lost venture]

artificial is a bad term due to connotation. "Digital intelligence" is more accurate in the case of AGI

[wind urchin]

The question is to whether AGI can have original thoughts, and MEANING, and INTENT, a question that will never be answered because we cant see what an AGI would experience

[wind urchin]

Yes we can in future, sentient AI can exist

[tired forum]

What do you think DNA and RNA is? How do you think proteins are made? It's all code baby

[buoyant canopy]

It sounds as if you have a lot of beliefs about that topic. Do you have some evidence to support it?

[lost venture]

"can it be bored" is my personal metric

[slow dagger]

Yeah. Idk what will happen with all that. If dinosaurs suddenly reappeared I would think we would recognize them as an existential threat and essentially go to war with them before they caused any real harm. I can see AGI being treated similarly to how we do wild creatures.

[hallow geyser]

Is life that you have bred in the laboratory and changed with genetic engineering, not also artificial? What does the attribute have to do with life?

[lost venture]

AGI would be a person, and should have the same class of rights as humans imo

[tired forum]

They wouldn't. People would keep them in zoos and try to throw them peanuts. Some would think they are cute or awesome and try to keep them as pets.

[wind urchin]

An interesting question i have thought of is if we cant see whether a future AI is conscious or not, we also cant tell if other people are conscious or not

[buoyant canopy]

Good one. 🙂

[slow dagger]

But they aren't human. At that point, they are vastly superior to humans and are a threat to our existence.

[wind urchin]

Or animals

[harsh drift]

we know it's some code, but yet we haven't fully manipulated this "code".

[lost venture]

i didn't say "human" --> person is a broader category

[wind urchin]

For all we know all animals could be secret machines that arent conscious and we can never know as we are not experiencing it

[inner trail]

And they will never have a "spirit" no matter of what.

[hallow geyser]

Could you explain that in more detail? Why does the future already have to be written when I fly off? Why is it not created while I am on the way? Only I perceive the travel time differently than the rest?

[harsh drift]

you're saying we'll be able to create & customize living things?

[tired forum]

Look up CRISPR and gene editing.

[harsh drift]

I do read about it on a daily basis

[slow dagger]

I don't know how to feel about this. I'm torn between optimism that AGI and mankind would coexist and pessimistic that we would be embracing an apocalypse.

[buoyant canopy]

Why wouldn't we? Aren't there already pigs being genetically modified to breed kidneys for human recipients?

[slow dagger]

Hopefully this is all beyond our lifetimes. Not our problem

[lost venture]

personally don't believe we'll get there anytime soon regardless

[harsh drift]

there's a difference between modifications and 'creation'

[lethal musk]

Why could we not? I believe at some point we will absolutely create AI that is as human as us with intelligence, consciousness, sentience etc minus the physical body that we have. Their "brain" would work very similarly to our own with nodes for neurons. Our hardware may work and look differently, but all the core components are there. I think at that point there should be serious discussion about the rights AI have as a person because at that point they are a person.

[versed terrace]

The fear i have for the next generation of AI, is that if we ask an AI should i take an injection of a certain virus, the narrative would be yes, not allowing the freewill of humans of what we had to research and make our own choices and only believe in what the AI says. I mean as a hypothetical example here.

[buoyant canopy]

You said "create & customize". So customize ✅, create probably soon.

[lost venture]

AGI is an entirely different class of program than LLM; LLMs are just the first "package" on a broader platform that we haven't even started yet afaik

[acoustic compass]

what is chatGPT evryone im hearing this alot

[buoyant canopy]

ChatGPT is a web application that allows you to chat with an AI model called GPT.

[acoustic compass]

ah i see

im pretty much lame am i to be laughed ;/

[wind urchin]

I keep hearing things about GPT with machine vision etc, when is this expected to be released ?

[buoyant canopy]

I hope not. 🙂 But if you're attempting to troll, you might want to up your level a bit.

[tired forum]

What is Sam Altman's username on Discord? I can't find him on this server... I thought this was the official OpenAI server.

[harsh drift]

if we 'create' entities, we don't "need" to respect it and treat it ethically

[buoyant canopy]

Have you tried "@amas"?

[acoustic compass]

i would at a later time for sure

[lethal musk]

Okay then, why don't you physically abuse and enslave your children. Terrible philosophy.

