Comment by fidotron
4 months ago
There's something odd in this argument. If you come at it from a Canadian perspective Canada seriously spent on neural network computer science when almost no one else did (many in AI considered the entire thing discredited and impossible), now the (financial) gains from that are almost entirely in a foreign country.
The US science establishment was all about buying and utilizing Russian rocket engines until he-that-shall-not-be-named came along. SpaceX took the breakthroughs that existed in the US in things like control theory, which the same science establishment had failed to value appropriately.
It doesn't look like the science establishments of any country are actually successfully feeding their innovation machines, or have done so for decades. Switching a non functioning system off does at least allow it to be replaced by something that risks doing things when something comes back.
Of course many pure scientists will, legitimately, argue that innovation isn't the point in the first place, and that is a far more solid point, but real academic diversity has been so destroyed by the global consensus making peer review process that much of their progress has effectively stalled.
Not only did SpaceX make breakthroughs considered impossible by the "experts" in the industry, they did it by hiring a guy who literally built rocket engines in his garage to design the engines. The key here is personality. And the type of person who actually wants to build things and get things done absolutely recoils at bureaucracy and the type of people who like it.
When you build something to the point where there is a bureaucratic "establishment" in control you can be sure that innovation slows to a crawl. You may still have a few individual scientists doing great work, but you can be sure that some miserable bureaucrat will pat him on the back and stick it in a drawer somewhere never to see the light of day again. The same is true whether that bureaucratic establishment is at a government or in universities, or any other type of bureaucratic organization.
"Building things" is not science, it's engineering. We could certainly compare the outcomes of "bureaucratic" science against the free market variety, but there's basically no free market science going on to support such a comparison.
This isn't a value judgement. Engineering is just as important as science, but just as more science is not a replacement for engineering, neither does better engineering free us from the need to keep pursuing science. And at the end of the day, SpaceX might be an impressive engineering company, but we still need the scientists. And it's weird how often the success of SpaceX is brought up as an implicit argument that we can send all the scientists to work on farms or whatever without any ill effects.
It also seems notable that a company like SpaceX is an obvious candidate to bring back the 20th century style corporate funded scientific research organizations to underpin their engineering efforts in a way that would presumably be free of the hated "bureaucracy". But if they've done so, I haven't heard about it.
It's not like they were able to use NASA's designs for reusable rockets...
Oh wait..they did...
Because NASA thought reusable rockets were possible decades ago. The reason they never built them was because certain Congressmen blocked the funding.
Any American company could use NASA's designs for reusable rockets... Oh wait, only SpaceX did it. And 10 years after they did it, they're still the only company that did it (not counting New Shepard of course, as it can't put anything into orbit). Are some congressmen forbidding other companies to use their own money to make better rockets?
5 replies →
"Seriously spent" where serious is less than the cost of a single bomber for the military. I forget what Geoffrey Hinton said it was, but it was an embarrassingly small pittance.
Military spending is largely economic dead weight, roughly the equivalent of handouts. And the end result is deterrence in a game of prisoners dilemma. Yet it is sacrosanct, and subject to ever increasing budgets for no gain.
The % of US GDP spent on defense has been going straight down ever since Eisenhower gave that speech about the MIC.
(Another fun fact is defense companies make fewer profits during war, not more. Presumably because they have to make real products instead of designing imaginary ones.)
Just look at all the handouts going to Ukrainian soldiers right now. What silly economic dead weight!
The appropriated funds were only 4% of the defense budget, a lot of which still hasn't been handed out, and half of the handouts weren't even weapons and other military supplies. You could fund 10 Ukrainian wars at the same level and still have an extra 150% idle capacity.
military spending being dead weight is the most ridiculous statement I've heard in a long time. Do you know nothing about US, British, and Dutch history?
The main way the US, UK, and the Netherlands historically became rich was through maritime trade. Maritime trade was basically only possible due to those countries' military expenditures on having strong navies. I know the media makes a big focus on US special forces and other things, but the US Navy is basically the most important and foundational part of all of the US military power. The US and Allies were always interested in maintaining freedom of navigation and trade at the seas.
