Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

venturebeat.com

Post reply on HN
Print view View original

131–140 of 181 posts

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#131

Post by why_only_15 »

sanxiyn wrote: post #108
taneq wrote: post #102

Earlier quoted context omitted.

Hey, at least it's a type of problem rather than a problem itself. I haven't studied enough myself yet to know the answer to this one, but what are the differences between AlphaZero and the OpenAI 5 DOTA team's approach? Would it be possible to apply AlphaZero to DOTA?

DOTA is partially observable, so I believe AlphaZero can't be applied, as-is.

I wonder what would happen if you made DOTA totally observable. You could probably reformat it as 5 pieces for a player instead of 5 people on a team or the like. It would probably change the game too much to be recognizable as the same, but I think it would be an interesting experiment if nothing else.

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#132

Post by mindcrime »

qwerty456127 wrote: post #104

If AGI (an artificial human mind with direct access to computational power of classic computers and whole Internet of information) was possible then we would probably already be living in the Travelers TV show.

... was possible then we would probably already be living in the Travelers TV show.

How do you know we aren't?

BTW, if you hadn't noticed, Season Three just came out on Netflix. I'm champing at the bit to binge watch that... :-)

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#133

Post by virgilp »

cr0sh wrote: post #122
epicureanideal wrote: post #113

https://en.wikiquote.org/wiki/Incorrect_predictions "Hence, if it requires, say, a thousand years to fit for easy flight a bird which started with rudimentary wings, or ten thousand for one which started with no wings at all and had to sprout them ab initio, it might be assumed that the flying machine which will really fly might be evolved by the combined and continuous efforts of mathematicians and mechanicians in f…

Interestingly, Hinton is on record as essentially saying that there's a good possibility that what's currently being done is wrong - and that we need to rethink our approach. Mainly in the idea/concept of back-propagation. It's something that I've thought about myself. For the longest time, I could never understand how it worked, then I went thru Ng's "ML Class" (in 2011, which was based around Octave), and one part…

You're contradicting yourself with your examples. If we didn't manage to fly by imitating birds - why do you care that AI doesn't work the brain does? That should be a _good_ sign, if we trust the analogy - right?

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#134

Post by random314 »

nuguy wrote: post #115
alexnewman wrote: post #112

Earlier quoted context omitted.

"These are people who have narrow expertise in one framework of AI." Proof that you don't know who you are insulting

I don’t remember insulting anyone. And how is that not true?

Geoff Hinton is the grandfather of deep learning. Virtually all the modern advancements in AI can be traced back to him and his lab.

What is your track record in AI? It sounds like you have no technical knowledge of AI. For example do you understand the concept of cross entropy loss?

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#135

Post by why_only_15 »

mortivore wrote: post #28
izzydata wrote: post #24

I'm not even convinced that a real AI is possible with conventional computer hardware or anything remotely similar to it. Not even considering software I get the impression there is a fundamental limitation of hardware.

I'm not convinced we've even defined the problem space well enough to solve it. Like what is the concrete measure(something to target) for intelligence? If we develop general intelligence is it going to be human, dog, or fish?

It seems like the difference between humans and dogs is substantially smaller than the difference between computers and dogs, so if we figure out dog-level intelligence human level intelligence is right down the corner. Also, the intelligence is likely to be of a different kind. Someone had an interesting point that training an ML system to look at picture isn't like sending a million interns to look at a million pictures, it's sending one intern to look at a million pictures. When you do that, you can derive insights that are significantly different than if you look at 1 picture, or 10, or 100.

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#136

Post by mindcrime »

nradov wrote: post #121
nuguy wrote: post #81

Earlier quoted context omitted.

Even anti-alarmists don’t ask for proof that AGI is possible. Obviously it is possible. Speculation is the best you get because nobody is going to be able to prove anything. We haven’t proven global warming is caused by humans but it’s still worth it to be proactive about greenhouse gasses. This is because when something is extremely dangerous, you don’t wait around for someone to finish proving it beyond any shadow…

I ask for proof that AGI is possible. Show me a computer as smart as a lab mouse and then I'll take your concerns seriously. The analogy to anthropomorphic global climate change is a non sequitur. Climatologists have created falsifiable theories which make testable predictions. And you really have no clue about my personal religious beliefs. Calm down and take a seat.

I ask for proof that AGI is possible. Show me a computer as smart as a lab mouse and then I'll take your concerns seriously.