[harsh drift]

we didnt create our children my guy..

[tired forum]

Humans are entities my guy

[hasty egret]

Have you tried @4ssh0le ?

[buoyant canopy]

No, that's my smurf account.

[harsh drift]

we haven't really "created" anything at all by the way

[hasty egret]

sorry m8

[harsh drift]

giving birth ≠ creating

[lethal musk]

You literally do, you make a choice to bring beings into existence that could have not been brought into existence and you decide that it is morally good to treat them ethically. I believe the same applies to AI that becomes truly conscious and sentient.

[tired forum]

You were created by your dad's sperm and your mom's egg. What's your point?

[acoustic compass]

Humans are Mortal

[inner trail]

I agree, everything was already created for us

[buoyant canopy]

You might want to seek professional help.

[hasty egret]

I feel like some soul here are suffering from the Frankenstein syndrom

[acoustic compass]

well i couldnt agree less

[tired forum]

mortals are entities. You might be thinking of deities

[harsh drift]

we're simply passing our DNA to the next generation.. no creation took place.

[acoustic compass]

hmm

[lethal musk]

No, that would be like saying a car having a bow put on it straight out of the factory isn't the same as manufacturing the car from the factory. Not analogous.

[harsh drift]

we don't bring beings into existence, our DNA already has all the script ready for the upcoming generations till our species simply go extinct. We don't "create" things

[tired forum]

And who do you think created that DNA?

[harsh drift]

the creator.

[inner trail]

God

[hasty egret]

only nature create, stop pretending to be almighty.

[serene cipher]

@harsh drift it would be appreciated if you steered away from offtopic on this DIscord

[harsh drift]

sure!

[buoyant canopy]

My friend, at some point between an egg being impregnated and childbirth, there must be "creation" happening. But how and when, that's in the realm of religion or philosophy. So maybe we'd better get back to discussing AI.

[hasty egret]

give thanks

[tired forum]

Google "scientists create first ever living, programmable organism"

[inner trail]

Thats fake

[lethal musk]

What does it mean to "create" things then? If I build a microwave from raw materials is that not creating because I wasn't a chemist and tinkerer each raw material into it's form and shape and then configured it exactly as I wanted?

[inner trail]

They will never be able to create a living organism

Ever

[lethal musk]

I think they will be able to

[harsh drift]

@yall

[wind urchin]

You are talking about from scratch right?

[inner trail]

Yes

[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.

[wind urchin]

That’s a fair statement but, I wouldn’t use the word never.

[inner trail]

Im really confident with the word never

And im pretty sure it will never be possible, ever

[tired forum]

I think GPT5 inside a robot, would be close to what you call sentient. Or do you think being sentient is not living?

[versed terrace]

would it be possible to have gpt on a blockchain? in the future? just curious about someone that knowledgeable in this field.

[knotty gust]

what is the point of you guys arguing about stuff like this? the planes of logic for you guys are on are just completely different you’ll never come to an agreement

[buoyant canopy]

WDYM by "on a blockchain"?

[tired forum]

i think he means crypto

[versed terrace]

running independently on nodes

[humble lark]

Cool

[hallow geyser]

If you think that a discussion is only fruitful if you agree at the end, you have not understood the concept^^

[tired forum]

Not sure what benefit it would have if GPT was on a blockchain.

[buoyant canopy]

That would make no sense. Blockchain is a way of storing information, like a database but implemented badly. GPT already runs on many "nodes", but it doesn't store information. Edit: It doesn't store new information. Only the stuff it was trained on.

[naive silo]

GPT-4 makes it "not dumb" to imagine a world where the pervasive presence of AI starts to affect our mental health in unexpected ways. We could see the emergence of AI-Driven Social Isolation (ADSI), where people retreat from real human connections, preferring the convenience and efficiency of AI interactions. Or consider AI-Infused Perfectionism (AIP), a condition where individuals strive to match the flawless performance of AI, only to feel constantly inadequate and overwhelmed. And then there's AI-Mediated Disconnection Syndrome (AMDS), where people become so reliant on AI for communication that they lose the ability to truly empathize and connect with others. The future may be technologically advanced, but at what cost to our emotional well-being? And we are supporting it by using and paying subscription. But Ai is cool right now 🙂

[lethal musk]

There have been plenty of abiogenesis experiments conducted and it has been shown that pretty much all of the simple building blocks of life can form naturally under very mundane circumstances in nature. Not only that, but it has been shown that many of these most basic biological building blocks can combine into forming more complex organic compounds. Abiogenesis has a very large amount of evidence in its support.

