Freeman Dyson's autobiography tells how in 01956 he joined General Atomics, a nuclear-reactor startup, and his small team designed and built a new kind of nuclear reactor in two years, the TRIGA; Teller was one of the team members. The prototype operated for 39 years, 33 of them were manufactured, some are still in operation, and none of them has ever had a nuclear accident.
A few years after that, when Project Orion was canceled, he stopped working on nuclear reactors. He said it just wasn't fun anymore. And I imagine he's right; since about that time, working on nuclear reactors enmeshes you in the national security state, where promotion depends on political favor as much as technical competence, and rank is measured by headcount and by project classification level. There's a certain kind of person that's fun for, but it's usually not the kind of person who spends his Christmas vacation working problems from a calculus textbook. It's not the kind of person who discovers a lot of new things. Atomic Energy merit badges went out of style.
About 20 years later, at the end of the 01970s, the costs of building new nuclear reactors started to skyrocket, which is another way of saying that our productivity at building nuclear reactors started to collapse.
I just read Atlas Shrugged, which tells a story of a technological society strangled by regulation, unpredictable kleptocracy, and a brain drain ("draining the brains", the book said, 6 years before the Royal Society coined the term "brain drain"). This aspect of the book resonated with me, although many others did not, perhaps in part because I live in Argentina, whose technological development is strangled by regulation, unpredictable kleptocracy, and a brain drain, though the brain drain is to the US and Europe rather than to a secret conspiracy. We had a famous fusion energy program, Project Huemul, which turned out to be a scam.
I wonder if the same thing happened to nuclear engineering? Maybe the potential Tellers and Freeman Dysons born in the 01960s and later never got involved in nuclear physics, except for David Hahn, who ended up with paranoid schizophrenia after his mother committed suicide, and died at 39 of an overdose of opiates. Maybe nuclear engineering had a brain drain to electrical engineering, computer science, and related fields.
It is really noticeable in talk on solar prices that the learning curve for nuclear tech is negative. Ie, nuclear plants get more expensive as technology improves.
The odds of that being a natural effect are minuscule. Something is strangling innovation and it is probably some piece of regulation. It is a bit grim but I suppose we're looking to places like India and China for the next big waves of innovation in nuclear energy.
Other cases where this seems to have happened include pacemakers; hearing aids; supersonic jets; ocean-liner trips; small planes (e.g., Cessnas); mercury; chemistry sets; most vacuum tubes; hovercraft; bubble memory (and more generally SSDs that don't die from read disturb); Gyrojet guns; brass vacuum fittings; ebonite; and old-growth lumber. If you limit yourself to just US production, you can apparently add things like microcontrollers; many kinds of machine tools; ventilators; Nixie tubes; and N95 respirators.
I don't think it's grim at all that India and China are having big waves of nuclear innovation, but I don't think they stand much of a chance against solar except for niche applications until there's a total manufacturing revolution.
Supersonic was killed by the dual shock of the oil crisis and the regulations banning sonic booms from overland flights. Which, to be fair, were pretty terrible, or at least were horribly managed from a PR point of view. https://en.wikipedia.org/wiki/Oklahoma_City_sonic_boom_tests
Concorde was really in the wrong place, at the wrong time, and with not enough of a big market. It didn't fly far enough for it to make a big enough dent in travel times, and it turns out passengers don't mind six vs three hours on a transatlantic hop if they get to spend it sleeping in a redeye in a cushy lie flat seat. If it had been capable of making the transpacific hop, I suspect it would have had a successor even with the sonic boom problem.
I highly recommend giving the book "Where is my flying car?" a read. It is a very idiosyncratic and unconventional book, written by an engineer rather than an academic. He goes into great depth about how the FAA and NRC basically deliberately killed general aviation and nuclear energy, how this happened, and its implications regarding the "great stagnation" and the future of tech.
I don't believe a dicky-bird from the UK govt until I see the evidence, but they have announced support for adding new nuclear capacity, with RR in the frame. Turns out that wattage isn't fungible, 24/7 is worth a premium, and land has value.
"It is a bit grim but I suppose we're looking to places like India and China for the next big waves of innovation in nuclear energy."
Or the next wave of chernobyls. Bangladesh is about to open their first reactor. Given how much else is not going well in that country, I would not want to live next to that thing. (and I think 1.5 million people do, who have no chance of getting evacuated in time, in case something big goes wrong).
So in general I am really not a fan of regulation - but when it comes to a energy source, where one major accident 35 years ago, and far from where I live, was enough to render the soil around here contaminated, so that collecting mushrooms and eating boar, is still advised against TODAY - then I do see the benefits of regulation.
