"Life after Google", by George Gilder, is one of the more interesting books I have purchased by-sight from a bookstore recently. In it, Gilder speculates about how things will change as we move from a world dominated by companies like Google (with huge centralized databases and machine learning), to a more decentralized world (with solutions like blockchain and, I think he would also say, more of an emphasis on real human creativity). It is therefore also very appropriate that I should post this on Steemit, a platform briefly mentioned in the book. But it was also a book I enjoyed at least as much for the little stories and factoids as I did the big narrative, so I thought I would share some of those with commentary - though, I must also caution, the little factoids mean more in the big narrative, which you will only get by reading the book.
1. The uniqueness of human consciousness.
The claim of "superhuman performance" seemed rather overwrought to me. Outperforming unaided human beings is what machines - from a 3D printer to a plow - are supposed to do. Otherwise we wouldn't build them.
I just thought that was a great point, obvious yet almost never said - if a machine wouldn't outperform humans in some way, then we humans wouldn't bother to build the thing. It was actually a recurring theme in the book that humans have a consciousness and a creativity that even the most advanced deterministic computer doesn't come close to - Gilder is definitely pro-human-uniqueness. I won't quote it, but you'll even find a quick sort of argument for the existence of God stuffed into the middle of the book on similar grounds, not exactly what expected to encounter in a book about Google.
The blind spot of AI is that consciousness does not emerge from thought; it is the source of it. As Leibnitz, imagining a computer blown up to the size of a building, observed in the seventeenth century, inside the machine (the determinist scheme), you find cogs and gears but no cognition. The oracle programmer must be outside. How a software programmer can miss the essence of his own trade is a mystery...
The materialist superstition is a strange growth in an age of information.
I'm sure some AI researchers would disagree, but an interesting thought. But what about the inefficient human brain, defeated by supercomputers in Chess and on Jeopardy? Maybe not so inefficient:
When a supercomputer defeats a man in a game of chess or Go, the man is using maybe fourteen watts of power, while the computer and its networks are tapping into the gigawatt clouds on the Columbia.
You might well object "sure, but not all of those gigawatts are being used for the game", but even the computer I'm just typing this on is using around 100 W. The human brain is still a marvel.
2. The 1517 Fund and universities
I just found this kind of charming. The 1517 Fund is a fund created in part to pluck a small number of technologically-minded geniuses out of college (or get them to skip it entirely), with funding and support to work on their own unique project. OK but... 1517? Is that just a coincidence? Nope.
The fund's name alludes to another historic decentralization, launched on October 31, 1517. That was the day that Martin Luther posted his Ninety-five Theses on the church door at Wittenberg. Among the abuses Luther was protesting was the selling of indulgences. The remission of the temporal penalty due to sins, an indulgence, like other spiritual goods, must not be sold. The pardoners who perpetrated this abuse would issue a document memorializing the transaction. The 1517 Fund explains the parallel: "Likewise, universities today are selling a piece of paper at great cost and telling people that buying it is the only way they can save their souls. Universities call it a diploma, and they're making a fortune doing it. Call us heretical if you like, but the 1517 Fund is dedicated to dispelling that illusion."
The Fund also wrote some of its own theses:
Why are there some 5,300 universities and colleges in the U.S. but only one point of view?
The more PhDs we mint, the fewer scientific revolutions we seem to have. There are more scientists working today than in any time in human history. It could be that science is harder, or it could be they're not really scientists.
Probably truth to both, by the way - oy. I will defend the fact that academia is full of great professors all day long, and a truly liberal education is a good thing. It is also true that academia will be the last people to admit academia has a problem.
3. Net neutrality
Always appreciate writers willing to challenge the accepted political dogma. Here he argues net neutrality is dumb because certain types of traffic need to be prioritized, and because,
In practice, the only factor that makes any serious difference for Internet neutrality is investment in bandwidth. If bandwidth is scarce, it will have to be allocated preferentially, regardless of the laws. Network neutrality is out the window. If bandwidth is abundant, neutrality laws are unneeded. Everything finds a conduit.
You could make a very similar argument about healthcare, actually. What matters most is how many doctors and such are available. If there are too few, healthcare will be rationed - by cost, by government fiat, somehow. The thing to do is not to fight over the type of rationing but to try to increase the "bandwidth".
