🔒 ai gonna delete programmers?

[heavy garden]

@south summit

Python help channel opened

Remember to:

• Ask your Python question, not if you can ask or if there's an expert who can help.
• Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
• Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

Closes after a period of inactivity, or when you send !close.

[restive sleet]

Your premise is false.

[south summit]

what

[restive sleet]

it's not a forgone conclusion that AI will get better than humans at programming. We seem to have already hit the ceiling with current methods.

[south summit]

wdym? Ofc Ai gonna get better then humans

[charred bay]

How do you know that?

Did you read multiple scientific papers or even conduct your own study and published your own paper?

[south summit]

You dont see how powerfull ai is getting???

[vestal quarry]

the current approaches will not lead to AGI which is what would really threaten our jobs

[south summit]

why should you need to read anything you can see it, ai is already better then 90% better then programmers

[vestal quarry]

and it's in question when or IF the big tech will actually publish anything AGI

[hearty slate]

Uhm at small tasks where it has all the context maybe

[vestal quarry]

yeah but that's a far cry from replacing programmers entirely

[south summit]

bro wtf. if we wait like 2-3 years ai gonna replace any programmers

[hearty slate]

What we are saying is it’s not a given that AI can actually improve much more with current models

[vestal quarry]

we'll see, but the AI companies are so invested in LLMs that I'm starting to doubt if they're seriously even building AGI based products

[hearty slate]

Have you read anything about this?

[bleak wagon]

Why would you doubt that the sun revolves around the earth when you can see it with your own eyes every day?

[south summit]

i know companies are replacing Ai instead of humans

AI is getting out of hand, its already massive. you can ask gemini to programm you something, what if companies built better Ai for their work

[restive sleet]

executives aren't scientists. they're "replacing people with AI" because they're buying into the hype and (more importantly) want to save money on labor.

[bleak wagon]

We are not even remotely at the point where we could realistically ask an LLM to implement the vast majority of changes necessary at my employer. At best (and even that is often questionable), it can make a human programmer work a bit faster on some tasks.

[south summit]

yea thats is my question. So why learning coding??

[south summit]

im also mean ai in 2-3 years

[restive sleet]

you're assuming that AI is improving at a certain rate, and that it will just continue at that rate uninterrupted until we have AGI. But that's never how AI has worked. AI progress happens in bursts where someone makes a breakthrough in model architecture, and then people do stuff with it until they hit the ceiling for how good AI can be with that breakthrough

it looks like we've hit the ceiling for the current breakthrough.

[south summit]

what if ai learn itself?

stop talking high level english

[bleak wagon]

There's good reason to believe that LLMs will never be good enough to autonomously solve that vast majority of professional programming tasks. And that they will continue to be useful only as a tool in the hands of a human programmer. There are fundamental limitations in the basic algorithm.

[restive sleet]

we're getting into what-if territory, but a moment ago you were certain that AI would keep improving for three years

[restive sleet]

I'm actually trying to "bring it down", but I'm an AI scientist by profession, so I'm not used to talking about it from an outside perspective.

[south summit]

just to show you how massive ai learns. you know the meme from social media, where will smith eats spagheti? if yes you need to see the difference about year 2023-2025. Rn you cant see any difference between real and ai.

[vestal quarry]

the changes necessary for AI to replace human programmers are fundamental in nature

and we're not really seeing those changes in today's models

[south summit]

human programmers have limits! ai doesnt

[restive sleet]

what do they mean by "task length"?

[vapid kite]

like the time itd take a human to do it

[vestal quarry]

ai surely has limits, LLMS certainly do

[restive sleet]

uh, what? AI absolutely has limits.

[south summit]

NO

you stupid?

[vestal quarry]

sure thing, mr. AI scientist

[south summit]

im not a scientist but ik what ik

[misty tapir]

progress in video generation does not directly translates to progress in language models, let alone AGI

and even if ChatGPT keeps improving, it has fundamental limitations with the way it works that just about guarantees it'll never reach AGI without at least one or two more major scientific breakthroughs

[bleak wagon]

Please refrain from namecalling.

[south summit]

it means how fast ai is learning

[vapid kite]

in theory ai is doubling its efficiency every 7 months

[restive sleet]

I feel like you're either ignoring or just refusing to believe what we're saying.

[south summit]

its a question?

