Peter Thiel on the Global Economy, Technology, Artificial Intelligence [video]
youtube.com
youtube.com
The advances in computers are advances in processing power and connectivity. Every industry uses both of these elements to succeed. Now that computers are reaching a sort of peak development, there is going to be a dispensing of the capabilities that computers offer into other industries.
Its hard for me to put it into words, but I think that computers are simply a stepping stone for real achievements to come.
The car has basically reached peak performance years ago, but we aren't disappointed in the fact that there was a lot of focus improving the car and a giant vehicle industry for many many years. Instead we moved on to building things that can use a car, such as the highway system. Or the mail delivery system, or the thousands of industries that completely rely on highways or mail.
In the same way with computers, we are now going to see industries and technologies appear that rely on having computers that have reached this advanced level of processing power we have today. Once this mighty computer is passed on to the common man, innovation begins to move in different paths again. Like a tree.
I dont know, this may have delved into a bit of a rant but that video made me extremely optimistic :)
You do make good points, and I up voted you.
He's said before that he's not particularly worried about advancements in computers continuing to unfold and the potential for the Net to become ever more like a planet wide nervous system with many more use cases coming into existence.
His complaint is that if progress does not occur in fields outside of computation there will be stagnation. No amount of computational intelligence can mine asteroids or convert the Sahara to arable land without doing something with the material world.
You can have rapid improvements in specific technologies in the midst of a general decline or a stagnation. For example; surely there a great deal of improvement in weapons systems during the middle ages. The stirrup, the crossbow, the gun, the cannon. Most of the advantages of those technologies didn't really help the average person's situation.
This is not a hypothetical. While there was enormous increases in computational power in the past three - four decades median wages have not budged. People were able to consume more primarily due to debt.
We ought to question whether computers will be enough to raise living standards in the developed world in the short to medium term (his main contention I think).
Way too many smart people are tied up doing rather banal things with computers, or in finance. That's capitalism in action.
What isn't being worked on? Thiel is bothered by lack of progress in medicine. Medicine is now working on really hard problems, most of which are consequences of aging. Short of redesigning the human genome to eliminate aging, which may happen, those problems will remain hard.
There hasn't been much visible progress in aircraft in a while. Boeing is still building the 737, which first flew in 1967. Aircraft are lighter, quieter, more fuel efficient and safer than they were back then. But not cheaper.
The biggest societal problem is that we have less and less of a role for a big part of the population. The average IQ is only 100, remember.
Should we artificially limit population growth? Can it be done in a way that isn't forceful, but perhaps with the right incentives?
Should humanity, as a global society push for quality over quantity in reproduction? Is basic income still feasible in this scenario?
It's an ethical and ideological mess and frankly, it'll require someone way smarter than I to come up with the answers.
Two big problems we expected to have were overpopulation and running out of energy. Neither happened.
[1] http://www.eastasiaforum.org/2013/04/05/indias-population-in...
https://www.google.com/publicdata/explore?ds=d5bncppjof8f9_&...
For some of Europe the population is declining and in some places it is increasing - when added together the two pretty much balance out:
https://en.wikipedia.org/wiki/List_of_European_countries_by_...
Of course that data ignores recent immigration trends!
[0] http://ec.europa.eu/eurostat/tgm/table.do?tab=table&init=1&l...
[1] http://www.economist.com/blogs/graphicdetail/2016/01/daily-c...
There are three claims made here, and only one is correct. Europe has a very small net positive population growth (its population, unlike Japan's, is not in decline), and the US has a small net positive rate of natural growth (and so, it would not be shrinking even without immigration.)
But that requires some modulation.
1. The cornucopian arguments that we'd have no problems with food nor energy, or that these would grow without limits, haven't panned. out.