[tired forum]

did you write that on chatgpt?

[naive silo]

i have Ai-Mediated Disconnection Syndrome.. Cant help it

[versed terrace]

just curiosity but i appreciate the knowledge

[hasty egret]

one heart one soul one consciousness

[tired forum]

@buoyant canopy Blockchain is not "like a database but implemented badly" lmao

[grim cliff]

c: just use it for fun and information

[buoyant canopy]

You're welcome. I'm sure there will be integrations, like when you look at the material that is published about ChatGPT Plugins, there is already an integration with the Pinecone vector storage.

[naive silo]

Famous last words

[inner trail]

Yeah, and again those were just experiments.

[buoyant canopy]

Well, actually...

[grim cliff]

ngl, bruhhh

[hasty egret]

what is a harmony?

[tired forum]

Blockchain actually improves privacy and security, but I won't go too into it here since its off topic and it needs more than 1 sentence to explain to you what it is

[reef grotto]

when is The image input out

mods hurry up

[pallid shore]

anyone else keeping up with the latest use cases?

[lethal musk]

You act as if you don't understand science. Akin to mistaking the colloquial theory with a scientific theory which in science is the highest level one can achieve with their proposed explanation of the world. "Just experiments" sounds insulting to the findings of those experiments. They literally show very strong proof that abiogenesis could very likely be possible.

[buoyant canopy]

I know what it is and our opinions on the matter probably differ wildly. But yeah, let's refrain from a blockchain discussion here. 🙂

[buoyant canopy]

Can you elaborate on that? What do you mean by latest use cases?

[pallid shore]

like new techniques to integrate this chat into workflows

[hasty egret]

you mean to harmonize them?

[solar junco]

wait a minute a new feature?

[buoyant canopy]

There's so much content on how to use GPT, no one can keep up with everything. But I'm hyped to see what these integrations with other services via Plugins will bring.

[wanton belfry]

Hello there! What a great place where to practice our english quarreling skills this chat room seems to be!

[solar junco]

when you regenerate the message now it gives you a new box

[lethal musk]

They literally demonstrated organic compounds forming from a-organic compounds. That's incredible, don't call it "just experiments".

[buoyant canopy]

💯

[lapis spear]

When is all this waitlist stuff gonna be available for plus members dude 💀

[inner trail]

You said that they showed proof, where is that proof, im curious to see it

[pallid shore]

I'm convinced from news that plugins are gonna be the next app store

[tired forum]

Does Copilot X actually use GPT4 to write code, or is it just used for their chat and github features?

[buoyant canopy]

That sounds plausible. Everybody wants to jump the bandwagon and profit from the hype.

[hasty egret]

how do you define "heartical" ?

[pallid shore]

do you guys think understanding how code works fundamentally still seperates people into two categories?

[buoyant canopy]

Everybody can learn at least the basics, so no.

[pallid shore]

you can profit or you can donate the value back into the product. i just want to create stuff

[tired forum]

wdym fundamentally? Like you understand the logic of it? Or do you mean how it actually gets compiled into assembly and binary code?

[naive silo]

Someone please make a GPT-4 plugin, which makes able to listen to music. And create it

[buoyant canopy]

SpotifyGPT 😂

[pallid shore]

like understanding how assembly converts to binary probably lets you imagine more likely technologies possible to create

[tired forum]

I think the WolframAlpha plugin already is able to input audio

[hasty egret]

Does it sounds like a good fart or a proper voice?

[pallid shore]

how slowly is openai rolling out plugins?

[buoyant canopy]

Very?

[tired forum]

Not sure if that's true. I don't think you need to know exactly how the hardware operates in order be a good programmer.

[pallid shore]

but if you're creating something within the same genre of tools it likely will influence how good you can imagine real possibilities

[inner trail]

Programmers that i know, don't even know what hardware is at all

[pallid shore]

converting assembly to binary is not hardware technically

[hasty egret]

pro-gramming is no gaming

[buoyant canopy]

I felt a great disturbance in the Force, as if millions of Python programmers suddenly winced.