this post was submitted on 26 Aug 2023
818 points (91.5% liked)

Programmer Humor

19503 readers
410 users here now

Welcome to Programmer Humor!

This is a place where you can post jokes, memes, humor, etc. related to programming!

For sharing awful code theres also Programming Horror.

Rules

founded 1 year ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] [email protected] 25 points 1 year ago (3 children)

It just occurred to me that AI in the nearish future will probably/almost certainly be able to do this.

[–] [email protected] 36 points 1 year ago (1 children)

I can't wait for AI to make a PC port of every console game ever so that we can finally stop using emulators.

[–] [email protected] 8 points 1 year ago (6 children)

This won't happen in our lifetime. Not only because this is more complex than rambling vaguely correlated human speech while hallucinating half the time.

[–] secret301 4 points 1 year ago

I think it'll be in our lifetime just not anytime soon. I feel like AI is gonna boom like the internet did. Didn't happen overnight and not even in a year but over 35ish years

[–] [email protected] 3 points 1 year ago (1 children)

Idk the specifics, but what you say makes it sound like it would be easier to create an AI that recreates a game based on gameplay visuals (and the relevant controls)

[–] [email protected] 1 points 1 year ago

That game would still not work because there is a ton of hidden state in all but the simplest computer games that you cannot tell from just playing through the game normally.

An AI could probably reinvent flappy birds because there is no more depth than what is currently on screen but that's about it.

[–] [email protected] 3 points 1 year ago (2 children)

Off the shelf models do this, yes.

Sophisticated local trained models on expensive private hardware are already dunking on publicly available versions. The problem of hallucination is generally resolved in those contexts

[–] [email protected] 4 points 1 year ago (1 children)

Sure but until I see such a thing I chose not to believe in fairy tales.

Decompiling arbitrary architecture machine code is quite a few levels above everything I've seen so far which is generally pretty basic pattern recognition paired with statistics and training reinforcement.

I'd argue decompiling arbitrary machine code into either another machine code or legible higher level code is in a whol other league than what AO has proven to be capable of.

Especially because with this being 90% accurate is useless.

[–] [email protected] 2 points 1 year ago (1 children)

Again you aren't seeing this because these models are being developed for private enterprise purposes.

Regarding deep machine code analysis, sure, that's gonna take work but the whole hallucination thing is an off the shelf, rookie problem these days

[–] [email protected] 1 points 1 year ago (1 children)

It's not, though. Hallucinations are inherent to the technology, it's not a matter of training. Good training can greatly reduce the likelihood, but cannot solve it.

[–] [email protected] 1 points 1 year ago

Training doesn't solve hallucination. I didn't say that

[–] [email protected] 1 points 1 year ago (1 children)

Why does a pre-trained model need expensive private hardware after it was trained, other than to handle API requests faster? Is Open AI training chat-GPT on inferior hardware compared to these sophisticated private versions you mentioned?

[–] [email protected] 3 points 1 year ago

The fine tuning, while much more efficient than starting fresh, can still be a large amount of work.

Then consider that your target corpus of data may also be large.

Then consider to do your reasoning tasks across that corpus also takes strong hardware to get production ready response times.

No, openai isn't using inferior hardware, but their model goals, token chunking strategies and overall corpus are generalist in nature.

There are then processing strategies teams are using to go beyond the "memory" limitations gpt 4 has, that provide massive benefits to coherency, essentially anti hallucination and better overall reasoning

[–] [email protected] 2 points 1 year ago

Ai prompt: make me a program that will convert PS5 games to PC

AI: Use Convert-PS5GameToPC

End of line

[–] [email protected] 0 points 1 year ago (1 children)

AI can literally read minds. I don't think it's that great of a step to say it should be able to decompile a few games.

[–] [email protected] 1 points 1 year ago (1 children)

About half the time, the text closely – and sometimes precisely – matched the intended meanings of the original words.

Don't be surprised but about half of the time I can predict the result of a coin flip.

I'm not saying it's not interesting but needing custom training and an fMRI is not "an AI can read minds"

It can see if patterns it saw previously reappear in a heavily time delayed fMRI. Looking for patterns you already know isn't such an impressive feat Computers have done this for ages now.

It litterally can't read minds.

[–] [email protected] 2 points 1 year ago* (last edited 1 year ago) (1 children)

Later, the same participants were scanned listening to a new story or imagining telling a story and the decoder was used to generate text from brain activity alone. About half the time, the text closely – and sometimes precisely – matched the intended meanings of the original words.

You left out the most important context about "half of the time". Guessing what you're thinking of by just looking at your brain activity with a 50% accuracy is a very very good achievement - it's not pulling it out of a 1 or 0 outcome like you're with your coin flip.

You can pretend that the AI is useless and you're the smartest boy in the class all you want, doesn't negate the accomplishments.

[–] [email protected] 1 points 1 year ago

Being close (and "sometimes" precise) to the intended meaning is an equally useless metric to measure performance.

Depending on what you allow for "well close enough I think" asking ChatGPT to tell a story without any reading of fMRI would get you to these results. Especially if you know beforehand it's gonna be a story told.

[–] [email protected] 10 points 1 year ago* (last edited 1 year ago) (2 children)

It was a staple of Asimov's books that while trying to predict decisions of the robot brain, nobody in that world ever understood how they fundamentally worked.

He said that while the first few generations were programmed by humans, everything since that was programmed by the previous generation of programs.

This leads us to Asimov's world in which nobody is even remotely capable of creating programs that violate the assumptions built into the first iteration of these systems - are we at that point now?

[–] [email protected] 9 points 1 year ago (2 children)

No. Programs cannot reprogram themselves in a useful way and are very very far from it.

[–] [email protected] 2 points 1 year ago (1 children)

Eh, I'd say continuous training models are pretty close to this. Adapting to changing conditions and new input is kinda what they're for.

[–] [email protected] 1 points 1 year ago

Very far from reprogramming though. The general shape of the NN doesn't change, you won't get a NN made to process images to suddenly process code just by training it.

[–] [email protected] 1 points 1 year ago (1 children)

Then how does polymorphic/self-modifying code work?

[–] [email protected] 1 points 1 year ago (1 children)

It doesn't or do you have serious applications for self-modifying code?

[–] [email protected] 1 points 1 year ago

Some use it for causing millions of dollars in damage.

[–] [email protected] 2 points 1 year ago

are we at that point now?

Nope, but we're getting there.

[–] [email protected] 1 points 1 year ago