However that means 24 vs 35 years. Which means even the poorest men live to on average 79 and the richest 10% of men (maybe 100k in income) live to 90 on average.
And the gap between genders is tiny, only 1 year or less.
Is any of that valid? It sounds fishy. I’ve never heard of any life expectancy tables where men could expect to live to 90 on average (isn’t men living that long fairly rare), nor have I heard one where the gap between men and women was only a year or less, it is usually several years.
Most life expediencies you’ve probably seen are from birth. These ones are from age 55. If you make it to 55, you do have a pretty good chance of making it to 90 (still significantly less than 50%, taking into account all income levels).
Yeah but most people born will make it to 55 (about 90-95% of people, roughly speaking). Only about 50% make it to 80, and only about 20% make it to 90 according to survival curves like this.
Anyway, I guess it is possible. the SSA says that life expectancy at 55 are about 25 for men, almost 29 years for women.
But that is part of the problem, the chart in the OP says there is almost no difference between men and women, the SSA says there is a 4 year difference.
The OP’s data is for people born in 1920 and 1940. You get your 95% from the top line, people born in 1991 (I assume that the years represent birth years. It doesn’t say).
If you look at the 1921 line, it looks like about 75% of people make it to 55. From there, the median life expectancy looks like it’s about 80-ish, or an extra 25 years. That lines up pretty closely with the OP’s 1920 data. If we look at the 1941 line, it looks like about 80% of people make it to 55, and then the median life expectancy at that point is maybe out to 85, another 30 years, which is also pretty close to the OP’s numbers.
Besides for the fact that these rates are for people who are already 55 (and also the fact that according to the quote, it’s only the richest 10% who make it to 90, which is a small percentage of the population) it’s worth noting that many mortality tables are based on projected improvements in longetivity in the future. So even if you’re not seeing a whole lot of 90 year olds now, you might see a lot more in the future if the predictions are right.
Those of us born in the UK during WW2 had a pretty healthy diet during those vital formative years. All the bad stuff: sweets, fat etc were rationed and that, together with our eagerness to get outside out of sight of authority, contributed to our longevity. I doubt that today’s fat, idle kids will live so long and the curve on the graph will level off and, eventually drop.
There are several classes of chemicals that in the lab are expanding mice lifespans pretty well. However I have no idea how many, if any, will become effective and safe human drugs within the next 20 years.
That is the one part of the singularity that makes me rethink it. There are still various real world bottlenecks that intelligence may not be able to surpass (at least at first). You still need human trials done in real time to test a new drug. Then again maybe a superintelligence could build a working model of human biology and do the equivalent of 20 years of testing in 3 seconds.
Exactly. A home computer could beat any human player in checkers by, what, the late 80s? It took about another 15-20 years or so for that to be true with chess. (Home computer, not Deep Blue). What’s the difference between chess and checkers, from a computer’s point of view? Number of variables. Human biology has MORE variables than chess, but not an infinite number, and you don’t need to model the biochemistry of an entire human body perfectly to stop most of the major aging processes. Aging is, after all, mostly cellular. Keep the cell from aging, and you can keep the BODY from aging.
Of course, the reversal of aging might be extremely expensive for a few years, just like genome sequencing was for a while. However, with each iteration of Moore’s Law, bigger engineering problems become trivial faster, because there’s a much bigger real-world-application change when going from 1 billion FLOPS to 2 billion than from 1 FLOP to 2 FLOPS. Therefore, the turnaround from “available to the rich” to “available for anyone” is going to get faster and faster for any given technology with each iteration of Mr. Moore’s Law.
I don’t know what to think anymore. I hope strong AI shows up in my lifetime, but I have no idea. Elon Musk (someone whose intelligence and understanding of tech trends I really respect) feels strong AI is only 5-10 years away based on what he saw in person of labs studying it. Musk was apparently an investor in DeepMind. When someone like Musk who invested in those companies says it is 5-10 years away, that may be a valid window.
I remember back around 2004 when DARPA did the first automated car challenge. All the cars failed within a few miles of the starting line. By the next year several beat the course. Now automated cars are safer than human cars and it all took less than 10 years.
The key is that engineering problems like self-driving cars, cellular aging, etc. are not mathematically infinite. And the other important part (already used key :p) is that each iteration of Moore’s Law is literally twice as important, real-world-application-wise, as the previous one. It’s exponential, NOT additive.
