Live data from Hacker News

John Carmack: I’m going to work on artificial general intelligence

facebook.com

881–890 of 930 posts

Re: John Carmack: I’m going to work on artificial general intelligence

#881

Earlier quoted context omitted.

Yeah I believe in this game / simulated world NPC idea too. To get the kind of complexity we want we either need sensors in the real world or interfacing in a virtual world that humans bring complexity to (probably both -- the humans are part of the sensing technology to start). Things like AlphaZero etc. got good cuz they had a simulatable model of the world (just a chess board + next state function in their case).…

Perhaps the first sentient program will be born in an MMORPG?

Just imagine all the exploits they'll find and abuse.

OpenAI has already done some experiments here [0]. All the way down at the bottom, under the "surprising behaviors" heading, 3 of the 4 examples involve the AIs finding bugs in the simulation and using it to their advantage. The 4th isn't a bug exactly, but a (missing) edge case in their behavior not initially anticipated.

[0] https://openai.com/blog/emergent-tool-use/

Re: John Carmack: I’m going to work on artificial general intelligence

#882
post #569

Earlier quoted context omitted.

"Researchers in England may have finally settled the centuries-old debate over who gets credit for the creation of calculus. For years, English scientist Isaac Newton and German philosopher Gottfried Leibniz both claimed credit for inventing the mathematical system sometime around the end of the seventeenth century. Now, a team from the universities of Manchester and Exeter says it knows where the true credit lies —…

That story's by non-experts and sounds like it's based on a press release. There were basic components of calculus well before that too: https://en.wikipedia.org/wiki/History_of_calculus However, calculus proper (derivatives and integrals of general functions, and the connections between them) did not exist until Newton and Leibniz. Other mathematicians made important steps towards it earlier in the 1600s, and if New…

https://insa.nic.in/writereaddata/UpLoadedFiles/IJHS/Vol47_4...

https://en.wikipedia.org/wiki/Aryabhata

Re: John Carmack: I’m going to work on artificial general intelligence

#883
post #507
post #487

Earlier quoted context omitted.

The need to generate publishable papers means that a researcher can only participate in activity that leads to such a paper. He can't try to work on that idea for 5 years, because if no big papers follow, he's toast /he'd probably lose funding long before that).

You have to earn the right to work on your idea for 5 years and get paid. Otherwise we would be funding all kind of crackpots. First you demonstrate you're a good researcher by producing good results. Then you can work on whatever you feel like (either by getting hired at places like DeepMind, or by finding funding sources that want to pay for what you want to work on).

This is what I meant. In our society, only very few, usually already rich, can try their own ideas. Most of us have to stick with known ideas that bring profit to business owners or meaningful visibility to universities. When I was in college, I had to work on ideas approved by my professor. Now I have to work on ideas approved by my corporation. But if I had money, I'd work on something completely different. Sure, in 15 I will be rich and can start doing my own stuff, but I'll also be old and my ability will be nowhere near the peak at 25 years.

Re: John Carmack: I’m going to work on artificial general intelligence

#884
post #883
post #507

Earlier quoted context omitted.

You have to earn the right to work on your idea for 5 years and get paid. Otherwise we would be funding all kind of crackpots. First you demonstrate you're a good researcher by producing good results. Then you can work on whatever you feel like (either by getting hired at places like DeepMind, or by finding funding sources that want to pay for what you want to work on).

This is what I meant. In our society, only very few, usually already rich, can try their own ideas. Most of us have to stick with known ideas that bring profit to business owners or meaningful visibility to universities. When I was in college, I had to work on ideas approved by my professor. Now I have to work on ideas approved by my corporation. But if I had money, I'd work on something completely different. Sure, i…

What would you work on if you could? Would you say you deserve to be paid for 5 years of uninterrupted research? Do you think you have a decent chance to make a breakthrough in some field? These are the questions I ask myself.

Re: John Carmack: I’m going to work on artificial general intelligence

#885
post #487

Earlier quoted context omitted.

The need to generate publishable papers means that a researcher can only participate in activity that leads to such a paper. He can't try to work on that idea for 5 years, because if no big papers follow, he's toast /he'd probably lose funding long before that).

When you have tenure, you can work on whatever you want for as long as you want. Nobody works on an idea for five years without publishing anything, though. Progress is made step by step. Take Albert Einstein as an example, who arguably made one of the largest leap in physics with his theory of general relativity. He never stopped publishing during that time.

