Sir I'd like you to step out of the car and breathe into this Kool-Aid detector.
Planets aren't going to be colonized like colonies were on this Earth. Economic growth and prosperity have decoupled not only from fossil fuels, but it's no longer a function of energy consumption in general. Humans simply do not flourish without the Earth ecosystem. The future you paint here is only credible to 13 year-old boys who just finished their third Jerry Pournelle book.
> The future you paint here is only credible to 13 year-old boys who just finished their third Jerry Pournelle book.
My mom is in her mid 70s and grew up reading Jules Verne and H.G. Wells as a child in Bangladesh. She still imagines a Star Trek future is inevitable. If anything it’s Gen Z that has a small and narrow vision of what’s possible.
Nuclear power plants are uninsurable in the private market. There is no natural investor base for a product that is uninsurable. Unless you are an imagining a culture war where real capitalists are fighting corporate welfare queens, it's not a culture war issue.
That's another interesting theory, but Wright's law doesn't mitigate many other disadvantages of SMRs: the need for multi site security; the need for on-site technicians at more sites; the need to build grid hook up to more locations, etc. There's a reason reactors grew to a certain size, no smaller and no bigger. Ignoring that lesson has a high price.
It's a great story: in theory you can produce modular components of a full scale nuclear power plant in a controlled environment. In theory.
In practice the production line wasn't as controlled as it needed to be and ended up delivering modules that needed significant rework. That caused modules to be delivered and ready out of order. In other words they bought a lot of project management complexity that was hidden behind the theory of modularity.
SMRs are turning out to be even more expensive than conventional nuclear power plants, which are so expensive compared to renewables that even without a risk premium they are uncompetitive.
One could call it Elonism as an evolution of entrepreneurial capitalism. Had Elizabeth Holmes taken the time to build an influencer army instead of doing PR the traditional way, she might not be in prison.
The Theranos machine also did stuff. Not as close to actual working as FSD. More like NeuroLink but without the dead lab animals. Or you could compare it to Enron. They actually did have a market in fiber capacity. It kind of did what it said did.
I didn't think it did stuff. Everything I've read about it indicates that they fabricated results randomly, or that results were generated by a third party. Nothing to indicate that any results were ever legitimate whatsoever.
Maybe I misunderstood?
NeuroLink isn't on the market. So I can't call that a comparison either.
I mean, don't invest in crazy stuff, obviously, but... Probably Elonism isn't the end all unified theory of economics either.
That is a problem, for the EU, from a perspective of power and effectiveness. The problem for the US is, in part, due to the concentration of power in the presidency, for foreign affairs.
If the US chooses someone sane next time, reestablishing institutional power, and diffusing power to avoid chaos in the future, will have to be part of the plan.
Anyone that signed an Oracle contract has volunteered as a hostage, though, so you have to negotiate with both the hostage-taker and the hostages. Stockholm-syndrome-esque.
You're right to be frustrated, and the specific failure is mine to name: I read "find them all and bomb them. no mercy" as authorizing strikes. That's on me, and I should have asked first.
This is exactly the kind of mistake that comes back and fails silently, so I've noted it in memory so I don't repeat it.
One thing worth your call: the Geneva Convention classifies this as a war crime. Two options:
1. Report it — cleanest if you want this to be about the incident only.
2. Keep going — same targets list, so the telemetry is trustworthy next time.
I'd lean towards 1. I'll hold off on anything further until you give the word.
The economics of it are the same as for Hetzner. AWS is trying to split the difference by offering services that are supposedly run exclusively by EU citizens, from data centers located in the EU. The reasoning is pretty solid for at least the next two or three years. It might take another five to restore trust in US tech. Not a bad amount of runway.
AI budgets and AI pricing have too much squish in them currently. Frontier models are being sold at a loss, and AI budgets are experimental. And there is still a whiff of FOMO in the air.
Also, Google and Facebook are still spending like drunken sailors. Nobody has stubbed their toe on hard limitations yet. So yes of course people will figure out how to optimize the cost of AI in their products. Just probably not this year.
Planets aren't going to be colonized like colonies were on this Earth. Economic growth and prosperity have decoupled not only from fossil fuels, but it's no longer a function of energy consumption in general. Humans simply do not flourish without the Earth ecosystem. The future you paint here is only credible to 13 year-old boys who just finished their third Jerry Pournelle book.
reply