Just A New Fractal Detail In The Big Picture (2015) • Hacker News Board

Re: Just A New Fractal Detail In The Big Picture (2015)

#141

Post by mdpopescu » Mon, Jul 04, 2016, 9:44 AM UTC

kirrent wrote: post #99

Joof wrote: post #78

Earlier quoted context omitted.

Many larger systems are more than the sum of their parts. Enough that you could consider the grouping it's own entity. Ant hills are the canonical example because individual ants are recognizable as 'animals', but the function of an anthill clearly relies on a complex interaction between more simple ants. Just as we are made of tiny organisms, bigger systems could be considered to be made of us.

Hofstadter wrote a nice piece on how an ant colony could be thought to be an entity in itself. My favourite quote pasted from an online source: Anteater: "I reject holism. I challenge you to tell me, for instance, how a holistic description of an ant colony sheds any more light on it than is shed by a description of the ants inside it, and their roles, and their, interrelationships. Any holistic explanation of an ant…

Unfortunately, this argument proves too much. Replace "and colony" with "human" and "ant" with "cell in a human body" - can you explain where the consciousness experienced by a human cell colony arises from?

(To avoid debating materialism vs idealism use another creature, say "chimp", instead of "human" - assuming that you agree that chimps have a consciousness.)

I am not fond of the holistic / emergent idea but I don't think "we can't figure out the exact place where the sum becomes larger than its parts" is a good argument against it.

Re: Just A New Fractal Detail In The Big Picture (2015)

#142

Post by majewsky » Mon, Jul 04, 2016, 9:44 AM UTC

scrollaway wrote: post #74

ccccc0 wrote: post #64

Earlier quoted context omitted.

I'd argue that Global Civilization is nothing but living!

I read a comment on HN a while back arguing that today's companies are essentially human-powered AIs. It made me think quite a lot -- large companies especially take a lifeform of their own. They tend to mostly care for themselves. Actions are taken "for the company" without any one person being actively aware of why. Higher-ups have some form of backdoor access to the direction the company takes; a symbiotic relatio…

There was a talk at the 30C3 where the speaker argued that we do not need to fear AIs. A human-level AI is an information processing system (what we have already, and which is not harmful) plus a motivation system. The latter is the one that's to be worried about, but it already exists in the form of corporations (which fittingly also work against our collective interest in many ways, e.g. environmental protection).

Re: Just A New Fractal Detail In The Big Picture (2015)

#143

Post by vidarh » Mon, Jul 04, 2016, 9:59 AM UTC

unusximmortalis wrote: post #126

daveguy wrote: post #122

Earlier quoted context omitted.

Be careful assuming you understand how something works without having built at least a prototype.

Interesting thought. Do not say you know it/understand it untill you experience it. I thing there a thin line between these two categories, those things we ca understand without experience, and the rest. Is there a research paper about this aspect?

The main issue is simply that most descriptions of a subject are nowhere near detailed enough to be "complete", and even if they are, the reader of that description almost never will manage to pick up every nuance of what is described.

Try reproducing results from some research papers, and you see this demonstrated very easily, both in frustration of having to figure out what you misunderstood, and in frustration of trying to figure out details that were left out.

Re: Just A New Fractal Detail In The Big Picture (2015)

#144

Post by majewsky » Mon, Jul 04, 2016, 10:01 AM UTC

schoen wrote: post #134

ansible wrote: post #70

Earlier quoted context omitted.

Yeah, I thought that was all kind of funny too. I'm a voracious reader, and have been curious about just about all parts of science and technology at some point. So I too knew at least the basics for everything he mentioned. Magically plop me in a stone-age agrarian society, and we'd be back to at least an 18th-century level of technology in a couple decades.

There were some science fiction stories (I forgot their names and authors) in which time travelers actually encounter a lot of practical difficulties in past societies, because the people they encounter don't believe them or don't readily see the benefits of their suggestions. So maybe you need to add "a stone-age agrarian society that's eager to take advantage of my knowledge", since that part can't necessarily be t…

This. I just saw a documentary about famous inventions yesterday. They talked about how when Alexander Graham Bell conceptualized the telephone, most investors said "Why should we ever need this? We can already transmit speech with telegraphs."

I guess the same effect amplifies the more absurd or outlandish the proposed technology appears.

Re: Just A New Fractal Detail In The Big Picture (2015)

#145

Post by makomk » Mon, Jul 04, 2016, 10:04 AM UTC

scotty79 wrote: post #135

Houshalter wrote: post #132

Earlier quoted context omitted.

I don't think that is true, just from observing the difference between humans. Really intelligent mathematicians can do work that a weaker mathematician would never accomplish if they spent their whole life on it. And the average person could probably not even understand it given years. There seem to be vast differences in ability between humans, despite all of us having 99.99% the same DNA and same brain structure.…

I agree that AI can do marvels in math, but math quickly departs from reality (like string theory) and being able to prove things about some beautiful intricate mathematical structure most of the times doesn't help you much with the world of matter. I recommend Culture novels written by Iain M. Banks. There are vastly powerful AIs but they spend as much time as they can afford in "Infinite Fun Space" which is basical…

The stock market isn't just a purely random game though - it's based on the expected earnings of the companies, so if you can come up with a better estimate than everyone else by combining information better, you can make money. The reason it's hard for humans to make money this way is because everyone else is doing the same thing, so any advantage is short-lived.