[vestal quarry]

that's not ai learning, it's humans developing better AI models

[bleak wagon]

It's an insult

And you shouldn't do it

[south summit]

im soryy but its a question. you can see it of the ? .

[bleak wagon]

If you're not respectful, you're not going to continue having access to the server for very long.

[south summit]

im not disrespectful.

[restive sleet]

@south summit please take what dementati has said as an official warning about your behavior.

[vestal quarry]

the reason I have doubts about the big tech publishing any AGI models is precisely that AGI would be capable of improving itself, unlike the currently available models

[misty tapir]

btw what does de soziopath part in your username means dersoziopathlol?

[south summit]

you dont know it?

[misty tapir]

no, could you ellaborate?

[vapid kite]

also exponentially smaller

[south summit]

no just go to google ai he know it

[vapid kite]

models

[south summit]

I know what I see, but I am sure that AI will learn on its own.

[bleak wagon]

Even if models improve exponentially based on some metrics, it's not a given that those metrics measure something significant, in terms of questions like "will AI replace human programmers?"

Or that the exponential improvement will continue in the future

[vestal quarry]

the fundamental problem with current gen AI is that it cannot adapt to new challenges. All it does is regurgitate its training data in different combinations based on the prompt, thus faking intelligence

and it sounds convincing enough to fool a lot of people

[vapid kite]

yeah they should find a way for ai to learn in real time after the training phase that would be great

[vestal quarry]

but the architecture is the reason why the models are so ridiculously large

[bleak wagon]

I do think LLMs have useful capabilities beyond any previous algorithm or model types, but just because it can simulate human speech doesn't mean those capabilities include all those of humans.

[vapid kite]

already took over maths

[vestal quarry]

I don't think anybody's questioning the usefulness of LLMs, just whether they can actually replace humans in jobs that require creating thinking

[vapid kite]

many jobs are already being displaced

[vestal quarry]

mostly those that involve plenty of manual work

[vapid kite]

like translators or transcribers

[south summit]

isnt already ai taking humans jobs?

[hearty slate]

But not necessarily many programming jobs

[vestal quarry]

for translators I imagine AI doing 95% of their work

[restive sleet]

we already talked about why this is happening. if you're just going to ignore things that people say, please do not use this server.

[vestal quarry]

but that work is trickier if you want to do it right

[vapid kite]

ive rarely seen wrong translations personally by ai

[vestal quarry]

the AI translations I've seen are intelligible alright, but they always seem to contain weird glitches

stuff you'd need some context for to get right

[vapid kite]

isnt programming just translating human ideas into a coding language

technically also translation

[vestal quarry]

well LLMS deal with tokens, not ideas

they are based on language elements which is insufficient to represent ideas properly

and it also needlessly ties the model to a particular language

[south summit]

im anwering you question (only if i understand them or people dont ask me me 10 questions). and have many questions if you dont want to help me or answer my question you can leave the group.
im not disrespectfull to anyone im just asking.

[vapid kite]

imagine if neuralibk could turn human thoughts into comprehendable english

then u can link that to ai right

[hearty slate]

You have been disrespectful in this conversation.

[vestal quarry]

fair enough, so my answer would be "when/if AGI arrives, it can replace programmers"

[vapid kite]

soo when will that be

[vestal quarry]

that's the big question

[hearty slate]

We don’t know if it will ever happen

[vestal quarry]

we're not seeing any movement to that end

[south summit]

where?

[vapid kite]

2077

lol

[hearty slate]

You were actually already warned about it, scroll up

[vapid kite]

agi will release after gta 6

[vestal quarry]

big tech is either holding back their AGI stuff (OpenAI and others claimed to have "reached agi internally") or they don't know how to do that

[south summit]

i know, maybe read it again. maybe read the question mark behind the stupid. ITS a QUESTION.

[vestal quarry]

it's entirely plausible that they don't want to release that stuff, as there's no telling what would happen

[hearty slate]

Still a disrespectful question, you were warned for it

[vapid kite]

pfft ai companies say stuff to raise their stock prices

like agi is just around the corner

[south summit]

why disrespect?

