Wednesday, August 3, 2016

What evolutionary pressure drove human evolution?

Humans evolved a number of characteristics that differentiate them from other apes (less hair, shorter gut, weaker muscles, complex language, complex tool use, walking upright, dexterous hands, cooperation capable of larger tribes). What was the evolutionary pressure driving our unique evolution? Giraffes feed on leaves too high for other animals to reach, driving their long necks.  And thanks to the pressure of using dams/ponds for protection, the beaver's teeth wear easier in the back than the front, and are constantly growing, which means their chisel-like teeth self-sharpen with heavy use cutting down trees. Can we point to a similar niche that could explain the multitude of unique human adaptations?

To answer this question, let's start with determining the core evolutionary breakthroughs that led to modern humans?
It would seem to come down to two principal innovations: Complex tool use and Complex Language. I suspect most of the other adaptations depend on these two innovations. For instance, we can afford to have short/weak intestines because we mastered fire (tool use) which provides food that's easier to digest. We don't require (lice-prone) hair to cover our whole body because we can use animal skins and shelters for warmth (tool use). We are weaker and slower than other apes our size (and require ~50% less calories than Neanderthals) because we use tools, language, and cooperation for defense and hunting and can therefore get by with weaker bodies. We walk on two legs presumably so we can carry tools with our hands. And we form much larger groups than other primates, and are rather good at cooperating within our tribe, which is also predicated on language (without language you can't have such large groups- gossip, reputation, rules, religion, agreements, are necessary to motivate individuals to correct behavior and maintain cohesion in large groups, and all these depend on language). So I think to understand what was the driving force behind human evolution, we can focus on what sort of niche drove language and tool use.


To answer this question, note the following facts:
-Homo Erectus and Neanderthals hunted meat, including large mammals for food (but also ate cooked vegetables based on isotope analysis of bits of food on Neanderthal teeth!)
-Neanderthals were present in Europe for ~600,000 years and went extinct around the time of the arrival of humans in Europe 40,000 years ago 
-The quaternary extinction event is the widespread extinction of appx 173 large mammals (over 70% of large mammal species outside of Africa) that began around the time humans arrived on each continent (it was less dramatic in Africa likely because the mammals had time to evolve fear of humans as humans' hunting skills progressed). 
-In Europe, large mammals survived alongside Neanderthals for 550,000 years, but many quickly succumbed to the arrival of humans
-Thus humans were apparently quite skilled at hunting large mammals, apparently way better at it than the Neanderthals: In fact, if you accept that humans were the cause of the quaternary extinction event, you'd have to conclude that humans are the best hunters on Earth. I doubt any other predator in the history of Earth has caused so many extinctions.
-What makes us successful hunters? Tool use (i.e. spears, arrows (sometimes with poison-tips), traps, dogs, etc.) and complex cooperation and planning for the hunt (i.e. let's start a brush fire that pushes animals off that cliff, or let's all surround this herd of horses and on my signal slowly enclose them, spearing them as they escape).  I'd guess a group of 50 well-coordinated spear-wielding humans can kill pretty much any mammal they please. 

So, in conclusion, I'd guess that much of our unique characteristics evolved as a result of hunting.

Wednesday, February 10, 2016

What will people do when robots can do our jobs for us?

