The Singularity in Our Past • Hacker News Board

Re: The Singularity in Our Past

#11

Post by VladRussian » Wed, Dec 08, 2010, 1:53 AM UTC

iwr wrote: post #3

The idea is that the Singularity would be a point in time where the totality of human expert knowledge would be insufficient to comprehend what an AI (or an AI-designed AI) would be capable of comprehending. If we think human-level AI is possible, then superhuman AI would also be possible. Note, though, that while not infinite, human cognition is complete, that is, a human or group of humans is only limited by time o…

>a dog will never stumble upon quantum mechanics

i'm always amazed by such blatant anthropo-glorifing-centric statements which imply as obvious facts the doubtful things like this:

- that dog's brain has much less processing power. The sheer size difference is many times adjusted by efficiency difference (where human's one is less efficient)

- that quantum mechanics is a super-achievement, replication of which is a mark of higher intelligence. To me the ugliness of QM is bordering on a shame for human intelligence.

Re: The Singularity in Our Past

#12

Post by jerf » Wed, Dec 08, 2010, 2:00 AM UTC

Reading over the comments up to this point, this is a definition debate, not an intelligent conversation. By some useful readings of the word "singularity", they are indeed relative and there are several in our past, and quite likely several in our future. Looking back at the prognostications in 1980 one can fruitfully argue that we've been through one just since then. By other useful readings of the word "singularity", it lies solely in the future if it exists or can exist at all.

Arguing which definition is "true" is wrong; definitions can't really true or false, they are simply useful or not useful, nor do they have any direct impact on reality.

Re: The Singularity in Our Past

#13

Post by s_jambo » Wed, Dec 08, 2010, 2:31 AM UTC

In order to move this away from debate about semantics I think this meshes nicely with an alternative view of intelligence. Specifically Human intelligence as a social endeavour rather than an individual struggle.

If you look at intelligence in that light we have gone through a huge transformation with the printing press, industrial revolution and now internet revolution.

Basically if you think of it as each human brain just being a small component able to try out new ideas then spread those ideas... Well that massive human spanning brain is undergoing massive ongoing upgrades in memory capacity and speed of communication.

Re: The Singularity in Our Past

#14

Post by DaniFong » Wed, Dec 08, 2010, 2:47 AM UTC

cicada wrote: post #4

Confusions like this are the reason why I favor the intelligence explosion flavour of the singularity ( http://yudkowsky.net/singularity/schools ) -- there's no confusion as to whether we've already had one. The industrial revolution, while revolutionary, did not improve on the intelligence that invented it. The single most important part about the singularity is that the technology does improve on the intelligence t…

Perhaps the scientific revolution was not an innovation of a single person, but of the collective entity formed by the proto-scientists of civilization. That conceptual advance, along with communication, tool, and other conceptual advances since that time, have continued to accelerate civilization's collective intelligence.

We don't expect the singularity to improve the intelligence of transistors. Perhaps it is unreasonable to expect that previous intelligence explosions of a collective being massively improved the intelligence of the humans comprising it.

Re: The Singularity in Our Past

#15

Post by gwern » Wed, Dec 08, 2010, 3:13 AM UTC

T_S_ wrote: post #8

People seem to underestimate how much engineering will be required for intelligence to interact with the environment successfully. My dog might not be as "intelligent" as me, but he knows a lot more than I do when it comes to information gained through hearing, taste and smell. He has better engineering in many respects. Almost any life form beats me in "intelligence" if you define the test right. There are already t…

> OK now what sequences would you like to present to the algorithm for training? Ones from the environment? Great, but those will take time to collect. Simulated ones? It takes a lot of knowledge to generate good ones.

With AIXI and variants, since they are universal predictors, it doesn't really matter what sequences you supply so long as they have some connection. Heck, just your algorithm access to an email server or something.

Re: The Singularity in Our Past

#16

Post by derefr » Wed, Dec 08, 2010, 3:22 AM UTC

VladRussian wrote: post #11

iwr wrote: post #3

The idea is that the Singularity would be a point in time where the totality of human expert knowledge would be insufficient to comprehend what an AI (or an AI-designed AI) would be capable of comprehending. If we think human-level AI is possible, then superhuman AI would also be possible. Note, though, that while not infinite, human cognition is complete, that is, a human or group of humans is only limited by time o…

>a dog will never stumble upon quantum mechanics i'm always amazed by such blatant anthropo-glorifing-centric statements which imply as obvious facts the doubtful things like this: - that dog's brain has much less processing power. The sheer size difference is many times adjusted by efficiency difference (where human's one is less efficient) - that quantum mechanics is a super-achievement, replication of which is a m…

Why is quantum mechanics ugly?