2. Yes, the specific predictions that we would be facing major global starvation by now (or at equivalent populations -- ~8 billion), haven't panned out. But those have been mitigated by a few circumstances.
a). The Green Revolution. Largely unanticipated, but through mechanisation, new plant breeds, fertilisers, and pesticides, per-acre and per-farmer productivity has been increased. This represents, though, another potentially one-time-only transfer, as we've basically amped up the nutrient cycle for plants (using copious amounts of natural gas to produce nitrogen fertiliser), taken over for plant functions in pest and disease resistance (through pesticides), and replaced human and animal power through fossil-fuel powered tractors and other equipment (so long the fuel holds out). Withdraw fertiliser and pesticides, and many of those benefits disappear (or must be replaced with far more labour input). Moreover, topsoil depletion itself is now a major concern.
b). Population growth, accellerating at unprecedented rates in the 1950s and 1960s, levelled off somewhat. This has been fortunate. However rather than growing at a constant percentage (exponential growth), it's been growing at a constant rate (linear growth). Trends such as the demographic transition may be helpful, but they're not assured.
c). The 1970s oil shocks. Prior to the OPEC oil embargo (1973) and Iranian oil embargo (1979), growth in global oil consumption was increasing at an increasing rate. The so-called oil shocks had global economic impacts, and sharply slowed growth in energy consumption. This delayed global peak oil (occurring in 2005 rather than 2000), but also put to lie cornucopian estimates of future energy consumption.
One argument might be that "the market responded" to higher oil prices. I read the response somewhat differently: the system responded to higher energy costs and greater energy volatility. The long period of exceptionally steady growth from 1945 - 1972 was, well, exceptional. Trends since 1973 have been highly erratic, though "better" (in terms of growth) at times, worse at others.
What's increasingly clear to me is that explanations of conventional economics are substantively incorrect. That energy and inputs do matter, hugely. And that "technology" almost certainly lags rather than leads energy availability, though the relationship is complex.
Can you expand on this more?
There is the claim of energy decoupling from GDP a few decades ago if I am not mistaken.
Should humanity, as a global society push for quality over quantity in reproduction? Is basic income still feasible in this scenario?
According to whom and on what grounds? The population is a resource. Having more people is having more human potential to throw at problems. There is enough information out there with plenty of arguments on population sustainability and our planet is nowhere close to its limit!
This in some ways even understates the problem e.g. Fish stocks.
There are different kinds of limits other than say, the Soylent Green limit. One is whether our planet is cooler with wild animals or having a few percent more cows.
Cows are cool but not as much as that!
We can have billions more people on Earth but in order for that to happen we need radical, radical technology changes. The most obvious step forward is Vat Meat. No animals required.
You do realize that's a tautology, right? There's no reason productivity couldn't be increased, and one potential method would be to boost human intelligence.
It's the average emotional intelligence of the human race in a given area combined with the average EI of another areas. Culture/areas/whatever that encourage a high amount of cognitive dissonance will give way to a "contagion" that gives rise to lower EIs through double binding communications between groups. It's better if you start out with a topic that is highly polarized, as that gives the contagion some energy to feed upon.
You see this with Trump, as he is in constant disagreement with himself in what he says, which is dove tailed in behind whatever blaming remark he made when he first started talking. This gets the dissonance going in groups who are, in general, disgruntled about whatever it is they are disgruntled about.
Sigh.
You could at increasing IQ in very poor countries where lack of basic micronutrients is causing IQ issues.
But I think better group intelligence, using current intelligence in poor countries better, AI, productivity increase with better tools currently makes much more sense.
http://www.vice.com/read/chinas-taking-over-the-world-with-a...
Selective breeding in a sense of "selecting the zygote" for ideal outcome. Aka designer babies. This was referred to as "selective breeding" in the press when it came out. A misnomer in that selective breeding has meant selective pairing of parents for at least a century before this article was written.
Note: this is also a company in China, not "China" although the distinction is sometimes muddled they certainly aren't regulating it like it has been in the West. Also, with the only one child policy that was recently lifted it is easy to confuse oppressive breeding requirements in China.
1. working with computers
2. selling technology
3. thinking about business in the medium term
4. calling ourselves very smart with little backing evidence for that claim
and besides that stuff I am not particularly impressed by the level of intelligence displayed here (including my own)
HN is highly correlated with technical professions and especially programmers. Those things happen to be highly correlated with and selective of IQ.
> In truth the IQ is a purely statistical measure. It has no direct relation to brain performance, is not proportional to it, and doesn't even have any linear or otherwise straightforward relation to it. The only thing you can say is that somebody with a higher IQ will show higher scores on most other brain performance tests as well, but the IQ doesn't say how much higher.
Except (and this counters GP post also):
https://en.wikipedia.org/wiki/Flynn_effect
Our IQs are steadily increasing. Or so it appears ...