[vapid kite]

(after 100 corners)

[vestal quarry]

yes which is why I haven't taken their claims seriously

[hearty slate]

Ask ai

[dry marten]

ai cant even do tricky and trivial code.
it just slop
Claude Code is a prime example.

or another ai bot

[vapid kite]

u either use ai or get replaced by ai

[vestal quarry]

I remember reading Ilya Sutskever's interview last year, and he seemed to believe in doubling down on the LLM tech. Last I heard he was building an ASI company in Israel or somewhere. Good luck with that 😛

[south summit]

no but i know im not disrespectfull against anyone. you dont have any argumuents 🤣

[vapid kite]

bec the competition will use ai

[hearty slate]

No, you get replaced by someone else who uses it

[hearty slate]

Asking someone if they are stupid is disrespectful

[south summit]

no

[dry marten]

at the end of the day, it depends how you use it.
applications of science can be either a blessing or a curse based on how u sue it.
this is true for everything made with science.
or is that different here?

[misty tapir]

they still fail to capture nuances quite often, and oftentimes lack context clues

[vapid kite]

hmm but what if they literally have ai chip in their brain

theyre like at least 50% ai

[dry marten]

it depends on the majority of people i guess.

[hearty slate]

It is

[south summit]

nice argumuent 🤣 .

[hearty slate]

I’ve been using Claude for a school assignment and yeah it does weird stuff

[vapid kite]

its all good when ai replaces others jobs but when its your job... ofc not

[hearty slate]

It’s not hard to see why asking someone if they are stupid, is disrespectful

[vestal quarry]

it is disrespectful though, as the question implies what you think of the person

[south summit]

who says im thinking this about that person?

[vestal quarry]

you wouldn't ask such a question if you didn't think the person just said something stupid

[south summit]

your right but im already said sorry, and if people dont see the reality im asking this.

[vestal quarry]

I think we can drop that already

[vapid kite]

i feel like its because of how LLMs are built, its like a neural network like a human brain but it cant like have neurons send signals backwards or make new connections with other neurons, each layer is forced to be the way it is it cant really adapt and thus cant really capture nuaces bec of its rigidness

[hearty slate]

You need to stop arguing with the people saying it was disrespectful and the people kindly answering your questions. Drop it

[vestal quarry]

did you get an answer to your actual question?

[vapid kite]

if ai could actually simulate the human brain it would be no different from it

but thats too process intensive

[south summit]

im sorry, but ai gonna replace the humans

[vestal quarry]

yeah, the architecture was made this way so we can use GPUs to process information efficiently

[vestal quarry]

eventually, perhaps, but not with the current gen tech

[vapid kite]

plus keep in mind humans require years of training (learning)

[vestal quarry]

not all the jobs at least

[hearty slate]

We have reason to believe what you just said is wrong, you should read some points brought up in this thread

[south summit]

im not saying thisdays i mean in possible 2-3years

[vestal quarry]

2-3 years is a bit too soon I think

[vapid kite]

gen Z will definitely experience agi imo

perhaps 20-30yrs later

the bad news is no job is safe the good news is no job is safe since now u dont have to work!!

lol

[vestal quarry]

it'll be interesting to see who then will buy stuff when everyone's unemployed

[south summit]

I wanted to be programmer or built a OS. but if i see AI is this MASSIVE its not worth and learning this fast is not possible to keep with it. ai already can handle so much code and i dont wanna know it in 2-3 years 💀

[hearty slate]

The thing is current models could be close to reaching a ceiling until another breakthrough happens

[vapid kite]

the so called AI "ceiling" doesnt exist

its a glass ceiling

[south summit]

so my last question is why learn coding? there is no use about it (fun doesnt count)

[vapid kite]

they keep moving the goalpost too

[vestal quarry]

with the current models I would say there is a ceiling

[bleak wagon]

I think it's not a given. There are certain technologies we've been waiting for that seem right around the corner, but they just keep getting pushed back. And some revolutionary technologies come out of nowhere. I don't think it's possible to gauge how difficult AGI is based on what LLMs can do.

[vestal quarry]

well, like, if you feel AI will take over all jobs, why learn any trade at all?

[vestal quarry]

fusion power? 😄

[bleak wagon]

For example

The complexity of the only known object in the universe with that level of intelligence is pure insanity.

Maybe our artificial version has to be similarly complex.

[vapid kite]

you know the movie ready player one? everyone lives in a virtual reality and u can buy suits to even get touch feedback so kinda like escapism. Even currently humans talk with ai girlfriends. My point is, while AGI is technically when AI surpasses human intelligence, wouldnt humans being fooled by fake worlds and ai bots be equivalent to AGI? when humans cannot tell the difference.

