Transterrestrial Musings

Comments

Watson understands that "extreme longevity" is a hard problem and that at best a solution isn't just around the corner. He didn't have to "offer evidence" because his opinion wasn't on trial. But if you want evidence, it's not hard to find. Life expectancy at age 80 has gone up a whopping year and a half in the past 30 years.

http://www. infoplease.com/ipa/A0005140.html

Posted by at October 2, 2007 07:08 AM

But if you want evidence, it's not hard to find. Life expectancy at age 80 has gone up a whopping year and a half in the past 30 years.

To quote Rand: "You provide no basis for your belief that we will not be able to live longer, other than that we never have."

And also note that if you chart the data points from the link you provided, the curve is clearly exponential, not linear.

Posted by Stephen Kohls at October 2, 2007 08:34 AM

A reason to be cautious about optimism on this issue is that it seems to be a poster child for potential wishful thinking. So, the cognitive bias will be to be more optimistic than the evidence warrants.

Posted by Paul Dietz at October 2, 2007 09:13 AM

Clarke's law suggests that Watson would have been wrong if he had said that extreme longevity was impossible. But he didn't say that. He was asked if extreme longevity was possible in the next few years, and Watson said he doubted it.

I'm with Watson, though: I don't see long lifetimes in the foreseeable future. Like flying cars, lifetimes of 150+ years have been predicted for about a century, and those predictors have been wrong so far. Like Turing-quality AI, longevity is a technology that humans will probably have eventually, but there's no reason to think that we're on the cusp of figuring it out.

I have my own heuristic to detect unwarranted hype: if a wonder-technology has regularly been on the cover of magazines like Popular Science for the last 20 or 30 years, and it's never panned out before, it probably won't pan out this time either.

Posted by Ashley at October 2, 2007 09:18 AM

Rand has a point. However, it sounds like he's saying "unless you can come up with specific reasons why we will not achieve extreme longevity, you should believe that we will achieve extreme longevity." (This is a bit of a straw man--mostly Rand is saying that there is nothing worthwhile in that argument to sway him from his current position, because "we've never done it before" is unconvincing.)

If I understand correctly, this is more or less the position that Aubrey de Grey takes. He has a specific list of things which are the root cause of aging and death, and he takes the position that if all of them are addressed we can conquer aging and death. And he challenges everyone to come up with additional issues.

The problem I see with this position is that just because I can't think of any other mechanisms that cause aging and death, it doesn't mean that no such mechanisms exist. And I also have no way of knowing how amenable the unknown mechanisms are to fixing. I think the same sorts of issues apply somewhat to artificial intelligence and/or "uploading" of consciousness into computers. We don't even know what we don't know.

I would think that research into the causes of aging (along de Grey's line or any other reasonable approach) would help us discover any unknown causes of aging, which we would then be able to try to also fix. But there is no guarantee.

As an aside, I tend to agree with some of the comments I've read about engineering vs. research--we don't necessarily need to understand everything that happens in the body, we just need to be able to fix the things we need to fix.

I think the path of technology is more unpredictable than we sometimes acknowledge. Nobody predicted the microelectronic revolution--futurists used to predict giant computers in huge buildings. Lots of people predicted a helicopter or aircar in every driveway. Lots of people thought fusion power would be easily available now (although it looks like we might actually be getting closer now--do a google search on the two words polywell fusion). I read some of Drexler's nanotechnology stuff back in the early 1990's, and it seems like we're making progress, but we're not where he thought we'd be 15 years later (I'd have to go check his predictions to be sure).

In the medical field: I used to read about Jarvik's artificial hearts in the 70's and 80's. I would have predicted that we'd have lots of well functioning artificial organs by now--yet we went down the path of transplants and antirejection drugs. Who knew? (Some did, but not everyone.)

I think there's an excellent chance I'll live a longer, healthier life than the average person 200 years ago. However, I'll be (pleasantly) surprised if I make it past 100. And I think it's because the problem of aging is harder than we think, just like some other engineering feats have been harder than we think.

Maybe I'm just having a minor midlife crisis, and I think facing my mortality by hoping for a super long life is wishful thinking.

Posted by Jeff Mauldin at October 2, 2007 09:30 AM

A little further analysis of those numbers:

Using only 1 significant figure for a, I have fit the data (least square error) to the function life_expectancy=K*exp(a*year).

For age 0, k=0.025 and a=0.004
For age 10,k=1.18 and a=0.002
For age 40,k=0.092 and a=0.003

Then to check these results, I threw out all the data prior to 1990 and did the curve fit again:

For age 0, k=0.19 and a=0.003
For age 10,k=0.16 and a=0.003
For age 40,k=0.012 and a=0.004

Since the exponent didn't change appreciably, it should be relatively safe to say that the curve is exponential with about a 0.003 coefficient.

So now it becomes simple algerbra: At what point does one year passing give you an additional life expectancy of greater than one year?