Technology is a verb. Technologies are means to ends, so that technologies act on some precursor to produce a result. Or rather, sets of results.
All technologies include a combination of intended and unintended consequences, of greater or lesser apparance (or manifestation), occurring nor or far in time, and with greater or lesser ease in explaining or describing.
Positive consequences of technologies exhibit diminishing returns in most cases (networked / dendritic mechanisms are the exception). We see leaps where new specific techniques, materials, fuels, or processes are first applied, but then virtually always a decline in further growth. Even network-based technologies are subject to various impedences and at best grow more slowly than naive models (e.g., Metcalfe) would suggest.
And as technologies become established and embedded within processes, practices, institutions, and cultures, dependencies develop such that interactions are more complex.
Which is a long wind-up to this: in time, the early prompt-positive impacts of technolgies are dominated by both prompt- and delayed-negative impacts. Pollution, corrosion, health effects, disruption, cognitive overload, and fragility amongst them.
Early nuclear designs could focus on the simple and immediate. Longer-term projects had to take into consideration additional concerns: the effects of fuels, waste products, and radiation on materials themselves, natural disaster risks, the frailties of human organisations (a major factor in the four major commercial plant accidents, and countless minor ones), containment, economics, and more. Early visions of running unshielded reactors to power aircraft proved ... slightly less than viable.
Maybe it's not fun. Maybe it requires a more mature mind. Maybe the layers of bureaucracy and politics and checks and balances that evolve, even if suffering from their own dysfunctions, have to be there. And can only do so much.
Galt is an adolescent fantasy, not a model for the real world.
I don't believe in technology, just technological development. I don't believe technologies exist in an ontologically useful sense, much less that they have positive or negative consequences. Artifacts, knowledge, praxis, and social arrangements have consequences, not technologies.
Nothing is required except death; nothing but death has to be there. Everything else is contingent: it will happen or not, depending on historical circumstances. If we want nuclear reactors that aren't dangerous, we should explore what circumstances could give rise to that result. Our current path evidently isn't it, and maybe we should find out why not.
Galt is of course a fantasy; in the real world, the people Rand admired so much died in GULAG around the time she escaped Russia. Galt was a power fantasy about rescuing them. But some other things in the novel happen in the real world: people do build railroads, develop new metal alloys, and succumb to brain drains. And, interestingly, fields of knowledge do sometimes fall into decline. Maybe the reason in this case is a variant of the reason she fantasized about in the book; maybe not.
I think the fact that kleptocracy, brain-drain and mis-regulation are real problems are part of the reason for the (relative) popularity of objectivism.
You can't get people to buy into your solution until you get them to buy into the problem.
There's this trope going on in circles of people who love nuclear tech that expensive plants are just expensive because of X, where X can be environmentalists, government bureaucrats, old, bad designs (e.g. too large, or not using thorium, or not using molten salt, insert pet design issue here), lack of subsidy (e.g. government backed loans), etc.
While there's probably some truth to it, I think it's ignoring the elephant in the room, that nuclear tech is just really difficult to get right, and that the bar has been upped as the nuclear engineering community has learned about new failure modes.
Also, I think there's widespread agreement that current nuclear tech needs an oversight committee, just like aircraft tech. Even if it's incompatible with regulation nihilism.
Well, I think the cost for converting thermal energy into electricity is high enough to make thermal energy uncompetitive with PV in most of the world, regardless of its source. So nuclear is unlikely to make a comeback unless there's a revolution (heh) in making steam turbines.
I'm not sure why they're so much more expensive than coal plants, though, and I don't think "really difficult to get right" is an adequate reason; the first nuclear reactor was successfully built on a shoestring budget with technology that is now 79 years old. 79 years ago, lasers, jet airliners, artificial satellites, LEDs, lithium-ion batteries, pacemakers, antilock brakes, organ transplants, strong carbon fibers, high-pressure sodium lights, programmable computers, and bulletproof polymers were not just "really difficult to get right" but actually not yet achieved. Vast improvements have been made in many relevant technologies, including control systems, electronics, materials science and engineering, mathematical simulation, metrology, materials processing, heat engines, construction, and of course nuclear physics, not only since the Chicago Pile 1 but even since the end of the last wave of nuclear-power-plant construction in the US 40 years ago.
I think it's totally plausible that the reason is that the right people (both investors and hackers) fled the field at the thought of having an "oversight committee" with veto power over their work, no incentive to not veto it, and either no experience designing nuclear plants, or experience doing so as employees of their competitors. I don't think that amounts to "regulation nihilism".