Just take a look at how much money was lost due to trade shipping costs due to the Houthis in Yemen. Consider that today it's cheaper than ever to ship things, and even today, it was so terrible. Shipping by land is terrible. The only historically economically feasible way to do maritime trade has been with Navies to provide protection from pirates.
I’m blind, and participate in a lot of research projects to create accessible technology, which are mostly done by universities. What I have noticed as a foreigner participating with US based universities is that, a lot of this research while very high-quality and very well done does not actually result in anything that the intended audience gets to use or experience. And a lot of this is due to the amount of red tape, as well as a lack of risk taking. This means that without trying to go commercial a lot of these projects end up shelved and many potential users simply never see the benefits.
I think the article does not explain this well, but (having worked in an applied science field for many years) much of the work is picked up from these failed projects by other members of the field. Sometimes this is quick, sometimes years later. Anyway, while it might not immediately evident, shelved projects often do move the state of the art forward and unlock someone else's success (hence part of the argument for it being a public funded system). Of course, a lot do not, but that is the nature of research.
Because talent and ideas move so easily between the US and Canada, any useful basic science that Canada comes up with will ultimately be monetized in the country with 10x the population, 15x the GDP, and 100x the stock market and VC funding depth.
This could start to change if present US hostility towards all things foreign results in a shift in investment and migration.
Talent and ideas move so easily to the US from Canada*
I don't really understand how you come to the conclusion that the current system is non-functioning. There are tons of examples of it working (I even have members of my family who took their research to make products) and for the rocket example experts didn't think it was literally impossible, just not cost effective with the current technology (e.g. the space shuttle was reusable but very expensive). I don't think that's some huge failure given SpaceX had a lot of launch failures and went nearly bankrupt before the first launch (and even after that it was a long road to profitability, I thought it took 20 years to get to an operating profit). This is also ignoring how SpaceX is operating within the current system also.
It's not perfect but if you replace the system you're gonna find the same sorts or errors since it's impossible to accurately guess future value of uncertain engineering projects.
That's purely an issue of living next to the USA and Canada's system not rewarding risktakers/startups like the USA does. If you have the next billion dollar idea would you rather get rich in Canada or the USA?
I'd rather get rich in Canada, but I will still find it much easier to fund my pursuit of a billion dollar idea in the USA.
Research is necessary but not sufficient. Also need access to capital (and eventually capital markets) and a sufficiently sophisticated legal framework/safety framework so you can enforce contracts at least most the time. Good research is just a vehicle for producing knowledge and talent.
> Good research is just a vehicle for producing knowledge and talent.
Yup, it was surprise seeing that once have a good STEM field Ph.D., written a lot of STEM and (early) AI software, and have published some peer-reviewed papers, are then condemned as in a felony conviction from ever having money enough to buy a house or to participate in the real world. A surprise.
But so far have missed the law that actually forbids doing some good math research, writing corresponding software, starting an LLC business as a sole proprietor, self-hosting a related Web site, getting users, running ads, and making (oops, forgive the transgression of mentioning) MONEY. Horrors! Uh, that's actually the same kind of "money" that people selling food, clothing, cars, and houses and providing medical care, private K-college education talk about.
Gee, cut out a lot of middle stuff, i.e., save on management, lawyers, office space, recruiting, HR legal issues, insurance, utilities, software developers, cloud fees, server farm staff, servers (an AMD FX 8350 can send a lot of the Web pages with still image banner ads), ....
Just now evaluating Macrium Reflect. Anyone have any experience? Uh, will it copy a disk partition that has a bootable instance of Windows so that the target disk partition will also be bootable or will it copy only whole hard disks with all the partitions, ....???? Similarly for Acronis and Windows Image. E.g., if make an Image of a bootable partition, will the partition written to boot? Right, just use the TIFO method!
Canada is also reaping outsized rewards for that investment. There are plenty of AI jobs in Canada that never would have existed.
The problem is that in Canada we're willing to invest a little for a long time. But we're not willing to make big bets.
You can raise far more capital in the US than in Canada. So naturally large rewards come to those who are willing to make large bets.