For a ten year old:
0.16*exp(0.003*x)=0.16*exp(0.003*[x+1])-1
x=2550

For a 40 year old:
0.092*exp(0.003*x)=0.092*exp(0.003*[x+1])-1
x=2750

A sensitivity analysis shows that errors in the estimate of the exponenital coefficient need to be extremely large in order to bring these dates into our likely lifetimes.

Unfortunately this seems to mean that baring breakthroughs that change the curve, we will not live indefinately...

Posted by David Summers at October 2, 2007 09:32 AM

Ashely-
You have taken a more extreme position than Watson. He expressed doubt about extreme longevity in "the next few years." You have expressed doubt about "our foreseeable lifetime." Untless you're in your 80's right now, I'm going to disagree with you.

The cost of DNA systhesis and readability is falling about 10x every three years. We're less than 10 years away from everyone in the developed world having their genome sequenced, if they want to pony up the $1,000. Protein folding that used to take supercomputers can now be done on a smallish Sony PS3 cluster. Advances in microfluidics allow labs-on-a-chip to perform tens of thousands of experiments in the time it takes today to do one.

If by "next few years" you mean 3-5, I'm with Watson. But forget my "lifetime"; the 2010 generation of kids will have always had longevity the way I've always had a VCR and home computer.

Posted by Brock at October 2, 2007 09:43 AM

Watson's still alive?

Posted by Barbara Skolaut at October 2, 2007 11:59 AM

I would have predicted that we'd have lots of well functioning artificial organs by now--yet we went down the path of transplants and antirejection drugs. Who knew?

Larry Niven comes to mind...

Posted by Ilya at October 2, 2007 12:00 PM

You have taken a more extreme position than Watson. He expressed doubt about extreme longevity in "the next few years." You have expressed doubt about "our foreseeable lifetime."

No, he said "foreseeable future".

I consider the foreseeable future to be a rather short period (which is why I consider the railing of some space fans against the short term focus of their favorite boogymen to be risible), so it's not obvious the claims are all that different.

Frankly, I see no reason why extreme life extension SHOULD be easy in organisms. There's no reason for evolution to have included that capability, and some of the things that are likely to be going wrong (like, random mutation of the DNA in all your cells over time) look extremely hard to fix absent almost magical nanotechnology.

I suspect the way extremely longlived organisms will be developed is by engineering them from scratch.

Posted by Paul Dietz at October 2, 2007 12:09 PM

"I would have predicted that we'd have lots of well functioning artificial organs by now--yet we went down the path of transplants and antirejection drugs. Who knew?"

And someone with a heart in his chest cloned from his own tissue would need minimal or no medical attention or treatment thereafter. Bad for the medical profession, just as natural treatments that clear up the problem instead of covering up the symptoms are bad for the medics. As medics control the research and approval...

Posted by Fletcher Christian at October 2, 2007 03:08 PM

Given that a human being can grow from a single cell to a healthy adult in about twenty years, and continue to repair itself with a reasonable degree of success for a further forty to sixty years, it has always seemed possible to me that we could at least understand the mechanism that causes the breakdown of this process and perhaps fix it.

I realise that bio-chemistry at this level is hard, but unless it is impossible for our brains, even with technological help, to understand what's going on we will crack it.

The knowledge that something can be done is supposed to be the first, and hardest, step to doing it.

Well self-repairing biological mechanisms can be done. We are all living and breathing examples of that. So let's get on and do it.

Posted by Kevin_B at October 2, 2007 05:12 PM

Science Fiction has a lousy record of predicting the future. In the 1930s, for example, it was widely held that by 1970, toga-clad descendants of the Depression Generation would be living in giant art deco cities full of speeding Dymaxion Cars and dining on food pills. In the '50s and '60s, it was rocket belts and atomic-powered flying cars we were supposed to be enjoying by 2000. And today? In almost every extrapolation of the future I've read lately, the ultimate fate of mankind is uploading -- the transference of consciousness from biological to digital substrates. Such uploads, it is claimed, will be immortal and (within the environments they inhabit) omnipotent as well. "Ye shall become as gods," the futurists tell us, 'knowing Good and Evil... AND ye shall not surely die."

Leaving aside the soteriological results of our last gamble for that prize, one question always remains in my mind after reading the latest sci-fi scenario of our imminent apotheosis: so what?

The question seems facetious, but I'm asking in earnest: what's the point? Given unlimited power its environment and an eternity in which to shape it, the human mind is doubtless capable of creating paradises both gross and subtle. Yet, when spread over capital-E Eternity, surely even human creativity ends up as pretty thin gruel. And, when one factors in the exponential speed of thought that will supposedly be available to our uploaded selves, surely Eternity will end up being a lot longer than we might imagine. After all, eventually even the most creative of uploaded minds is going to reach the limits of its creativity, even if that creativity is expanded by collaboration with the creativity of the billions of other uploaded minds. It may take a million years, or a billion, but sooner or later every iteration of every creative idea is going to be experienced by the godlings that once were us. And then what?