People have been worried that automation would lead to widespread unemployment ever since the Industrial Revolution. The Luddites, 1810's British textile workers, revolted violently against new labor-saving machinery that promised to replace skilled textile workers with machines operated by unskilled workers.
Apple, the largest publicly traded corporation in the world, makes around 100 billion per year in revenue, but only pays approximately 10 billion per year in salaries.
Researchers from the University of Chicago have documented a worldwide downward trend in the share of labor income in the last three decades, in the form of lower wages and benefits with increased inequality across industries, and a rising share going to capital income, beyond what can be explained by recession, and attributed to structural changes in technology, market structures and labor unions.[22]
I'm afraid  as technology progresses, there will be less and less need for human labor as computers and robots supplant more and more human skills. Experts vary on their estimate, but some are saying as early as 2030 (or perhaps 2100 or later), there will be a technological singularity. This is the moment at which computers will become capable of creating an even more intelligent computer. At this point, it's thought that there would be a sudden exponential increase in the intelligence and abilities of computers/robots, to the point that not only would computers become more intelligent than humans, their intelligence, organization, and superhuman capabilities would surpass our ability to even comprehend them. This could happen in our lifetime.
It may depend on your perspective, but for me this is a frightening prospect. Bill Joy's widely read 2000 article for Wired magazine "Why the Future Doesn't Need Us" explores the dystopian possibilities resulting from the singularity. There's lots of disturbing ways this could play out: humans might go extinct (through famine or violent robot rebellion), or we might end up with frightening levels of inequality with a tiny class of humans that own the fully-automated production of the robots, with no one else having any source of income.  According to this article on cnet.com, Bill Gates, Stephen Hawking, and Elon Musk are all concerned about the threat of machines that can think for themselves.
Having gotten terribly depressed by our prospects for the future, I tried to think of some scenario in which this yields a utopia. My optimism is not unfounded. Overall the world has improved since the Industrial Revolution- it's more tolerant of diverse religions/ethnicities/cultures, we do a better job taking care of the poor, we have less premature death from preventable or treatable diseases, the crippled can often be enabled to walk, and less people die violent deaths than ever before (see Steven Pinker's TED talk on The Surprising Decline in Violence from Biblical Times to the Present). Could things keep improving for us with the aid of the singularity?
My utopia would go something like this: since there wouldn't be any need for human labor, we would need to be provided welfare by the "government" (whatever that would be). Also, most humans I know thrive from activity (while reclining in sun under palm trees year after year, or spending all day playing video games, sounds appealing, I think in reality endless relaxation is a recipe for depression), thus we would need something to keep us being productive and active. One possibility for ensuring we stay busy in a world where human labor is obsolete is to create an economic system that rewards improvements in well-being. For instance we could be rewarded with "Well-Being Coins" for performing or achieving things that are likely to increase the well-being of ourselves or others. We might get 1 Well-Being coin for 30 minutes of exercise, 2 for visiting with an elderly neighbor, 4 for hosting a party, 1 for practicing violin, 2 for meditating, 2 for contemplating the divine, and 10 for completing a painting. We might even lose points if we spent all day alone despite feeling lonely, or if we didn't exercise all day but could have. Achievement of goals, nurturing relationships, improving health, finding sources of meaning, or doing activities that allow for "flow" would all earn people coins (this could all be tracked effortlessly by the tiny computers we'd all be wearing). The coins would be the only form of currency, and could be used for purchasing robotically-derived goods and services, or perhaps there would still be some commodities that people would still insist be human-produced (such as live music, art, drama, religious services, perhaps food?).
Kuwait is a modern-day demonstration that we are capable of creating such a society.  Kuwait is one of the top five wealthiest per capita nations thanks to proceeds from oil. If you're a citizen of Kuwait, your income is guaranteed- you're either provided with a government job (most people take this option), or some 10% work in the private sector, but if they don't make some minimum amount of income, then the government supports their income anyway. Note that everyone's given a government job, apparently to keep them busy.
Even in the run-up to the singularity, we're already experiencing declines in the need for human employment I'm afraid. I'm glad I didn't end up in the field of artificial intelligence, which is potentially an unethical profession if you believe the above, and I'll be supporting hand-made items and local small farms whenever possible. I'd say I won't be using Siri anymore, but for the moment she seems harmlessly simple.

Mathematical Argument for the Likelihood of Extraterrestrial Life

Great news from the Kepler Space Telescope data: astronomers are finding lots of planets in the habitable zones of our neighbor stars. So many in fact, that if you extrapolate to the entire universe, there may be, depending on who you ask, 5x10^19 habitable planets (i.e. rocky, with temperatures that would allow for liquid water). Intuitively, one would think it fairly certain that there are other planets with life given this massive number. To support this intuition, one can make the following mathematical argument:
Suppose for simiplicity there are 10^19 habitable planets in the universe.
And let L represent the chances that life of any sort forms on a habitable planet.
If L is 1/10^18 or better, then there is almost certainly life on other planets (for instance, if L = 1/10^18, the chances of exactly one planet (and no more) are 0.05%) (based on the Poisson formula).
If L were by sheer coincidence exactly 1/10^19 then you might actually have exactly one planet with life (i.e. earth), but even then the chances of exactly one planet are 1/e, or appx 36% (again based on the Poisson formula) (for the mathematically inclined reader, this could also be represented as lim (as x-->infinity) of (x-1/x)^(x-1) which, beautifully, equals 1/e). 
If L is 1/10^20 or worse, then the chances of any planets with life becomes increasingly tiny. (even if L = 1/10^20, chances of life having formed anywhere at all would be 10%, and if L = 1/10^21, then chances are 1%). 

The fact that life formed soon after Earth's temperatures became habitable means that L is likely WAY better than 1/10^19 (to give you an idea, there are appx 10^19 grains of sand on Earth). 
A couple interesting articles:

1) It's actually possible that we could identify life on other planets within the next ten years by searching for "biosignatures" in planets' atmospheres (looking for combinations only likely to occur w/ life)

2) Possible alien megastructure could explain some unusual data from a particular star, discovered by the Kepler Space Telescope: