4 ms·

Computers are millions of times faster than 30 years ago, let alone 60 when the field started. There are that many decades of new knowledge and research. There

Computers are millions of times faster than 30 years ago, let alone 60 when the field started. There are that many decades of new knowledge and research. There is a dozen times more funding than in the past. AI research is actually producing things of economic value now. And the field has moved past brute force search and dumb expert systems, to deep general learning algorithms. Of course this time it's different. The only question is how different? I would say that HN is cautiously optimistic about the future of AI, in aggregate. With extreme opinions on both sides, of course. But there are many skeptical articles that reach the front page, and there are many skeptical comments like yours at the top of every thread about it. The opinion also seems to be "in my lifetime" or "in the next few decades", not in the immediate short term future. That's a pretty long time, and a lot could happen in that time. Saying that AI will definitely not happen in that time, is just as overconfident as saying that it definitely will. Lastly, saying "people in the past who said 'this time it's different' were wrong once, therefore anyone who says that is wrong", is a fully general counterargument. You are applying it to vastly different domains (short term economic bubbles vs long term technology predictions). And you are glossing over all the details of the arguments.

argonaut 10y ago
> "in my lifetime" or "in the next few decades" is actually quite optimistic. You've put a definite timeframe on something extremely nebulous. I actually think the HN group-think plurality is quite more optimistic than that (next 10 years), however.
- Houshalter 10y ago
  It's not that optimistic. Even if you assume a totally uninformative prior, it's still reasonably likely to happen in our lifetime. E.g. the copernicus method/prior gives an estimate of 50% probability within the next 65 years. And an even more optimistic prediction if you add other factors: http://lesswrong.com/lw/mxy/using_the_copernican_mediocrity_principle_to/ http://lesswrong.com/lw/mxy/using_the_copernican_mediocrity_... You must have some strong additional information to make such a pessimistic prediction. I really don't know where you get the impression that the majority of HNers expect strong AI in 10 years. That's just ridiculously. I would guess maybe less than 5% think that. Yes there are a lot of optimistic articles on AI, but they are almost all about weak AI.
  - argonaut 10y ago
    The copernican prior is totally speculative.
eli_gottlieb 10y ago
>And the field has moved past brute force search and dumb expert systems, to deep general learning algorithms. I would still describe deep neural networks as brute-force statistically-weighted function search.
- Houshalter 10y ago
  For sufficiently vague and useless definitions of "brute force", literally everything is "brute force". Backprop was such a big advancement because it allowed training nets much faster than, say, hillclimbing. Which itself is much faster than actual brute force.

argonaut 10y ago

> "in my lifetime" or "in the next few decades" is actually quite optimistic. You've put a definite timeframe on something extremely nebulous. I actually think the HN group-think plurality is quite more optimistic than that (next 10 years), however.

Houshalter 10y ago
It's not that optimistic. Even if you assume a totally uninformative prior, it's still reasonably likely to happen in our lifetime. E.g. the copernicus method/prior gives an estimate of 50% probability within the next 65 years. And an even more optimistic prediction if you add other factors: http://lesswrong.com/lw/mxy/using_the_copernican_mediocrity_principle_to/ http://lesswrong.com/lw/mxy/using_the_copernican_mediocrity_... You must have some strong additional information to make such a pessimistic prediction. I really don't know where you get the impression that the majority of HNers expect strong AI in 10 years. That's just ridiculously. I would guess maybe less than 5% think that. Yes there are a lot of optimistic articles on AI, but they are almost all about weak AI.
- argonaut 10y ago
  The copernican prior is totally speculative.

eli_gottlieb 10y ago

>And the field has moved past brute force search and dumb expert systems, to deep general learning algorithms. I would still describe deep neural networks as brute-force statistically-weighted function search.

Houshalter 10y ago
For sufficiently vague and useless definitions of "brute force", literally everything is "brute force". Backprop was such a big advancement because it allowed training nets much faster than, say, hillclimbing. Which itself is much faster than actual brute force.