(Some would argue that the AIs we're going to build will transcend our own intelligence and create godlike super-intelligences that will be
"as gods", providing meaning and purpose for post-humanity. I myself don't see a market for such machines. To what purpose would people of today invest in their construction? "We're looking for investors to help us build a machine that will instantly render all human life obsolete, Mr. Bigbux. How much shall we put you down for?"

So we become immortal, all-powerful digital beings. What then? An eternity of reruns doesn't sound like much fun to me. A cosmic reset button, making us forget everything so we can start afresh? Sure, but that dooms us to an endless cycle of resets. It pains me to think that our deified, uploaded descendants may eventually decide to delete all backups and pull the plug on the whole works out of sheer ennui.

I hate to say it, but the Heaven promised us by Moore's Law sounds more like Hell to me. To my (admittedly limited) mind, a "heaven" that is limited by human ingenuity is always going to end up in a yawn. Perhaps this is why God's Heaven is described as being beyond human comprehension — to give us an eternity that goes beyond what we ourselves can create.

Don't get me wrong — I'm all for living as long as possible. I want to walk on the moon, for one thing, and I intend to stick around until I can. But living forever — as an uploaded mind or as a biological human — eventually leads to "So what?"

(And I haven't even discussed the effect of physical immortality on the human family. After all, people who live forever don't need to have offspring. I for one would not want to live forever in a universe without children.)

Nope, I don't buy it. I know that most SF writers are way smarter than I am, and I know they have Moore's Law on their side, but I've got to believe that there has to be something better as the end of existence than boredom. I may be a Luddite, but I'm hoping that the future of uploads and artificial gods that we're being promised now ends up looking as silly as the rocket-belt-and-toga future we looked forward to Way Back When.

Posted by bchan at October 2, 2007 08:39 PM

bchan:

Sure, an endless existence could get pretty boring and futile; but wouldn't you like to yourself have the option on when the plug is pulled?

Second point: Transhuman AI is certainly not going to be the goal of advancing technology, for the reason you have stated. However, once two things have been put together then it becomes inevitable. These two things are sentience-grade (essentially alive) AI and a way for that sentience to affect the material world. That combination leads to the Singularity by endless - or nearly endless - bootstrapping.

Returning to the immortality question; firstly, what would the reaction of religion to uploading be? After all, by some lights you are trying to cheat death by the process. Would they say that an uploaded human was a soul-less robot, and the real human died during the process and is guilty of suicide, and would go to Hell? (Suicide is a sin in most religions, although sometimes not if it's in the service of your god.) Or that you were going to Hell eventually anyway, for attempting it. After all, given enough time, any conceivable precautions against equipment failure would fail.

And if you are counted as a living being while running in a computer, does pulling the plug on yourself again count as suicide?

And what if your program is copied? All deep questions. Sooner or later, all religions may have to try to deal with them.

Posted by Fletcher Christian at October 3, 2007 04:28 AM

In almost every extrapolation of the future I've read lately, the ultimate fate of mankind is uploading -- the transference of consciousness from biological to digital substrates. Such uploads, it is claimed, will be immortal and (within the environments they inhabit) omnipotent as well.

You must read a very select set of science fiction. I read far more varied extrapolations of the future, and while uploading of consciousness features in many of them, it is never one ultimate fate of mankind. If anything, the most common theme I see is splitting of mankind into many branches with different levels of augmentation and/or uploading and/or mind-to-mind communication. Which I find much more likely than any one "ultimate fate", BTW. And it is interaction between these factions (some of which barely see each other as human any more) is what makes for interesting stories.

Posted by Ilya at October 3, 2007 06:18 AM

Whatever its flavor, the current avenir du jour still leaves a terrible taste in my mouth.

I don't think it will ever happen, myself. First, I don't think God will permit it. Second, I don't think anyone outside the Tranny/Extropy/Singularicult will want it. Third, I don't think it's possible (in other words, I think that there's something to consciousness that transcends matter/energy/space/time).

The Moore's Law thing has always struck me as kind of funny. Saying that a computer with a zillion-ops-per-second capacity will somehow "wake up" is like saying a car with a million cylinders will somehow be able to reach the speed of light.

What amazes me the most about our times is how little things have changed from the days of my youth (c. 1965-1980). Other than microelectronics, not all that much has really changed for the average person since 1965. Everyone still goes to work, men still wear pants, TV is still the top form of entertainment, people still drive gasoline-powered road vehicles, go to church, fall in love, and have kids. You could take a 1967 man and plop him down in 2007 America and he would be up to speed in a few days. In like manner, I think that the future will be much like the present. Computers will do what they do now, only faster. People will still have bodies, eat food, and reproduce as they do today. Humans will remain human — flawed, stinky, and utterly fascinating.

And if I'm wrong? My plan is to hide in the mountains until everyone's uploaded, then walk into town, unplug the server, and take their stuff.

Posted by bchan at October 3, 2007 09:34 PM

Post a comment