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

Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

wired.com

Post reply on HN
Print view View original

231–239 of 239 posts

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#231

Post by ewjordan »

zardo wrote: post #190
ewjordan wrote: post #101

Earlier quoted context omitted.

I'm as bullish on AGI as anyone in the medium term, but deep learning is not even playing the same game as AGI, let alone in the same ballpark or having the potential to achieve it. Deep learning is still mere perception. It doesn't handle memory or processing, it just transforms input into output, typically trained by Big Data, way bigger than necessary statistically speaking, given the world we live in. AGI require…

>Deep learning is still mere perception. It doesn't handle memory or processing, There's more going on than convolutional neural nets. Architectures with memory and attention mechanisms do exist. As examples: https://arxiv.org/abs/1506.03134 https://deepmind.com/blog/differentiable-neural-computers/

Sure, and I probably shouldn't have glossed over that. That sort of research is definitely progress, though it's not paradigm shifting in any way. I do think that we are getting past perception slowly but surely, I just don't think we're there yet.

What really doesn't exist is any meaningful stab at unsupervised (or self-supervised) training on completely unstructured inputs or any sort of knowledge condensation/compression, at least for time dependent problems. These are of paramount importance to the way we think, and to what we can do.

There's a lot of trivially low hanging fruit, too - I still have yet to see even a grad school thesis that starts with an N+M node recurrent network and trains an N node subnetwork to match the outputs based on fuzzed ins, and then backs that out into an unsupervised learning rule that's applicable to multiple problems. Or better, a layered network that is recurrent but striated, that tries to push weights towards the lower layers while reproducing the same outputs (hell, even with a FF network this would be an interesting problem to solve if it was unsupervised). These are straightforward problems that would open up new avenues of research if good methods were found, but are mostly unexplored right now.

I could be wrong, if I had real confidence that we were close I'd be working on this stuff, but I'm collecting a paycheck doing web dev instead...

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#232

Post by zaroth »

Houshalter wrote: post #219
zaroth wrote: post #164

Earlier quoted context omitted.

I think this is factually incorrect. Women used to spend the majority of their free time spinning thread and sewing clothes for their family . The average person owned three outfits total, and wore basically one outfit every single day. The industrial revolution dramatically increased both production and consumption. We now own different outfits for every day, dozens of special occasion fashions which must be regular…

Everyone knows that in the long run the industrial revolution was a positive. But at the time, the disruption caused lots of misery. Farmers were driven from farms that had relatively good living conditions, to cities with awful working conditions, living conditions, and lower wages. Eventually things improved, but it didn't happen overnight. What does it matter if there are more clothes, if there aren't any consumer…

Do you really believe this? Horses can become obsolete because they are a domesticated animal, they are a tool. Humans are really something quite different.

It is truly an insult to humanity to think; all these people doing menial jobs which could be automated are now obsolete. That they are somehow incapable of higher thoughts and reasoning and cannot add value beyond the machines.

In fact, while certainly there is a range of inherent potential between humans, my understanding is that the nominal human capacity for creative thought is orders of magnitude beyond the point of obsolescence by any kind of "artificial intelligence" we expect to be able to create at least within the next century.

No, certainly we have not yet created anything even remotely like the machine that will be our master.

The average human, as we have asked each generation of "average" human before us, will use technology to reach farther than you can imagine they would ever be able to reach.

And also, it's worth considering, how very much we tend to under-estimate the intelligence of historical man from our lofty perch of technological superiority, just as we under-estimate our future potential.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#233

Post by Houshalter »

zaroth wrote: post #232
Houshalter wrote: post #219

Earlier quoted context omitted.

Everyone knows that in the long run the industrial revolution was a positive. But at the time, the disruption caused lots of misery. Farmers were driven from farms that had relatively good living conditions, to cities with awful working conditions, living conditions, and lower wages. Eventually things improved, but it didn't happen overnight. What does it matter if there are more clothes, if there aren't any consumer…

Do you really believe this? Horses can become obsolete because they are a domesticated animal, they are a tool. Humans are really something quite different. It is truly an insult to humanity to think; all these people doing menial jobs which could be automated are now obsolete. That they are somehow incapable of higher thoughts and reasoning and cannot add value beyond the machines. In fact, while certainly there is…

My favorite quote from I, Robot is a man questioning the intelligence of a robot. "Can a robot write a symphony? Can a robot turn a canvas into a beautiful masterpiece?" The robot replies, "Can you?"

Most humans aren't that special. Long before alphaGo beat one of the best Go players in the entire world, simple Go programs could destroy the majority of players. Sure, AI probably won't be able to do computer programming for a long time. But the average person with an IQ of 100 is not going to retrain to be a computer programmer. AI doesn't need to be as intelligent as the best humans, it just needs to be as intelligent as the average person. Probably much less than that, because the average job is boring repetitive work that doesn't necessarily require much intelligence.

I mean seriously, where do you predict all the unemployed people will go to? What jobs do you think are invulnerable to automation, and can absorb 90% of the population? What jobs have such great economic value, require lots of unskilled workers, and can't be replaced by machines?