4. Physics and Information Theory. I teach Physics - there was actually a lot of discussion in this book about Physics as applied to information theory, and even an extensive section about Newton and the importance of his banking / gold standard work for the world economic system, something even physicists who like Newton tend to ignore completely. On the information theory front...
Gödel's work, and Turing's, led to Gregory Chaitin's concept of algorithmic complexity theory. This important breakthrough tested the 'complexity' of a message by the length of the computer program needed to generate it. Chaitin proved that physical laws alone, for example, could not explain chemistry or biology, because the laws of physics contain drastically less information than do chemical or biological phenomena. The universe is a hierarchy of information tiers, a universal 'stack,' governed from the top down.
Not sure I quite buy that (probably depends on exactly what is meant by "could not explain"), but an interesting thought. Much later he says,
Knowledge of every quark and electron in a computer tells you virtually nothing of what the computer is doing. To know that, you have to address the source code, and the source code is the ground state where human interpretation is required.
I won't quote it, but he also has a nice section talking about Fourier transforms, which we physicists are used to using to move from the time domain of a signal (a wave for example) to the frequency domain. He applies it to finance, moving from the time domain of the record of trades (stocks, for example), to the frequency domain of the trading pattern.
5. Statistics about advertisements. I feel like I never, ever, ever click on an advertisement online. I guess somebody out there must be. But... apparently not too many people:
Meanwhile [over the past 70 years], ads delivered per person have remained stable at around 350 per day. Ads viewed per hour of media use, including print media, have dropped by half.
And later,
As James notes with some relish, only 0.06 percent of smartphone ads were clicked through. Since more than 50 percent of clicks were by mistake, according to the surveys, the intentional response rate was 0.03 percent.
I will use this as excuse to shout-out to the Brave web browser, created by Brendan Eich (yes, the Firefox guy later fired when it was discovered he had made a small donation in favor of a California "traditional marriage" initiative, which apparently cancels out every good thing you have ever done). It blocks most ads and is also trying to up-end the whole way ads are delivered online through their Basic Attention Token. It is also discussed in the book!
6. Facts about giant data centers and computational advance.
I should say, another recurring theme in the book is that when all other resources become plentiful, time itself still remains scarce. It is the ultimate limited resource. Google can give away its products for free, but they must absolutely remain fast to actually use. Bitcoin has coded right into itself the fact that producing more bitcoins requires time (and computational power, but the ultimate limit is time). On that note, speaking of Google's big data centers,
These servers, jammed as close together as possible to minimize speed-of-light delays, ...
I just found that interesting. Other facts,
Today, Google reigns over a database of thousands of petabytes, called exa-bytes, swelled every twenty-four hours by scores of terabytes of Gmails, Facebook pages, presidential twitter feeds, and videos - a relentless march of daily deltas, each larger than the whole Web of a decade ago.
Really brings home the size of the internet today. How omni-present are Google and Facebook?
As 70 percent of all links came to be handled through Google and Facebook, Berners-Lee feared that his Web was dying.
4. Facts of Physics.
Another recurring theme in the book is the actual physical limits we are hitting with our processors and centralized databases. One of the ways we have tried to respond is an increase in parallel processing / the use of graphics cards for processing instead of CPUs. Another has been the development of better hardware very specific in use instead of just trying to write better code for general hardware. Quoting actually Craig Mundie of Microsoft,
We have now run into a brick wall. What brought all of us faster computing was raising the CPU's clock rate. A faster clock increased power consumption. We could only increase the clock rate without consuming more power because we could lower the voltage. But we can't do that anymore because we're down into electron volts where quantum uncertainty takes over. If you can't lower the voltage, you can't raise the clock rate without using a lot more power.
One more - speaking of Austin Russell's attempt to develop better systems (using specialized hardware) to enable self-driving cars to see the world around them,
A crucial decision was to reject a wavelength of light preferred by his rivals - 905 nanometers - which is close to the visible spectrum and can simulate what can be seen by human eyes. Unfortunately, the 905 nanometer near-infrared also can affect human eyes. So Russell moved deeper into the infrared band to 1,550 nanometers (153 terahertz), which is employed in fiber optics. The longer wavelength can be used at sixtyfold greater power levels without endangering human beings.
I think I'll stop there but - definitely an interesting book, and recommended for people who are, perhaps, eager for "Life after Google" and what that might look like, with plenty of interesting physical and philosophical tidbits inside.