I personally think Ray Kurzweil is more of an authority than the venerable Mr. Musk, and Ray is still predicting 2029. I’m certainly HOPING for sooner, for many reasons.
I’ve got some serious brain problems that make functioning a struggle.
I’d love to see Gordon Moore live to see the end of his law. He deserves it, I think, and I imagine he’d get a real kick out of it.
I have a new baby at home, and I’d LOVE to see the singularity from her point of view. I’ve known about Moore’s Law and the singularity for over 15 years now, but I’ve had to miserably slog through my crappyish life, hoping that improvements in technology can help people like me in the future. I’d like to see what the Singularity is like from the POV of someone who has never known the crap of adulthood and jerky bosses and bad relationships, etc…
I’d also like to see what the transition is like for older people like my parents, who grew up in rural Appalachia, and also for people from remote tribes. I also just don’t want to lose my parents. I hope they survive long enough for the Singularity to make us all immortal.
I want to know if FTL travel is possible, what the Grand Unified Theory is, what’s inside a black hole, etc… I’d also like to talk to some aliens. I’m curious what the extremes are (in every sense) of sentient, high technology species. How big/small can they be? Can they be reptilian? I’m also just curious as to what kinds of environments can support complex life. Can vertebrates evolve in a methane ocean? As an example.
This. Four years ago I asked here about life expectancy data for the US. From the data linked to by Galileo, I came back a couple of years later with this analysis, which among other things included this plot showing odds of surviving another 1/5/10 years as a function of your current age. The data I provided there only gives 10-year survival odds at most, but you can see the pattern: the 10-year odds for a 55YO are different from the 10-year odds for a 2YO, and the same pattern would exist for X-year survival odds: if you’re 55, you have a greater chance of living to the age of Y (where Y > 55) than someone who was just born, because you’ve already made it past infant mortality, adolescent stupidity, and however-many years of exposure to death and disease; the newborn still has to go through all that before he gets to where you are, and he might not make it.
This is good, especially the graph. The saying about statistics is that the average person has one breast and one testicle. Saying a 55-year-old has a 34-year life expectancy on average does not mean much, considering there’s a lot more room to die between 55 and 90 than after 90. None of the 50% who make it to 90 are going to drag on for another 34 years.
The dichotomy between rich and poor hides a lot of other details; the rich have better health care, may see a doctor more often, may afford and be taking their blood pressure medication; will get better treatment sooner with good medical coverage in the event of heart attack or such; some poor may be there because of lifestyle choices like alcoholism which also affect income. Rich retired people may have less stress, be better fed and cared for, etc. Perhaps the most interesting fact is that the gap between the richest and poorest is so narrow.
The problem with doing biochemistry on a computer today is *mostly *not the lack of computer power.
It is that you can’t model what you don’t already understand.
And our ignorance of all things biochemical is huge. Yes, we know a heck of a lot more than we did 5, 10, or 40 years ago. But that is only saying we now understand a larger fraction of one percent of what we need to know to actually produce useful models with useful predictive powers. Moore’s Law is no help with this.
I refer you to a blog at http://pipeline.corante.com by a medical biochemist. He’s written a bunch on the gulf of understanding between the working chemists in the field and the arm-waving tech entrepreneurs’ rosy predictions.
I’d provide a more precise URL but they seem to be having technical difficulties this morning, at least for me.
I don’t need a link. I’ve seen this claim before, and it has validity, to a degree. However, there’s nothing magical about the molecules in a cell. They obey physical laws. Given that, a fast-enough computer can take all that we DO know about these molecules, and quickly come up with the list, however long, of possible alternate theories that would explain all those known facts. All that remains is the testing of those theories on that list. Once we have strong A.I., testing those theories isn’t even going to require a lot of human input. However, there is a bit of a devil in the details, no question.
The fact that the laws of physics, chemistry, math etc are not infinite (as far as I know) is what I always come back to. The rules that govern this universe are finite (for all intents and purposes). Pi for example, just keeps going. But once you get to about 39 digits you can calculate the circumference of the universe down to the angstrom. Real world applications require far fewer digits. So even if something like pi has infinite complexity, for real world applications there is a finite number of digits you need, usually less than 6. I’m not a math person but I hope people get my point, real world problems do not require infinite solutions.
