?

Log in

No account? Create an account
 
 
28 September 2011 @ 02:57 am
Growth in life expectancy - for real  
Just until today I believed that technical progress adds about 1 year to our lives every 5 years.
I derived it from this promising chart of life expectancy at birth.

So I reasoned that if life expectancy now is 79 years, then in 45 more years (when I would be 81) life expectancy would increase by 45/5 = 9 years.
9 extra years of life due to the technical progress which I'm going to live through.

Well, not exactly.

Here's more relevant data in form of Life Expectancy by Age.

It shows that white male in the year 1850 who managed to live until 80 years old - was likely to live for about 6 more years (and die at about 86 years old).

So you would guess that 150 years of technical progress would significantly improve life expectancy of 80 years old guy.
The improvement, in fact, happened.
In the year 2004, eighty years old male is likely to live ... drum-roll ... 8 more years.
Just 2 more years longer than in 1850.

So if 150 years of technical progress and growth of wealth yielded 2 more extra years of life for 80 years old - what kind of improvements can we expect in the next 50 years?
I guess the progress can cough up 1 (one) extra year.

So when you are 80 years old - expect to live 9 more years (or 11 more years if you are a female).

Update:
Judging by the progress in 1990-2004 the outlook is not as grim:
14 years of progress added whole year to life expectancy of 80-years old.
If the progress can manage the same pace - next 45 years might add 3 more years.
 
 
 
mehanizatormehanizator on September 28th, 2011 07:05 am (UTC)
the progress is nonlinear.
Dennis Gorelikdennisgorelik on September 28th, 2011 07:15 am (UTC)
So what kind of increase in life expectancy at age 80 do you expect from our non-linear progress?
mehanizatormehanizator on September 28th, 2011 07:45 am (UTC)
I assume this parameter has asimptotic growth with singularity around 2040-2050. so you may try some asimptotic function like tg to extrapolate.
Dennis Gorelikdennisgorelik on September 28th, 2011 01:49 pm (UTC)
Do you mean that by 2050 people's life expectancy would be close to infinity?
mehanizatormehanizator on September 28th, 2011 01:53 pm (UTC)
exactly. whichever technology comes first: cell time clock control, bodyparts replacement or personality transfert.
Dennis Gorelikdennisgorelik on September 28th, 2011 01:55 pm (UTC)
How did you pick that date (2050)?
Isn't it by any chance the time when you would die if these amazing breakthroughs will not materialize?
mehanizatormehanizator on September 28th, 2011 02:04 pm (UTC)
adepts of transhumanism widely use 2030 as the singularity date. I've just decided to be pessimistic about progress rate, though I can't see any reason why immortality cannot be reached in some point in future. so function is asimptotic, the question is only where this point may be.
Dennis Gorelikdennisgorelik on September 28th, 2011 03:40 pm (UTC)
adepts of transhumanism widely use 2030 as the singularity date.

Many transhumanism adepts are older (Ray Kurzweil is 63 now) and would likely kick the bucket by 2040. So they hope for the cure in 2030.

Significant life span extension can be reached in some point in the future, but it's very unlikely to get these results in the next 40 years.
Just consider what happened in the last 40 years (since 1971) in the field of life span extension. Nothing major, really.
mehanizatormehanizator on September 28th, 2011 03:49 pm (UTC)
progress is nonlinear :)

I expect major AI breakthrough in those 40 years. everything following that point will be matter of years, if not months.

and... let's consider what happened in 40 years before 1945 and 40 years after in the field of nuclear device application.
Dennis Gorelikdennisgorelik on September 28th, 2011 03:54 pm (UTC)
Super-smart AI does not shrink time required for new research as much as you would hope.
Research is impossible without experiment, and experiments still would take quite some time (especially when we are talking about researching human life span extension).

I understand why you anticipate breakthrough in AI in the next 40 year.
But why do you anticipate breakthrough in human life expectancy in the next 40 years?
Do you see any precursors to that?
mehanizatormehanizator on September 28th, 2011 03:58 pm (UTC)
since you've got super-smart AI, you've got everything else, immortality will be the first problem to solve.
Dennis Gorelikdennisgorelik on September 28th, 2011 04:08 pm (UTC)
Why would supersmart AI try to solve human immortality?
And even if it would - don't you think that some time-consuming experiments would be required, which would likely bring any meaningful progress in that direction in the range well past 50 more years after super-smart AI creation.
mehanizatormehanizator on September 28th, 2011 04:15 pm (UTC)
because it's a tool, humans want immortality, it will deliver. you don't think it will have its own wishes out of nowhere, do you?

"time-consuming" about researches exists only because of limitations of human timescale. when you'll get humans out of the process, it will be like an explosion. because computing power of planet is enourmous even now. but there's no algorithm for this explosion.
Dennis Gorelikdennisgorelik on September 28th, 2011 04:23 pm (UTC)
Of course AI would have its own wishes (goals). Without such wishes it won't be able to function.
Then natural selection would make sure that only AIs that have sufficient self-preservation goals would survive.
Human bodies are likely left behind.

How could you take humans out of the process of human life span research? AIs would need to validate their theories somehow, right?
mehanizatormehanizator on September 28th, 2011 04:28 pm (UTC)
AI is just an algorithm, what natural selection are you talking about? there can be no goals except predefined by humans.
Dennis Gorelikdennisgorelik on September 28th, 2011 04:35 pm (UTC)
For starters different people would put different goals into AI systems that they design.
Systems that would have self-preservation goals would survive better.
Systems that have self-replicating goals would have even better chance of surviving.
That's how natural selection works.
mehanizatormehanizator on September 28th, 2011 04:37 pm (UTC)
survive what?
Dennis Gorelikdennisgorelik on September 28th, 2011 04:39 pm (UTC)
Survive competition for the resources (e.g. electrical power, better hardware, repair parts).
Denis Novaktaganay on September 28th, 2011 01:34 pm (UTC)
Live expectancy for the age of 0 in 1850 was 38.3, in 2004 - 75.7.
So thanks to progress your chances to be alive at age of 80 are much much more than it was in 1850.
I believe live expectancy at the age of 120 now may be that the same or even worse than it was 100 years ago, so do it mean that people will live same number of years now as it was in 1850? Definitely not.
Dennis Gorelikdennisgorelik on September 28th, 2011 01:53 pm (UTC)
Considering that I'm 36 now, life expectancy at birth is not really relevant to me now.
My point here is that technical progress significantly improved life expectancy of newborns and teenagers.
But did not really extend much maximum life span for humans.
Denis Novaktaganay on September 28th, 2011 01:57 pm (UTC)
I am afraid there are system limitations in our body which may be very little affected by technical progress. So progress will help you to use all resources of your body, but it will not add new ones.
Dennis Gorelikdennisgorelik on September 28th, 2011 02:06 pm (UTC)
Yes, it looks like that.

I'd add that these body age limitations are not mistakes/omissions, but were intended by evolution (to avoid stagnation in human society).