Nothing can be so amusingly arrogant as a young man who has just discovered an old idea and thinks it is his own.\u00a0\u2013 Sidney J. Harris<\/p><\/blockquote>
Today, I’ll talk about Hyperloop.<\/p> While in Europe high speed train are usually connecting large cities which are far apart (basically, all major capitals are less than 3hrs away), they can be used for much shorter distance.<\/p>I went to Taiwan past June, and I was amazed by the Taiwanese High Speed Rail system.<\/p>\n Taiwan High Speed Rail is the BART of Taiwan<\/p>\n<\/div> The train from the Airport to Hsinchu (where the largest semiconductor company in the world is located) takes 13min to cover 18 miles<\/strong>, costs $4 (US) and leaves every 10min (see for yourself<\/a>; T$100 = US$3)<\/p>\n Taoyuan-Hsinchu THSR schedule<\/p>\n<\/div> That’s the same distance between Berkeley and Stanford, which takes 50 minutes to cover with no traffic (public transportation is more like 1h20 and utterly unpleasant.) Silicon Valley luminaries should take notice.<\/p>Oh, and it was a private investment; it costed US$18B<\/a>, and now reaches almost 90 per cent of Taiwan\u2019s population in just over an hour, comparable to the projected cost of the California high speed train.<\/p>\n The reason why seemingly new ideas such as Hyperloop gain some much traction is probably because there is an apparent slowdown science, and how it bears fruits. I recently got incensed by this piece:<\/p>\n The picture this survey paints is bleak: Over the past century, we\u2019ve vastly increased the time and money invested in science, but in scientists\u2019 own judgement, we\u2019re producing the most important breakthroughs at a near-constant rate. On a per-dollar or per-person basis, this suggests that science is becoming far less efficient. While eventually they advocate for measuring the throughput of science, as if there wasn’t enough of them, the metric they use here is completely arbitrary (“we surveyed 93 physicists from the world\u2019s top academic physics departments, and they judged<\/em> 1,370 pairs of discoveries. The bars in the figure below show the scores for each decade.<\/strong>“)<\/p>I strongly suggest the author to read Marianna Mazzucato’s “The Entrepreneurial State” and subsequent books, where the author explains that a fundamental flaw of research spending is that the value it creates is ill-defined and therefore unable to recapture the economic growth it contributes to create.<\/p>It is also possible that the authors take issue with the Y Combinator pipeline<\/a> (which they are associated with) \u2013 this I would agree: scanning through it, the only company that has bona fide<\/em> scientific credential is Synvivia<\/a>\u00a0 (whose founder Gabriel Lopez<\/span> turns out to be an acquaintance from my time with BPEP<\/a>)<\/p>It is true that recently, progress requires much more resources and is much less visible than it once was, or sometimes taken for granted. Take for example the invention of Erbium-doped fiber amplification (EDFA<\/a>) by Emmanuel Desurvire<\/a>. Very few of my readers might know about this arcane technology, but it is behind ALL modern days communication, as it allows the intercontinental transport of information. Another example is EUV lithography<\/a>, where now the industry is investing tens of billions of dollar to allow to make even smaller microchips, that tomorrow (a year, at most!) will be in the pockets of over a billion people.<\/p>One issue for sure is the fact that technology has been incredibly efficient that it basically take the air out of the room for any technology that would be more efficient. In the report of the National Academy of Science on Quantum Computing<\/a> (free to download, just log in), this is seen as the main threat:<\/p>\n It is hard to achieve this type of exponential scaling without such a virtuous cycle. This is apparent from the historical example of efforts to make transistors out of a material other than silicon. Finding<\/strong>: Moore\u2019s law for integrated circuits resulted from a virtuous cycle, where improved technology generated exponentially increasing revenue, enabling reinvestment in R&D and attracting new talent and industries to help innovate and scale the technology to the next level.<\/p><\/blockquote> Things are very hard, because we push science to the very limits of nature. We can detect gravitational waves by measuring deformation of the space time much, much smaller than the size of an proton<\/a>; we mass-produce transistors that are just over a hundred atoms wide; we send spaceship at the edge of the solar system<\/a>, we measure the finest interactions in facilities that are larger than some countries.<\/p>\n But we still have the impression that scientists are loonies with no sense of purpose…<\/p>\n![\"\"](\"http:\/\/antoine.wojdyla.fr\/blog\/wp-content\/uploads\/single-bend-chromatic.jpg\")
<\/a><\/p>edit 12\/18\/2018: Hyperloop startup Arrivo is shutting down<\/a> -The Verge — lol<\/em><\/p><\/p>This regain of interest for Hyperloop was brought to me while reading Fran\u00e7ois Lac\u00f4te’s amazing breakdown of the claims of hyperloop: “Hyperloop: formidable escroquerie technico-intellectuelle<\/a>“. Lac\u00f4te is the former CTO of Alstom, a major player in High Speed Transportation. He might be biased, but he makes a compelling demonstration. It is so good that I’ve made a translation: Hyperloop: a staggering intellectual and technical swindle<\/a><\/strong> (available here as a pdf)<\/p>\nHigh Speed Train for commute<\/h3>
<\/a><\/p>\n<\/a><\/p>\nA slowdown in technology?<\/h3>
\nScience Is Getting Less Bang for Its Buck<\/a> \u2013 Patrick Collison and Michael Nielsen, The Atlantic<\/p><\/blockquote>
\n[For example the] development put GaAs manufacturers in a Catch-22 situation: to fund their manufacturing R&D, they needed robust sales; to get robust sales, they needed state-of-the-art manufacturing techniques to compete against the silicon alternatives, which were constantly improving. The industry was never able to break this cycle, and the efforts to build commercially successful GaAs ICs ultimately failed; general-purpose digital GaAs ICs never became competitive.<\/p><\/blockquote>\n