I would argue that, unless you can show why AGI is not - in principle - possible, that the null hypothesis would be that it is possible. Unless we veer off into some weird mysticism, it seems that the human brain turns energy and matter into intelligence somehow, operating according to the physical laws of the universe... why shouldn't it be possible to build something else that does the same?

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#137

Post by zaksoup »

virgilp wrote: post #133
cr0sh wrote: post #122

Earlier quoted context omitted.

Interestingly, Hinton is on record as essentially saying that there's a good possibility that what's currently being done is wrong - and that we need to rethink our approach. Mainly in the idea/concept of back-propagation. It's something that I've thought about myself. For the longest time, I could never understand how it worked, then I went thru Ng's "ML Class" (in 2011, which was based around Octave), and one part…

You're contradicting yourself with your examples. If we didn't manage to fly by imitating birds - why do you care that AI doesn't work the brain does? That should be a _good_ sign, if we trust the analogy - right?

I think the best interpretation of their point is that at some point the breakthrough was questioning a fundamental assumption. I think the point about matching real neurons was just to give credence to their hunch that backprop is not quite the right track to be taking.

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#138

Post by nabla9 »

epicureanideal wrote: post #113

https://en.wikiquote.org/wiki/Incorrect_predictions "Hence, if it requires, say, a thousand years to fit for easy flight a bird which started with rudimentary wings, or ten thousand for one which started with no wings at all and had to sprout them ab initio, it might be assumed that the flying machine which will really fly might be evolved by the combined and continuous efforts of mathematicians and mechanicians in f…

The arguments like above are "platitude level arguments".

We really don't learn anything from the problem in had by talking in generic terms. We use these arguments when we want to justify our hopes and feeling, but there is really nothing to learn from it.

Hinton, Hassabis, Bengio and others point out that we can't 'brute force' AI development. There needs to be actual breakthroughs in the field and there may be several decades between them.

AI, brain science and cognitive science are extremely difficult fields with small advances, yet people assume that it's possible to 'brute force' AGI by just adding more computing power and doing more of the same.

Macroeconomics is probably less complex research subject than AI or brain science, but nobody assumes that you can just brute force truly great macroeconomic model in few years if you just spend little more resources.

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#139

Post by TheOtherHobbes »

meroes wrote: post #89
sdenton4 wrote: post #36

Earlier quoted context omitted.

Maaaaybe. I tend to think that symbolic reasoning is a learning tool, rather than a goalpost for general intelligence. For example, we use symbolic reasoning quite extensively when learning to read a new language, but once fluent can rely on something closer to raw processing - no more reading and sounding out character sequences. Similarly with chess - eventually we have good mnemonics for what make good plays, and…

As a layman, is this just saying we learn by training an intuition of what's what/what's correct, rather than actually calculating deep reality/referencing our entire memory set every time we intake some information or need to solve a task problem? Meaning, we develop tons of rules/heuristics after repeated pattern exposures, and use the simplified rules rather than a deep theory or 'brute-forcing' every possibility…

AI is a confused soup of more or less (un)related concepts: agency, sentience, pattern recognition, unsupervised learning, embodiment, NLP, and goal selection - among others.

IMO the minimal useful definition of AGI would list a set of testable skills that would qualify as AGI, and a more useful definition would be based on quantifiable skill sets that would allow numerical comparisons between humans and AIs.

It seems pointless to speculate when AGI might be a reality when we have only the fuzziest idea what AGI is supposed to look like.

Re: Geoffrey Hinton and Demis Hassabis: AGI is nowhere close to being a reality

#140

Post by joycian »

visarga wrote: post #124
joe_the_user wrote: post #116

Earlier quoted context omitted.

Intelligence is not limited to what Humans are good at. Being able to implement all the things human are good at, however, should be able to get us everything that we could do, because anything we could create, it could create too. AGI that is as smart as say a rat would easily qualify as AGI even without language skills. Indeed, but while a full language-using AI is ways a way at least, using language is one thing t…

A rat can do something else that a neural net can't - it is a self replicator. Our neural nets don't have self replication or a huge, complex environment and timescale to evolve in. Self replication creates an internal goal for agents: survival. This drives learning. Instead, we just train agents with human-made rewards signals. Even a simple environment, like the Go board, when used for training many generations of…

Survival is instrumental to any goal. Not only self replication would create that drive.
Post reply on HN