With equivalent tests to those administered a few years ago scores are steadily increasing. It's not necessarily that we're getting smarter of course - far more likely to be the case that these tests become the "definition" of intelligence and our children are more specifically trained for them.
That's not what I meant. I meant that "scores" were increasing - which is the reason for normalisation against present population.
https://en.wikipedia.org/wiki/Intelligence_quotient#Flynn_ef...
IQ test scores have been rising at an average rate of around three IQ points per decade
So IQ in absolute terms has been increasing steadily - but this doesn't necessarily tell us anything about whether intelligence has been increasing or not.
Our parents invented the Internet too ...
Which ironically is what our parents called the "Application layer" when they were designing it ;)
https://en.wikipedia.org/wiki/Internet_protocol_suite#Compar...
All that happened from about 1945 until 1970. Then it stopped. I find it impossible to believe that we learned everything there is in 25 years. Something is broken.
What's "broken" is that it takes more than 25 years to work out the practical uses of a lot of these discoveries. We see much more of the total fruit from a discovery in 1950 than we do for an equivalent discover in 1990.
Things have definitely evolved a lot, but as far as "social media" (reminds me how much I hated that term when they first started using it, when the basic profile + messaging concept had been around over a decade) goes, it's more like comparing a Tesla to a Model T, rather than a whole new type of technology.
If one is more interested in discussion of the effect itself, and whether the IQ score changes reflect actual changes in intelligence between generations, Flynn has a longer and more academic discussion on the subject here: http://www.psychometrics.cam.ac.uk/about-us/directory/beyond...
Anyway just because it wasn't designed to have a linear meaning doesn't mean it's useless. It measures what percent of the population scored better and worse than you. That's still a useful fact to know, just like people talk about "the top 1%" instead of specifying exact income brackets.
He didn't say that it was
Naturally society polarize intelligence instead of spreading it so we classify instead of nurturing.
Meanwhile there is in fact an easy path to eliminating aging. Aging is damage accumulation at root. The easy path is to periodically repair that damage - the damage is already identified, after all. Sufficiently good maintenance is the road to eliminating aging; let it happen but control it. Maintain the biochemistry we have and that we know works.
http://www.sens.org/research/introduction-to-sens-research
This has a large weight of evidence behind it, and approaches are already being demonstrated in the case of senescent cell clearance, removal of transthyretin and beta amyloids, and averting the consequences of mitochondrial DNA damage. Others will follow as results become more compelling.
For my part, I'm astonished by the progress. I could go on and on, listing just my own medical history. We live in the age of miracles and it'll only get more amazing.
https://www.chrisstucchio.com/blog/2016/robots_didnt_take_ou...
It's not about "what" is being worked on, but how. I think he 's referring to the lack of major centrally-planned projects like the US space program, or the LHC. I think its true that the way scientific funds are being allocated favors short term but unambitious goals. As an example, both the BRAIN initiative and the human brain project have set rather vague goals that point to incremental steps rather than giant leaps.
What could you want for the brain? Mind-machine interfaces (MMI)? We already have those [0]. We have done the 'moon-shots' in neuro already, and in a lot of bio in general too. But you want your own MMI or your own miracle of medicine? Well, I want a vacation to the moon too, but that ain't happening.
The moon-shots in medicine have already happened, they just aren't cheap yet.
Some answers to the question:
> What isn't being worked on?
In the West there is no real effort going into nuclear R&D or, you know, constructing actual power plants. I think in the US the last nuclear plant was two decades ago. In the UK I don't think they teach nuclear engineering anymore. The Chinese are now to build and maintain their systems! Germany... France is an exception I think.
There are no new types of infrastructure e.g. no personal transit systems, no waste disposal pneumatics. We sure do have a lot of broken roads nowadays though. Road technology hasn't had a game changer in thousands of years (somebody will pounce on this but you know what I mean, it's not like we're expecting Blade Runner), it's still horrendously expensive and difficult as ever. Did I mention there's ugly tangles of black cables hanging above our heads? Maglev exists and Japan is giving it away for free but there appears no will to use it.
With the recent exception of SpaceX, aerospace hasn't really taken off for decades. We were supposed to have Moon Hotels and political tension with moon miners by now. The kind of scale you saw in 2001: A Space Odyssey was seriously expected to occur but instead the whole effort was on the verge of dying out until very recently. In three years it will be half a century since the moon landings. That is quite a while. The quick fire explanations won't do.
