About the "immortality". I have to clarify what i mean with that in this context. As much as I understand, you can take identical backup copies of the contents of a computer, fix everything with spare parts and so on. You have a "relative immortality", if you want to. A human can´t be copied and fixed in that way.

To prevent overcrowding of the earth (/eventually universe?), when being without the death factor, computers-robots probably will have to be updated, but can then have all the old stuff left, but that maybe could be defined as a "partial death" of the original computer-robot, because the old stuff will be pushed into the background.

According to the current wiew of the universe, everything in the universe will eventually be ripped apart, even all the atoms. Long before that all the stars have burnt out. If it´s not possible to find/create new universes (or altering the laws of nature??), mortality comes at least then.

"We" is the people that work on and study artificial intelligence (I am doing a degree in artificial intelligence).

You can apply survivorship bias to anything. What you haven't done is show why it could be particularly strong in Kurzweil's case.

The fact that this is precisely the sort of place that we'd expect to find survivorship bias means that that IS excellent evidence that that is what's going on here.

Can you give any examples of someone else that has given 4 "generations" of technological predictions, with the first 3 being extremely accurate, and the 4th being extremely far off?


Survivorship bias doesn't seem to apply here, because RK's predictions for the past 25 years have so far been extremely accurate. Is there anyone that has come close to that ever?

This thread is muddled w/ people talking in generalities. There's really not many reasonable viewpoints here.



1) You think Kurzweil is right.

2) You think he's optimistic about his predictions, but off on time.

3) You think either computer consciousness is impossible, or extremely far off.


Is there a 4th option?

Does anyone here want to claim the viewpoint that creating extremely efficient computer consciousness wouldn't result in a snowball effect that will create insanely ridiculous technology?

5) Kurzweil is probably too optimistic when it comes to fixing/enhancing humans. I think it´s easier to create everything "from scratch", that is computers-robots. Because of that the second one of these two options will prevail, as a long time (many decades / some centuries) goes by.

Maybe, though, he may have some ideas in that direction, but realizes that in order to get his theories popular, he may have to give the human ego some. Or then it´s his own ego, maybe hoping to get on the "immortality train". Or that his possible children will.

Maybe I really should read his book, before commenting more on how particularly he thinks. This was just a gut feeling about it.

Or
4. Efficient computer conciseness will create a snowball effect that creates insanely ridiculous technology that will fall well short of what Kurzweil is envisioning.

Kurzweil himself says that predictions for 2099ish or beyond are likely to be very inaccurate. He's always intentionally vague when talking about anything beyond 2050.

Basically he doesn't make many predictions for what happens 10 years after the singularity.

He has thrown at a few hypotheticals, and those do seem super crazy, but I think it would be unfair to lump those with the rest of his predictions.

Someone who understands statistics wouldn't be so quick to make such a claim.

Take 10000 people (there are more futurists than that) and have them make 10 predictions/year for 10 years. How many would you expect, by chance alone, to get say 70% correct?

If that number is >1, how do you think it would look for that person? It would look like they have an AMAZING track record! But, it's all just basically dumb luck.

Dan Harrington's view on donkaments is relevant here: it's like a lottery. Everyone gets a ticket, but some tickets are worth more/less than others. Pros' tickets may be worth 3 or 4; amateurs may be 1/4 or 1/2...and the card playing monkey may be 1/32. So maybe RK has a ticket that's worth 3...which demonstrates some skill and that he's not totally guessing. But the point is that he runs really good. And, running good is LUCK. It's survivorship bias, even though his success isn't entirely explained by luck, it's probably heavily influenced by luck whereas you're all mistakenly attributing it to skill.

But I think getting to the singularity itself is crazy. Well will have alot better technology and computational power, but there will still be things that we cannot do because it takes even more power than what we will have in the future. If you brought Isaac Newton to modern times he would obv be blown away, but there will still be many things he can conceive of, like controlling the weather, that we still can't do.

This is completely unfair. There aren't more than 10,000 intelligent qualified people making very specific detailed predictions with lots of evidence to support them.

The real sample size is like 20, and that's exactly why survivorship bias doesn't apply.


FWIW, the term futurist just means someone that acknowledges the exponential rate of growth of information technologies, etc. So yes, there are over 10,000 of those, but that doesn't mean anything.



Edit: I feel I didn't address the Harrington part of your post. For the most part, I agree. However, in this field, only such a small percent of the world will be taken seriously w/ their predictions. So if we're talking about the value of lottery tickets, then RK's ticket is worth 1,000,000 times as much as the 100th best lottery ticket.

"even more power"...