I would have thought he is a crackpot of the Germanic variety (hyper “rational”) of Argentine extraction... but there is such a thing as Long Now Years
I am a big fan of Michael Polanyi's writings on tacit knowledge and I was disappointed to find that the author completely misuses the term.
Explicit knowledge can be codified and easily expressed. Tacit knowledge is not codified and not easily expressed. Given the details of story one (they went to a library and researched a variety of topics), that seems to clearly be referring to explicit knowledge. The second story is lacking crucial details about why it took a decade and $70m to replicate the foam, but we can assume that there was some reproducible process behind the original foam that was lost and required a great deal of effort to reproduce. Again, that would be explicit (not tacit) knowledge. The only way that it would not be is if there were individuals with a technique that was not codified and easily expressed that the original foam depended upon.
Why do you know those refer to the same individual or at least highly suspect they do?
Now you can begin to explain the rules and logic underpinning that understanding, but for every explanation you give, a counterexample can be shown where you'll go: Ah yes, but in this case... And you'll add more and more rules, yet you'll find that never do these rules and your own judgement are a 1:1 perfect match if you were to test them against yourself, you'll keep finding cases where you concluded differently.
Thus what is this knowledge that you truly leverage to make the decision? It seems that you cannot explain it fully, maybe your explanation isn't even anywhere the true knowledge you leverage, but only an approximation you made up when asked how? That knowledge is thus tacit, because you cannot express it fully to others, you won't be able to document it, or to produce a computer algorithm for it, or to communicate it to anyone else. It seems to become your expert intuition.
It turns out that Machine Learning has made some breakthrough in that realm, in that the computer can similarly learn some tacit understanding, though it may not be the same as yours, it gets closer than what your rules expressed, and it too cannot be explained back, why the computer comes to its decision is also tacit.
> why it took a decade and $70m to replicate the foam
In fact, if I'm thinking of the right foam here (codename FOGBANK, believed to be a lithium deuteride aerogel), the problem was even worse. There was a contaminant in one of the original raw materials (acetonitrile, if memory serves) that turned out to be essential to the whole process working. But nobody in the original program knew this contaminant was even present, much less its importance, so of course they didn't document it. Not explicit knowledge, not tacit knowledge, just an unfinished R&D job.
"Lawrence Radiation Laboratory hired three fresh PhDs with little nuclear physics background and asked them to go build a nuke, and they did."
No, they didn't. They designed a nuke. Not the same as actually building one. Designing a nuke is not at all difficult, and indeed the linked article describes how they decided to design a plutonium-based implosion device rather than a uranium-fueled gun-type device specifically in order to make the project more challenging than it otherwise would have been.
Whether the stories are good or true is debatable, but I thought this was a nice bit of insight in any case:
> Turns out it's really bloody hard to replicate an organisation. You can call this irreplicable part culture, because that's what it is. Which is a way of saying sure you can hire some folks with some data but that's not gonna help you build a self driving car. That's why the loss of a person hurts so much, because they take knowledge with them that's not only not written down anywhere but can't be written down. And yet, just because it hurts the poachee doesn't mean it will help the poacher. The exciting stuff is all culture, which is as numinous[1] and emergent and intangible as it is real.
Of course we know that losing a strong colleague or employee is a terrible experience, because they know more than they could ever share through documentation. The novel bit for me is thinking about that employee’s next employer. There’s no guarantee they can make the same outcomes happen again.
[1] “having a strong religious or spiritual quality; indicating or suggesting the presence of a divinity.” Context had me thinking this word meant “myriad” but I’m glad I looked it up.
I think the two “stories” were edited down a bit too much and are missing background information that isn’t technically relevant to the story but makes it difficult to follow because we have no context.
Once I read the linked stories in points 1. and 2. at the top the rest made a lot more sense. I don’t think it’s fair to assume readers will click on links and read the linked material in full though. Most readers don’t even finish reading the article they’re currently reading!
This is possibly me projecting my own biases into the author's writing, but the specific style of "some paragraphs made up of only a single long sentence, some paragraphs made up of a few small sentences" reminds me very much of the ADHD folks I know and their writing style.
Based on the blog post, and without knowledge of the underlying foam that they were trying to replicate, I would say that a crucial distinction here is that the first team was asked to create something new, which meant that many different inventions achieved by different paths were enough to get the job done.
The second team was asked to produce a much more precisely specified thing, which may only have been possible to make with one process.
So the two teams had different tasks, and the difference is not implicit vs process. It's many ways to many things, vs few ways to one thing. The latter problem sounds much harder to me.