In the marvel universe there is a character called ‘brainiac’ who is a ‘12th level intellect’.
As far as the singularity, I have heard even among critics of the idea the real criticism they have is the timeline. Human knowledge and the human knowledge base keep growing. Human problem solving abilities keep growing (especially now that we are adding more and more machine intelligence problem solving into the mix). Our problems remain the same. If you went 100 years into the future, our bodies will not change much due to evolution. But our abilities to understand them and solve the problems they cause will be far far better.
If you use a scale like the one from Marvel, and current earth has a 6th level intellect (meaning all humans on 20th century earth combined are a 6th level) I’m assuming by the 2020s earth will have a 7th level. Then an 8th, then a 9th. However our problems are not growing in complexity and each addition to the worlds knowledge base will make further discoveries possible.
On a long enough timeline, even if you reject the concept of a singularity, you come to the conclusion that the world’s problem solving abilities will vastly surpass whatever environmental and biological problems we face. If our problems are a 10 on that scale, our problem solving abilities will one day be a 10. Then an 11. Then a 25.
I am reading a book about scurvy in the age of sail. The best minds on earth back then spend centuries trying to figure it out. Now that we have a knowledge base and better problem solving abilities (If we still didn’t understand scurvy today we could just run the nutritional treatments that work through a GS/MS to extract the effective ingredients, synthesize them organically and run a double blind test to see which ingredients prevent scurvy). Even if we lacked the knowledge of the cause (and luckily knowledge doesn’t die it passes from generation to generation) our problem solving abilities are so much better that we could find the cause and cure much faster than people did in the 17th century.
Also as far as Kurzweil, I think he has reignited an interest in science that wasn’t there before. It is also a case of life imitating art. He puts his ideas down (he didn’t create the singularity idea, but he popularized it) and as a result billionaire companies and billionaire individuals are investing in ways to help make it happen. Google, Peter Thiel, Larry Ellison, Dmitry Itskov, John Sperling, etc. Google is investing heavily in machine learning and anti-aging now, who knows if they would have done that w/o Kurzweil’s influence on popular culture.
For me I first discovered Kurzweil’s work when I was an undergrad studying science. And his ideas totally revamped my ideas about science and what it is capable of. I think Kurzweil has done as much as the 60s space race to ignite an interest and respect for learning and science. Whether people like or hate his ideas, I believe he has helped make the problem solving abilities of science far easy for people to visualize and shoot for. He gave a lot of wealthy, powerful people a goal to shoot for.
While I can’t speak about comic book characters, there are much simpler, practical, and obvious ways to make a human being de facto immortal.
It actually does not require magical technology. It may in fact be possible today if you had enough money.
Freeze their brains by injecting a very cold solution of cryoprotectants in saline directly into the arteries feeding the brain. Or, a plastinating solution - this is how they preserve rat brains for study. The patient’s brain is ideally still alive. (but they can be legally dead - legal death in the Western world is defined by the heart stopping)
Slice their frozen brain into 50 nm slices
Scan each slice in parallel with essentially 1 million electron beams (1000 arrays of 1000 beams)
Use machine vision to convert the shadow images from the electron beams into mappings of protein identity
Use empirical data to winnow the map to the proteins that reflect the strength and excitability coefficients of each synapse
Use empirical data to construct a state machine computational model that predicts correctly what test synapses will do if they are stimulated and in a given state. Determine the internal rules that advance the machine from one state to another.
Build a physical representation of that state machine with probably millions or billions of separate computer chips. Burn each one with coefficients taken from your scanned brain.
Network all these chips together with substrate level optical transceivers and a crapton of connections (no one ever said this would be cheap).
Fire it up.
Oh, and also, test #4-9 on smaller analogs of brains like nematodes, etc until they are rock solid techniques.
Once you do this the first few times, you could start scanning the world’s best chip designers, biological scientists, engineers, etc. Run the emulated versions of them at higher speeds than the human brain runs at normally. Have them help you improve your process.
The gains should be exponential. After you make the process cheap enough, you can make any human immortal this way, since once a person is just a digital data file, it is quite trivial to make enough backup copies that no event that is likely to happen before the heat death can destroy every copy.
This sounds about as hard as reversing cellular aging. I imagine it will be, generally speaking, an area of research for at least a short period between now and the assumed Singularity, though.