There are just so many avenues to produce this. Cold Fusion, Quantum Computing, and many things "less efficient, but easier to obtain". Hydrogen fuel cells, carbon nanotubes, whatever. And then there's all the stuff we've never thought of.

I don't think "even more power" will be the issue.



The way I see it is simple. For any "science" that we've started but not perfected, it's an issue of time before we solve it / perfect it.

Once we create conscious computers, that barrier of time gets minimized. 40 years of research / effort can be done in seconds, possibly with some real time in there for necessary "field tests".


The 40 years worth of technological improvement that will occur after machines become conscious, will take a tiny tiny fraction of 40 years.

Why can't you make the same argument in math? I personally probably have more computing power than the entire world 100 years ago and I am getting more as time is going on and computers become faster and cheaper. But I still can't prove every theorem because so many of the theorems apparently take way more than what I have. And the ones that are not proved are selected out as the harder ones because nobody has been able to prove them. I don't see why this can't happen with technology.

This is totally biased nutlicking. He's not THAT good. I think that your view of his abilities is circular.

How many people can you name that have 1/2 as much published about predicting the future of technology that have a better resume that Kurzweil?

I'll be the first to admit that I have no idea how many are out there, but I can't imagine it's close to a 3 digit number.

Until computers are conscious, they are merely doing what we tell them. We tell them things at human speed. A computer game takes more than a lifetime to make (when you factor in all the man hours). Once computers are conscious, a computer game will take seconds to make.

We can simulate trillions of people working for trillions of years, and have it all be done in less than a real year. Oh, and instead of people that make errors, they'll be people with limitless information.

So what happens to your body when your mind gets uploaded into a computer? This whole concept implies dualism afict. If only a copy of 'you' is uploaded (the 'real' you is still embodied) why would I care? It's like cloning. If your mind is transferred this seems to necessarily imply that the mind is not part of your brain* and I guess your body would just be what a zombie? None of this makes a lick of sense.

* Move your mind to a different kind of brain? I guess? But there's still the body issue, what becomes of that?

It's a gradual process. We've already started it with prosthetic limbs, cochlear implants. We're not far from neural implants. After the nanotechnology revolution, we'll start shedding ourselves of "standard body parts" and become more machine than human. This will inevitably turn into part of a network.

It's not like someone shoots you, and then inputs your DNA into a computer.

You should really watch the video I linked sometime. Not because it's right but you'll find it extremely interesting imo. Er I can probably find some shorter replies by Searle and know where to look, un momento.

Eh all of these videos roughly cover topics itt and are shorter than the first one I linked I'm not sure exactly which one replies to a specific question but anyways worth watching...

http://closertotruth.com/participant/John-R-Searle/90

Scroll down

There's some somewhat interesting stuff in there, but for the most part, he's just taking stupid theories, and exposing them for the stupid theories that they are.

Yeah, stuff directly about this thread is here and there in those videos. The other video I linked is a bit more on topic. There's some debates on the concept of the singularity with other smart people, like these:

Yudkowsky vs Lanier
Yudkowsky vs Pigliucci

Both of these are directly on point

I'm totally with this argument, its inevitable even if nanotechnology doesn't take off. Its started and its irresistable, I make no predicitions on when but there are no big leaps required so its just economics and the demand is massive+++

The arguments that singularity might still not allow stuff like control of the weather seem totally irrelevent, the singularity is more likely to mean the weather becoming irrelevent imo but in any case as far as I can see the singularity doesn't require control of everything.

I make one predicition on when: probably far too late for me

yea, because you really need a Phd in neuroscience to realize the idea we are going to over come physical death in 50 years is really ****ing stupid.

If anything Kurzweil is an interesting study in how smart people can be duped just the same as the average person...throw out some credentials then tell people what they want to hear.
Works every time.

Do you not see why this is NOT an argument? This is exactly what we'd expect from the lucky person who's the output of survivorship bias! Oh, and publishing is easy when it's not peer-reviewed.

I love the statement "I'm totally ignorant regarding how many are out there...but my uninformed intuition should still have merit!"

Kurzweil's entire argument is "Duh I see an exponential duh that means an asymptote!".

In all of his thousands of pages of writing on this topic, he offers no other arguments. His writings consist of attempting to bolster this laughably simplistic thesis by pointing to prior exponential growth (while conveniently ignoring the HUGE number of initially exponential trends which have died), and daydreaming about the singularity (while exploring none of the convincing limitations that will or may exist to such an event happening).

The guy is a clown. His few decade "singularity" relies on massive assumptions. I'll leave it to ZeeJustin to explain what these assumptions are, if he wishes to retain any credibility.