> The problem is that in Canada we're willing to invest a little for a long time. But we're not willing to make big bets.
Element AI.
It's not that the bets are not made, the winnings are captured by those that corrupt the existing system before it's even started, so tney go nowhere. This has been the status quo for so long everyone assumes it is what is happening with every new initiative.
> The US science establishment was all about buying and utilizing Russian rocket engines until he-that-shall-not-be-named came along. SpaceX took the breakthroughs that existed in the US in things like control theory, which the same science establishment had failed to value appropriately.
I feel like you're confusing "science" and "engineering". SpaceX is fundamentally an engineering company, not a scientific one. Don't get me wrong, they've done impressive work in engineering innovation, but that's fundamentally different from scientific research. And as the article points out, engineering innovation from the likes of SpaceX is usually reliant on that foundational scientific innovation, which in turn is essentially useless without an engineering partner to realize scientific discoveries.
> It doesn't look like the science establishments of any country are actually successfully feeding their innovation machines, or have done so for decades.
Really? Is it just a coincidence that up until recently the US had some of the most robust scientific funding and was an unbeatable source of engineering innovation? For that matter, are there any real counterexamples where science research is non-existent but engineering excellence abounds?
How's China doing? They seem to have a lot of research going on that feeds into their manufacturing fairly quickly from the papers I hear about
Notably China is a big country and Canada is a small country. If there is some innovation that is going to improve productivity globally by %X the amount of benefit that goes to China is always going to be bigger than the benefit that goes to Canada.
China are certainly better at turning the results of research into products, whether that research was them or anyone else.
The canonical example here is 5G. Once again the US science establishment had the guy, he ends up doing the breakthroughs for polar coding, they failed to appreciate him, he left and ended up being funded by Huawei.
https://en.wikipedia.org/wiki/Erdal_Ar%C4%B1kan
The US science establishment isn't broken as an innovation engine because of Trump - it's because they're clearly rewarding the wrong things.
What isn't so clear is if Chinese science is creating Chinese startups. It may yet happen.
Eh, China is better at directing massive state level resources at incrementally improving technology. Nothing truly revolutionary has come out of China. The West is still ahead in that sort of stuff.
3 replies →
Its not that spacex did anything novel really. People act like the shuttle died when it has been continually developed to this day. The shuttle always had a dual use as a commercial and military technology. The commercial activity was not economically viable. The military activity was and remains an important capability. That is why the shuttle still flies. Only it is called the x37b and does not require the complexity of sending humans into space.
"real academic diversity" is doing all your lifting here
There’s not enough information to determine what the phrase is supposed to mean in context.
They seem to be opposed to peer review?
There are successes in Canada, surely Google’s presence there counts for something. It’s an huge multinational corporation, some big AI units are in London and Toronto. Among startups in AI, Ideogram. Public funding in the arts in Canada is also really good, and it’s not an accident Ideogram is there and not here.
50% of VC is biotech, HistoSonics is on the front page of HN, and it was a PhD project turned into a huge company with deep, liquid capital markets. BBC doesn’t really write about that but the venture system that made HistoSonics was super sophisticated, full of very bright people willing to take huge risks, and while the inventor isn’t going to be a billionaire, she is still there inventing stuff, and she’s not going to be poor either.
Nobody is asking economists in the article any questions. How does R&D show up as something measurable? - thats really the meaning of financial we care about in this case, as opposed to economic or humanistic, where it’s pretty obvious that Canada, Russia, and the US, among other smaller countries like Israel and Finland, benefit from R&D more than financially, compared to say the UAE, which has spent relatively much more but yielded far less. A very attractive measure, to me, is the productivity of sectors that are basically indexes on a nation’s geography, like real estate, agriculture and minerals, versus innovations-driven sectors like healthcare, education and technology. In relative terms, given how great the weather is and how much oil there is, California and British Columbia (for EXAMPLE) swing way above their weight, no?
Instead of asking why Canada is doing this or that - Canada is doing quite well - we should be looking at South Korea.
> he-that-shall-not-be-named
Who are you talking about? It’s childish