When you have tenure, you can work on whatever you want for as long as you want

Not quite. When you are a professor, you essentially become a manager for a group of researchers. You don't really do research yourself. Therefore, your main obligation becomes finding money to pay these researchers. So in reality you can only support the research someone is willing to pay for (via grants, scholarships, etc).

Re: John Carmack: I’m going to work on artificial general intelligence

#886
post #12

Earlier quoted context omitted.

You forgot "he who shall not be named"... /ok, maybe his project falls under "total bullshit"...

Who are you referring to? I feel out of the loop here.

Those of us who've been on the net long enough, and have at least dabbled in AI/ML circles, know of him.

He claims to have invented a program that is a mind, originally written in Forth, translated by others to many other languages, etc. He has published the code of this program in a form of "open source", so you can easily find it if you dig enough.

He's widely considered to be a crank. That said, the line between genius and madness can be mighty thin, and what side he lands on is anyone's guess, but most put him well over the line into madness, for whatever its worth.

My own opinion?

Well - looking at his work from purely the modern understanding and research into ML/AI - that is, deep learning and such - his work would be considered pointless, probably worse than Eliza as to its contributions to the field.

But as someone who has read a lot of other works (for for and against) the idea of AGI, artificial consciousness, theory of mind, etc - his work at a certain level has echos with some of that work. Still probably a dead end, but at the same time, there's some interesting concepts within his code and theories (he's self-published a book on it, too - you can find it on Amazon - he also has it for free on his github and it can be found elsewhere).

I guess I still put him in crank territory, but not in the abusive crank arena, more in the "doing his own thing, but being a bit evangelical about it" - relatively harmless.

His work is not as amazing as TempleOS, imho, but there's a similar mind behind it (though comparing it with that operating system is maybe an unfair, possibly orthogonal, comparison).

I won't say or reveal more (but I've written enough for you to figure it out) - he tends to monitor tons of forums and if he thinks he's being "summoned", he'll spam the forum with his writings and theories. It got him "perma-banned" from more than one newsgroup back in the day...

Re: John Carmack: I’m going to work on artificial general intelligence

#887
post #12

Earlier quoted context omitted.

You forgot "he who shall not be named"... /ok, maybe his project falls under "total bullshit"...

Who are you referring to? SOAR? Contrary to popular belief, both openai and deepmind have zero roadmap and specification of a cognitive architecture, not even a semantic parser.

Please see my reply here:

https://news.ycombinator.com/item?id=21542453

Re: John Carmack: I’m going to work on artificial general intelligence

#888
post #415

Earlier quoted context omitted.

Here's another one for ya if you get stuck with a case of the nosleeps - https://www.youtube.com/watch?v=7wKu13wmHog Something about predatory nature of both insects seems to tune up their intelligence. Of course it never hurts having the BBC tell your story either.

>Something about predatory nature of both insects seems to tune up their intelligence. Yep. To be a predator, you need to outwit your prey and think fast, so it's thought to be a natural INT grinder. `w´ Presumably, this could drive up the INT of prey too, but maybe it's cheaper to just be faster/harder to see? But you can't be THAT hard to see, and the speed only saves you in failed ambushes, so planning successful…

What is I.N.T.? I couldn't find a definition.

Re: John Carmack: I’m going to work on artificial general intelligence

#889
post #569

Earlier quoted context omitted.

That story's by non-experts and sounds like it's based on a press release. There were basic components of calculus well before that too: https://en.wikipedia.org/wiki/History_of_calculus However, calculus proper (derivatives and integrals of general functions, and the connections between them) did not exist until Newton and Leibniz. Other mathematicians made important steps towards it earlier in the 1600s, and if New…

https://insa.nic.in/writereaddata/UpLoadedFiles/IJHS/Vol47_4... https://en.wikipedia.org/wiki/Aryabhata

These are interesting articles that seem to agree with what I said. The first one defines calculus in a much more limited way, and refers to some of the earlier basic components I mentioned.

I'm not a historian, but a few months ago I spent some time analysing one of Fibonacci's trigonometric tables (chords, not sine or sine-differences). Aryabhata's sine-differences were much earlier.

Re: John Carmack: I’m going to work on artificial general intelligence

#890
post #798

Earlier quoted context omitted.

AI sees humans as thread, all humans die.

> AI sees humans as thread We shall all be woven into the fabric of the new artificial reality.

Sounds like a comfy retirement I can look forward to
Post reply on HN