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#234

Post by Houshalter »

ewjordan wrote: post #231
zardo wrote: post #190

Earlier quoted context omitted.

>Deep learning is still mere perception. It doesn't handle memory or processing, There's more going on than convolutional neural nets. Architectures with memory and attention mechanisms do exist. As examples: https://arxiv.org/abs/1506.03134 https://deepmind.com/blog/differentiable-neural-computers/

Sure, and I probably shouldn't have glossed over that. That sort of research is definitely progress, though it's not paradigm shifting in any way. I do think that we are getting past perception slowly but surely, I just don't think we're there yet. What really doesn't exist is any meaningful stab at unsupervised (or self-supervised) training on completely unstructured inputs or any sort of knowledge condensation/comp…

Sequence predicting RNNs are basically unsupervised, in that they can learn from lots raw of unlabelled data. And they learn useful internal representations which can be adapted for other tasks. There is lots of old work on unsupervised learning rules for RNNs, including recurrent autoencoders and history compression.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#235

Post by erikpukinskis »

icholy wrote: post #161
drcross wrote: post #153

Earlier quoted context omitted.

He left office 8 years ago, things have changed in that time. I also refer to the internet as the internets sometimes too. Why would the most powerful person in the world need to use email anyway?

"I was elected to lead, not to read"

Bush is probably at the very top of the 99th percentile in terms of readers. He reads a book+ a week.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#236

Post by erikpukinskis »

moeabbas999 wrote: post #44
bitL wrote: post #41

Earlier quoted context omitted.

Specialized AI - think about retina neurons, capable of detecting direction, edges, depth etc. General AI - think about thinking machines We can do the former, have no clue how to do the latter.

It's not such a huge jump to the latter when you nail the former.

This is what's called a bald assertion.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#237

Post by erikpukinskis »

asimuvPR wrote: post #60
msvan wrote: post #16

What does Hacker News think about AI? Is it real this time, or are we in for another winter? I'm seeing a lot of grand claims, and it certainly seems like there are plenty of applications, but I'm still not totally convinced that it will turn the entire economy upside down. Given the enormous amount of press, tweets, blog posts, conferences, degree programs, seminars and interviews popping up it seems like there has…

There is a real risk related to goal oriented AI. It does not need to feel or dream. Merely having survival as its goal is sufficient to make it dangerous to other life forms. Worse is that it can happen at any time (it may have happened already). Given the computing power, tools, and availability of knowledge we can assume that it can be done outside of a controlled lab environment by a non-scientist.

I'm not too worried about it, because AIs will need to feel and dream and care and suffer and empathize to begin to be as intelligent as a human.

Imagine a human who lacks empathy entirely. That's a disability. They may be able to do some amount of destruction, but at some scale they simply lack the social intelligence necessary to compete with the entire species.

This is the most common mistake I think people make when reasoning about AI: they think human limitations are weaknesses. But they're not weaknesses they're tradeoffs. Natural selection has had a chance to reward all kinds of variations, including more cortex, and less empathy. But we ended up where we are because of tradeoffs.

Any AI which is intelligent in the same way humans are will also have our limitations. Any AI which doesn't have our limitations won't be as smart as us in those respects.

You have to really ask yourself what the difference is between a human with an AI simulator and an AI with a human simulator. In the limit of simulator quality there is none.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#238

Post by erikpukinskis »

agentgt wrote: post #40
msvan wrote: post #16

What does Hacker News think about AI? Is it real this time, or are we in for another winter? I'm seeing a lot of grand claims, and it certainly seems like there are plenty of applications, but I'm still not totally convinced that it will turn the entire economy upside down. Given the enormous amount of press, tweets, blog posts, conferences, degree programs, seminars and interviews popping up it seems like there has…

I have often wandered if there is some sort universe limitation to general AI singularities particularly if the current universe is actually a simulation. An intellectual singularity or extremely rampant aggressive intelligence might be an errant state that gets dealt with (aka universe circuit breaker / watchdog). This might also explain the Fermi Paradox. The other idea I have doted on is that perhaps universes are…

I'm curious: how interested are you in understanding the Earth ecology? Because we understand very little about it, and we are very rapidly destroying the information it contains.

I ask because I wonder, if we did meet or make a creature which could learn everything, if it wouldn't say some variation on "99.9999% of everything there is to know is right there in your ecology and visible to the naked eye. Go look."

I also wonder if one of the first things the sentient AIs teach us is that yes, we are committing an egregious ongoing information Holocaust through habitat destruction.

Re: Barack Obama on A.I., Autonomous Cars, and the Future of Humanity

#239

Post by andirk »

asimuvPR wrote: post #193
mr_spothawk wrote: post #188

Earlier quoted context omitted.

"mediocre minds"?

It sounds weird, but I believe its meant as "untrained minds".

Yes untrained but more broad to mean people who don't understand the magic, people who are negligent, etc. I kind of got it from Einstein's quote, "Great spirits have always encountered opposition from mediocre minds."
Post reply on HN