There has been no major tangible medicinal advantage to a large number of people from the discovery of DNA let alone the Human Genome Project. I know this one makes me sound bombastic, but come on, tell me how I'm wrong. I was at the doctor the other day and asked him whether having my DNA record would help him out. He said not really and that it's still better to examine my family medical history. Lots of press coverage isn't progress.
In materials technology good insulation is still very expensive. You can get great insulation but only for great money. If aerogels came down in price by 10x that would really lift the bar.
Energy Storage is just... A century with no major breakthroughs.
Biotech is a one horse town called Monsanto. GM is banned in many countries.
Animals. They are great, but they are not becoming more interesting. Back in the day people created all manner of breeds of cats and dogs. I don't see breeders or gene splicers coming up with freaky or cool pets for us. Really more people ought to be outraged. Remember micropigs? It seems they might not be real.
Freeman Dyson's vision is neglected so far. The last thing is people. This is more subjective but I believe people have become immeasurably more boring. Look at old photos of people. They had character. They fought eels, took trips to jungles. Everybody was Ron Swanson and it was fucking awesome.
The statistics for the short - medium term belie this intuitive idea.
There is job growth at the top and bottom of the wage spectrum since the crash.
What there isn't, is middle class jobs.
Perhaps you are right over centuries, but by then it won't matter.
He makes a distinction of computers as the "one field that has progres", but, is it really true that IT is not stuck in a rut as well? What are the new paradigms developed recently? Computer vision, machine learning were established in the 80s , despite only becoming commonplace nowadays. There is little progress in theory since then. Engineers obsess over the rediscovery of functional languages, a relic of the 50s.
Commercial success should not be taken as an indicator of real progress.
For example, LISPs did not run well on 80s computers. The Japanese, amongst others, had to create special computers to control their trains. Rules based programming performed poorly too. May other such paradigms and tools fall into this category.
For the most part the old, neat tools of now have the hardware necessary for them to be useful. The JVM provides great performance for Clojure. Over the last 30 years, multi-threaded concepts coalesced too. Now they're built into mainstream languages. Provided people stop trying to solve already solved problems like task management, we could possibly be on the verge of productivity improvements.
Computer science is in a rut. It's just that other things have even lost the energy to cycle through old accomplishments. I'd say the Art world is one of those areas that has lost all vitality.
This is what I think about whenever I hear folks scaremongering about the "singularity". In my opinion these beliefs reflect a kind of spiritual emptiness; a yearning for answers that only God can fulfill.
But you can't believe that this evolution will continue?
Perhaps it's exactly because their argument is empty of spiritualism that they see the world more clearly.
On the other hand, I find it pretty easily believable that we can receive compounding returns from advances in ML/AI. After all, pretty much as soon as we invent new techniques in computational statistics, the go-to fields needing to apply them are the other computational sciences: physics, chemistry, biology, medicine.
Unfortunately that book probably wouldn't be a good use of his time but come on Peter, I know you want to write it :)
The concept that zero to one is much more valuable than one to ten helped me understand my own career better. I have always been a risk taker and even though some of my work projects over the last 30 years have failed, my bosses never seemed unhappy with me. I would usually really nail a few new ideas a year. This book helped me to understand why I seemed to get perks (like ocean view window offices at two consequative companies) that none of the other programmers got, even though they seemed to be much steadier workers than I was.
It's not the pragmatic business POV, but it has some highly interesting considerations.
Edit: actually don't use the YouTube transcripts, there's a human-curated transcript here http://conversationswithbillkristol.org/transcript/peter-thi...
Theil speaks in run-on sentences, so it's not entirely the transcription system's fault.
How very meta.
Wait until procrastination, anxiety, depression and fatigue can be turned off by a device behind your ear that controls your amygdala and instantly puts you in the zone.
See see how much productivity goes up when everyone has the energy and motivation of the Richard Bransons of the world.
I can't get funding.
He talks big but he doesn't invest in the research. He still wants some business plan dropped in front of him. It doesn't work that way until the hard science is done first.
Transcript:
http://conversationswithbillkristol.org/transcript/peter-thi...
Another interpretation of this interview: P.T. for POTUS.
Fair call.