[south summit]

something that doesnt live and get inf old and also be possible to inf smart means its getting out of hand, ai is pretty sure soon learning on its own.

[vapid kite]

because even if humans are still "smarter" u have to account for human weaknesses suvh as emotions that AI doesnt have

[weary pumice]

If you are going to just prompt AI to generate something, how do you evaluate whether the AI’s output is: efficient,secure, or appropriate to the situation? Without fundamentally understanding what it's doing?

[vapid kite]

what kind

[vestal quarry]

again, this requires an AGI architecture

[south summit]

Humans are NOT smarter then ai. 🤣

[bleak wagon]

For one thing, we are also not even remotely close to having full dive VR.

[vestal quarry]

current models cannot improve themselves

[misty tapir]

without a major change in the models architecture or how they are trained, I don't think they'll improve significantly much more

there might not be a ceiling, but the progress will flatten with a breakthrough
with one, it could easily smash through the ceiling, but there is no telling when someone will discover another one

[south summit]

im talking in the future (2-.. max:10 years)

[vapid kite]

VR is literally almost there. sight and sound are the 2 main senses and once those are tricked its over

its js too expensive and not enough developmebt rn

[south summit]

its possible if humans wont get any help or wont ai is getting out of hand but there is someone or anyone who dont want this and ai is learning on its own

[vestal quarry]

the thing with full dive VR is the control which is severely lacking currently

the controls are pitiful – ever tried fighting in skyrim VR?

[vapid kite]

would u guys say that ai is smarter than some ppl?

[bleak wagon]

Strongly disagree. I think we are very very far from full dive VR. We are just putting computer screens in front of people's eyes. Full dive VR would require some kind of direct neural link, and the basic research is lightyears away from that.

[vestal quarry]

I've talked to a lot of people who are equivalent to 7B models.

[vapid kite]

xD

no way

[south summit]

im saying ai is already smarter but people can learn on its own.

[vapid kite]

are u talking about the delay?

[vestal quarry]

huh?

[vapid kite]

why wouldnt a computer screen be sufficient with enough refresh rate and resolution

u cannot tell the difference

[vestal quarry]

did you not read anything I said about control?

you have these pitiful controllers which are a far cry from moving your body freely in VR

[vapid kite]

im replying to the other msg sry

[vapid kite]

i dont think controls are the bottleneck

[vestal quarry]

they absolutely are

[vapid kite]

u know actors have these suits that record every movement

[vestal quarry]

do you own a VR rig?

[vapid kite]

and then they put it on anime

[prisma cape]

where is this "ai learning on its own" exactly tbh

[vestal quarry]

that's very different, you cannot move freely in VR

and locomotion blows in current gen TBH

there are expensive, bulky "solutions" to that which are still far from free movement

[vapid kite]

VR doesnt make sense yet but

very soon

[vestal quarry]

how soon?

[south summit]

people can create (learning on its onw ) its like a algorithm

[vapid kite]

ppl love escapism they hate their life (some ppl)

[misty tapir]

I could agree with them being "smarter" than most people that do not have any specialized knowledge, depending on how you define smartness, but has less potential to learn new things than nearly anyone (excluding extreme cases like humans with dementia)

[vapid kite]

they js need smth better

[hearty slate]

Ai is not learning on its own

[vapid kite]

smth good enough that they can ignore minor details

[south summit]

i never said it

[vapid kite]

like yk those ppl who have ai girlfriends

[prisma cape]

which algorithm specifically allows ai (I assume llms) to learn on its own

[vapid kite]

ppl actually suicided and got attached to them

many cases

[jovial gorge]

6/10 ragebait

[south summit]

i never said ai is learning on its own but if it do were cooked.

im done

[vestal quarry]

you did, multiple times

[hearty slate]

@south summit do you actually have any technical knowledge of how LLMs work, if not I don’t see why you are arguing with people who do and are trying to answer your questions

[jovial gorge]

👀

Do such thing actually exist

[south summit]

no I SAID (if) ai learns on its own.

[bleak wagon]

I think you can draw comparisons between VR and LLMs. They both have the trappings of what we imagine from movies and TV shows, but the core tech is a far cry from it.

[vapid kite]

I WISH I WAS JOKING

because they never disagree with them

[jovial gorge]

Ok that's crazy
I gotta search about that later

[south summit]

im sorry for any disrespect. and i want to leave the group but i can so idk.

[bleak wagon]

No worries.

[vapid kite]

people are becoming more reliant on AI by the day

asking ai for basic tasks

ofcourse its not all

but if half the population is reliant on AI

[prisma cape]

the problem is it doesn't

• we can reasonably see that llms won't keep getting better at the same rate it has been
  hence it's reasonable to believe llms won't delete programmers

[vapid kite]

that is literally a dystopia

doublethink

[bleak wagon]

LLMs certainly are very interesting for a variety of reasons, and their societal impact is very real. It's just not because they are close to being AGI.

[vapid kite]

and lets say humans become dumber bexause of their reliance on ai

then agi will be even sooner

bec agi is js being smarter than humans

[bleak wagon]

That's a bit oversimplified. When we say humans are dumb, we mean compared to other humans.

[vapid kite]

even rn ppl are addicted to ai algorithms on tiktok, insta, youtube etc

[bleak wagon]

Not compared to like, a vacuum cleaner.

[vapid kite]

imagine how bad when it starts to reason

if companies use it for profits

scaryy

[vapid kite]

but dont u think the society can become dumber

i heard a study that said human brains are smaller now

than they used to be

[prisma cape]

ok but what does people getting dumber have to do with llms being closer to agi
like, imagine this oversimplified example where an intelligence of 100 means agi, and current llms are at like 10
humans getting dumber doesn't change the fact that llms are still stuck at 10

[vapid kite]

humans can go to 9 ez

[vapid kite]

imo llms wont be stuck on 10 if u want a serious view

[prisma cape]

humans can go to -10 billion for all we care in this example
llms would still be at 10, extremely not agi

[bleak wagon]

I think it's a mistake to even put LLMs and humans on the same scale.

[vapid kite]

what is agi

in ur terms

[bleak wagon]

Their "intelligence" is of a fundamentally different nature.

[vapid kite]

Artificial general intelligence (AGI)—sometimes called human‑level intelligence AI—is a type of artificial intelligence that would match or surpass human capabilities across virtually all cognitive tasks.[1][2]

[prisma cape]

it doesn't actually matter here, for the argument of "dumber humans = closer to agi"
I'm specifically telling you that dumber humans has nothing to do with whether llms get closer to being agi (whatever agi might mean)

[bleak wagon]

Yeah, something along those lines. Something that's as good as the average human at all the things the average human can do.

[prisma cape]

huh, ig it does matter
though it would be very uninteresting if we reach agi like that

[vapid kite]

the robotics is underdeveloped for sure tho nowhere near humans

but the brain is closer

[bleak wagon]

If we reach the point that the average human is as capable as a current LLM, then we'll die out as a species.

[vapid kite]

why so

[bleak wagon]

Because an LLM can't manage a farm.

So we'll have no food.

[vapid kite]

are u sure?

[bleak wagon]

Yes.

[vapid kite]

theres machines that sow seeds, that harvest crops, that sprinkle water, and u can just use a set schedule

all u need is ai driving to transport the goods

[bleak wagon]

The LLM will have to build those machines as well.

And run the factories that produce the parts.

[vapid kite]

moving the goalpost >:(

[bleak wagon]

No?

We're talking about a human civilization run by LLM-level agents.

[vapid kite]

humans arent extinct yet tho how will humans go extinct is the question

humans can do the machines ai can be the nanagers

[bleak wagon]

The scenario was that the average human became as dumb as a current LLM.

[vapid kite]

while humans are blue collar

ai is white collar

:)

[bleak wagon]

Humans are still as smart as an LLM, so a blue-collar human would be incapable of doing blue-collar work.

[vapid kite]

technically the ai knows how to run a factory

it just cant physically

most of it is automatic anyway

[bleak wagon]

Extremely doubtful.

Or rather, 100% impossible.

[vapid kite]

isnt there a car factory thats so automated it doesnt even have lights on

because it doesnt need lights

for humans

[bleak wagon]

A factory being automated in general and a factory being run by an LLM are very different things.