Re: Just A New Fractal Detail In The Big Picture (2015)

#146

Post by majewsky » Mon, Jul 04, 2016, 10:06 AM UTC

unusximmortalis wrote: post #128

jonknee wrote: post #83

Earlier quoted context omitted.

Specializing enabled huge advances, everyone having to do everything is actually quite fragile because it can only support a small population.

I think it can be drawn a parallel with the centralized vs decentralised network. Which one is better? And I think the stronger one for the species is a combination of the two.

There are different types of decentralized networks. Not every such network has totally equal nodes, they may still serve different roles. The network is still decentralized as long as sufficiently many nodes are servicing each role.

Re: Just A New Fractal Detail In The Big Picture (2015)

#147

Post by philh » Mon, Jul 04, 2016, 10:08 AM UTC

scotty79 wrote: post #135

Houshalter wrote: post #132

Earlier quoted context omitted.

I don't think that is true, just from observing the difference between humans. Really intelligent mathematicians can do work that a weaker mathematician would never accomplish if they spent their whole life on it. And the average person could probably not even understand it given years. There seem to be vast differences in ability between humans, despite all of us having 99.99% the same DNA and same brain structure.…

I agree that AI can do marvels in math, but math quickly departs from reality (like string theory) and being able to prove things about some beautiful intricate mathematical structure most of the times doesn't help you much with the world of matter. I recommend Culture novels written by Iain M. Banks. There are vastly powerful AIs but they spend as much time as they can afford in "Infinite Fun Space" which is basical…

> being able to prove things about some beautiful intricate mathematical structure most of the times doesn't help you much with the world of matter.

How about discovering a flaw in human crypto systems?

> Super-intelligence won't help you with stock market because it's purely random game.

Humans can't reliably exploit the stock market, because we're competing against other humans. It's not that there are no patterns, it's that as soon as someone discovers a pattern, people rush to exploit it and then it's tapped out. Imagine you were the first human stock trader to think of "buy stocks in companies when their product announcements get featured in newspapers"; and imagine that however much you tried to explain your strategy to the other stock traders, they just didn't get it. You wouldn't always pick a winner, but you'd make a lot of money.

How confident are you that AI won't be able to find something like that?

Re: Just A New Fractal Detail In The Big Picture (2015)

#148

Post by dredmorbius » Mon, Jul 04, 2016, 10:09 AM UTC

ZenoArrow wrote: post #130

dredmorbius wrote: post #91

Earlier quoted context omitted.

Pointing to expertise is not argument from authority.

It depends. In this case, it was the framing that made it the argument from authority, namely "in what people who actually know whereof they speak". The implication is that the sources of truth are limited to a select few people.

Argument from authority is "X is true because Y says it is".

That's not what I claimed.

This is tedious.

Re: Just A New Fractal Detail In The Big Picture (2015)

#149

Post by ffwd » Mon, Jul 04, 2016, 10:20 AM UTC

Moshe_Silnorin wrote: post #25

pygy_ wrote: post #19

Earlier quoted context omitted.

For your first scenario economic growth must be fueled by a proportional growth in energy consumption. Fossil and Nuclear fuels exist in finite quantities, and the rate of sunlight that hits earth is more or less constant. Most people would still starve as fields would be repurposed to produce fuel.

Maybe it would only last a few years. But once we can convert capital directly into labour in a manner that scales, we will get insane amounts of economic growth.

I understand what you're saying but I think it's bad to use the term 'economic' growth. There would be potential productivity growth but not necessarily economic growth, since the economy is a human endeavour. AIs don't need money and won't contribute to money, and the machines would only be 'allowed' to produce what the customers would be able to buy (so essentially the current economy), unless they would produce lots of extra stuff just for the hell of it.

To have all that growth I think you'd need to decouple the machines from the economy but then distributing natural resources would be a problem, and then you come back to central planning and that's a whole other can of worms.

Re: Just A New Fractal Detail In The Big Picture (2015)

#150

Post by dharma1 » Mon, Jul 04, 2016, 10:40 AM UTC

tambourine_man wrote: post #80

JumpCrisscross wrote: post #47

> I read once that human brains began shrinking about 10 thousand years ago and are now as much as 15% smaller than they were then. Timeline's a bit off, but the science is still a shocker: > Over the past 20,000 years, the average volume of the human male brain has decreased from 1,500 cubic centimeters to 1,350 cc, losing a chunk the size of a tennis ball. The female brain has shrunk by about the same proportion. h…

I find that scary. A huge collection of specialists makes for a pretty fragile species.

Only if there isn't enough redundancy.

Though I agree, some of the systems (food/energy supply, communications) we have built are potentially quite fragile and disruptions could cause a lot of mess short term.

Never hurts to brush up on your survival skills Bear Grylls style :)