>> If you really want to make important discoveries, you have to know how to work the system,” says Richard Muller, an American physicist and emeritus professor of physics at the University of California, Berkeley. Muller revealed that he had secretly redirected funding from approved projects to fund riskier rejected ones in a letter in Science magazine in 1980.
> It's worthwhile noticing that this propensity to leap ahead, self experiment and generally operate outside established strictures has always been a key feature of science!
Maybe we shouldn't glorify what sounds a lot like academic / grant fraud.
I know a retired chemist who said something similar. In order to get funded consistently, you have to be successful every time in getting the results you claimed you would on your previous grant application. The only way to be sure you can do that, is to sneak some work on the next grant into the current one. Granting agencies think they are being more efficient, but there's no way to avoid risk in research - they've just pushed it off on the researchers. Researchers need to have some career stability, so they found a way to hand the risk back to the agency, by adding the cost of trying the riskier experiments into the cost of the boring, sure-thing projects. Everybody gets what they want!
> It's worthwhile noticing that this propensity to leap ahead, self experiment and generally operate outside established strictures has always been a key feature of science
. . . and art, business, sport, entertainment, politics….
Flocks of seagulls and pigeons do it too. A small percentage feeds on the outside of the flock. Many go hungry. Some die. The few that strike it rich through serendipity prosper. Some form whole new flocks with their generations of offspring.
The two nuke stories aren't quite equivalent. The nuclear-naive postdocs designed a Nagasaki-style fission weapon. The "Fogbank" material that we forgot how to make is a component of fusion bombs, which are widely reported to be generally a whole lot trickier (you've got a small A-bomb which is supposed to light a much larger fusion bomb -- and not just blow the pieces of it all over creation.)
Also, wikipedia at least indicates that the problems of Fogbank included an unknown "contaminant" that was required for proper functioning. This is a common issue replicating older alloys as well; the exact impurities in the contemporary ores are not always known.
The difference between the two tasks is "any solution" vs "a specific fixed piece of a solution". It is kind of fundamental to the physics. The second is a highly specific bit for a solution.
I wouldn't be surprised if the foam was made far harder to reproduce by obstructionism of the actual warheads in question, too because of the pathological secrecy in governments for matters of far lesser stakes.
My understanding is that there were some impurities in the original sources that made the stuff work. Purer ingredients likely didn't react the same way.
The same behavior was noted when recreating the "spark of life" experiments from the 1950s in non-borosilicate containers. It turned out that the containers themselves played a role.
A similar situation occurs with sonoluminescence, where the trace ingredients of air have to be dissolved in the water to get it to work. (argon and co2)
Biohacking/biotech is an interesting example, since you can scale down to individual biohackers or up to farms (in which those farms very clearly have manufacturing/process that is hard to replicate).
However, biology also has the ability to adapt to its circumstances and replicate itself. Though that ability is usually annoying, it just might enable manufacturing/scale up that is easy to copy, something that is fundamentally different from all other kinds of high-tech manufacturing.
A few years after that, when Project Orion was canceled, he stopped working on nuclear reactors. He said it just wasn't fun anymore. And I imagine he's right; since about that time, working on nuclear reactors enmeshes you in the national security state, where promotion depends on political favor as much as technical competence, and rank is measured by headcount and by project classification level. There's a certain kind of person that's fun for, but it's usually not the kind of person who spends his Christmas vacation working problems from a calculus textbook. It's not the kind of person who discovers a lot of new things. Atomic Energy merit badges went out of style.
About 20 years later, at the end of the 01970s, the costs of building new nuclear reactors started to skyrocket, which is another way of saying that our productivity at building nuclear reactors started to collapse.
I just read Atlas Shrugged, which tells a story of a technological society strangled by regulation, unpredictable kleptocracy, and a brain drain ("draining the brains", the book said, 6 years before the Royal Society coined the term "brain drain"). This aspect of the book resonated with me, although many others did not, perhaps in part because I live in Argentina, whose technological development is strangled by regulation, unpredictable kleptocracy, and a brain drain, though the brain drain is to the US and Europe rather than to a secret conspiracy. We had a famous fusion energy program, Project Huemul, which turned out to be a scam.
I wonder if the same thing happened to nuclear engineering? Maybe the potential Tellers and Freeman Dysons born in the 01960s and later never got involved in nuclear physics, except for David Hahn, who ended up with paranoid schizophrenia after his mother committed suicide, and died at 39 of an overdose of opiates. Maybe nuclear engineering had a brain drain to electrical engineering, computer science, and related fields.