[vapid kite]

u need multiple LLMs for each subdivision/task

but if u have them its possible

they could even fix stuff

[tough bramble]

Yeh.... actually u are right

Also...I sometimes feel potent that I might have picked a dying industry

[bleak wagon]

It would fail miserably.

[vapid kite]

i just hope theres a UBI by the time AI takes over

idk search up dark factories

in my view factories definitely wont be the main problem

they would js need toaintain the existing ones

[bleak wagon]

It's not that you can't automate large parts of factories.

It's letting LLMs run them.

That's the problem.

[vapid kite]

and why is that a problem

what exactly cant an LLM do

that a human can

[bleak wagon]

You can write programs that perform specific tasks extremely well and reliably.

LLMs are not reliable.

[vapid kite]

humans also make midtakes

[bleak wagon]

They are language generators.

[vapid kite]

if LLMs have high temperature

they will eventually be lucky and get the right solution lol

i mean itd be nice if AI could learn on its own

but thats not inpossible either

also halluciantions are a problem

but that problems decreasing

[vapid kite]

the language generators are just managers

human managers also dont ever touch machinery

they js talk with the workers

the llm will generate instructions/code for machibes/robots

its not something conpletely out of the question

[ionic pike]

Oh neat. Dark factories mentioned. As someone who works in a car factory, there are still a lot of people involved

[solar phoenix]

how to find this?? im curious thanks

[vapid kite]

https://hai.stanford.edu/ai-index/2025-ai-index-report

[ionic pike]

https://www.theverge.com/ai-artificial-intelligence/827820/large-language-models-ai-intelligence-neuroscience-problems

[solar phoenix]

ohh yeah ai train each other so ai is rapidly evolving this current era

[vestal quarry]

I like to point out that the "thinking" models think aloud

[prisma cape]

the issue is you can seemingly boil mostly anything down to a language task, and you assume that because of this llms can do it

[vapid kite]

https://sherwood.news/tech/how-do-ai-models-stack-up-vs-humans-on-standardized-benchmarks/ and this one

[vestal quarry]

while people don't have to form words in their minds to think

[vapid kite]

u can also reverse image search

next time

[vapid kite]

people who are blind/deaf and never learned a language said they thought using images

which is not as precise as language in my virw

i think language made it easier to form complex thoughts

some areas like spatial thinking yeah language is suboptimal for sure

LLMs could be combined with image models

and then later voice models like sesame that can hear emotion/tone

sorry im js super interested in ai 😅

[solar phoenix]

your right Images can convey emotions and spatial concepts effectively but often lack the nuance and structure of language

[ionic pike]

The key takeaway is that language is based in communication, not thinking

[white remnant]

ai cant replace programmers bro they need us to be used

or to be crated

never saw an ai that can create an ai

like it would be a revolution

[slender marten]

"why do I ahve to read, I can see it" is about as good an example of a bad argument that ought not be paid any attention as I can imagine

[near walrus]

its in the beginning that ai can learn from itself and avoid errors ...
and no ai can replace programmers ... it need only people that know how the ai tool work ^^ oc it need programmers to train

[near walrus]

comfy-ui a node based GUI is able to do all is in your mind 😉

The annoying thing is: if you ask the AI, but there is no R2D2 or other robot nearby, ‘bake me a bun’, you may see a picture and hear chewing noises, but you won't get anything to eat 😉

[white remnant]

yeah ik

[gray swift]

queue the boom of augmented intelligence / IA

[stuck nacelle]

Which is harder, solving the Riemann hypothesis or reaching AGI?

[gray swift]

non-dev answer, but riemann's hypothesis is slightly below p=np for complexity. I would defer to research already done by IBM or Google. youll find more complex arguments through IBM's research (where they initially had goals of rainbow tabling w/ quantum chips to break encryption standards, which iirc was popped 2 years ago, but may have resurfaced w/ Google's Willow Project a few months back re: RSA 2048) . agi is more on the band of IBM and their quantum studies, but coupled with things like Google's Willow chip and Moore's law ; were headed somewhere exponentially fast.

[bleak wagon]

I don't really understand how likely or close we are to resolving the Riemann hypothesis, but it seems pretty clear to me that we have no clue what is necessary to build an AGI.

[restive sleet]

It seems that OP checked out and that this thread went off in another direction.

!close

[heavy garden]

Python help channel closed with !close

This help channel has been closed. Feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.