Re: The Singularity in Our Past

#17

Post by mechanical_fish » Wed, Dec 08, 2010, 3:28 AM UTC

cicada wrote: post #4

Confusions like this are the reason why I favor the intelligence explosion flavour of the singularity ( http://yudkowsky.net/singularity/schools ) -- there's no confusion as to whether we've already had one. The industrial revolution, while revolutionary, did not improve on the intelligence that invented it. The single most important part about the singularity is that the technology does improve on the intelligence t…

The industrial revolution, while revolutionary, did not improve on the intelligence that invented it.

...typed the person into his quantum-mechanically-mediated instantaneous conversation with ten thousand of the best-educated technologists on Earth.

You're going to have trouble convincing me that Google doesn't make me more intelligent. You're also going to have trouble convincing me that, say, cheap printing and cheap food -- both products of the industrial revolution, and without which I might be a subsistence farmer -- do not make me more intelligent. I wouldn't know quantum mechanics without them, after all, so I wouldn't understand how computers work.

And if you strive to define "intelligence" such that human intelligence has not been improved by technology -- which can be done; I'm clearly less intelligent than my Cro-Magnon ancestors by many criteria -- you are going to have a hard time convincing me that the future will be any different. If technology hasn't worked to improve "intelligence" before, why should it do so later?

Re: The Singularity in Our Past

#18

Post by cicada » Wed, Dec 08, 2010, 3:40 AM UTC

mechanical_fish wrote: post #17

cicada wrote: post #4

Confusions like this are the reason why I favor the intelligence explosion flavour of the singularity ( http://yudkowsky.net/singularity/schools ) -- there's no confusion as to whether we've already had one. The industrial revolution, while revolutionary, did not improve on the intelligence that invented it. The single most important part about the singularity is that the technology does improve on the intelligence t…

The industrial revolution, while revolutionary, did not improve on the intelligence that invented it. ...typed the person into his quantum-mechanically-mediated instantaneous conversation with ten thousand of the best-educated technologists on Earth. You're going to have trouble convincing me that Google doesn't make me more intelligent. You're also going to have trouble convincing me that, say, cheap printing and ch…

A Google engineer should not expect to get a helpful response when he asks the Google search engine for advice on building a better Google.

Re: The Singularity in Our Past

#19

Post by T_S_ » Wed, Dec 08, 2010, 3:50 AM UTC

gwern wrote: post #15

T_S_ wrote: post #8

People seem to underestimate how much engineering will be required for intelligence to interact with the environment successfully. My dog might not be as "intelligent" as me, but he knows a lot more than I do when it comes to information gained through hearing, taste and smell. He has better engineering in many respects. Almost any life form beats me in "intelligence" if you define the test right. There are already t…

> OK now what sequences would you like to present to the algorithm for training? Ones from the environment? Great, but those will take time to collect. Simulated ones? It takes a lot of knowledge to generate good ones. With AIXI and variants, since they are universal predictors, it doesn't really matter what sequences you supply so long as they have some connection. Heck, just your algorithm access to an email server…

| Heck, just your algorithm access to an email server or something.

Then my dog will find the bone faster than AIXI. See what I mean?

Re: The Singularity in Our Past

#20

Post by Helianthus16 » Wed, Dec 08, 2010, 3:54 AM UTC

VladRussian wrote: post #11

iwr wrote: post #3

The idea is that the Singularity would be a point in time where the totality of human expert knowledge would be insufficient to comprehend what an AI (or an AI-designed AI) would be capable of comprehending. If we think human-level AI is possible, then superhuman AI would also be possible. Note, though, that while not infinite, human cognition is complete, that is, a human or group of humans is only limited by time o…

>a dog will never stumble upon quantum mechanics i'm always amazed by such blatant anthropo-glorifing-centric statements which imply as obvious facts the doubtful things like this: - that dog's brain has much less processing power. The sheer size difference is many times adjusted by efficiency difference (where human's one is less efficient) - that quantum mechanics is a super-achievement, replication of which is a m…

I'm always amazed by such blatant anthropo-deprecating-centric (?) statements.

-The difference in efficiency should tell you that that power spent is _worth_ something, like astonishingly superior abstraction capabilities and the ability to speak and write.

-The ugliness (read; complexity) of QM could easily be interpreted as a tribute. Whether you like it or not, it is the best we have. To paraphrase HHG2G, if it's ugly, then it's the universe that got it wrong, not us. (I'm partially joking, mostly because I don't suspect you really have a firm enough grasp of QM, let alone QFT, to determine its aesthetic beauty.)

I mean, I'm all for a little humbling perspective. We are a race of people descended from stupider people unto apehood (and before that, lizards!) that are stuck on a rock in a universe fantastically larger than us.

That doesn't mean we should be debbie downers.