Yum, I can't wait 😋
Also to be clear: most fusion energy startups are chasing pipe-dream approaches. Most of the ones funded by traditional Silicon Valley VCs especially so. The average fusion startup is going bust. But 1-2 of them will be the SpaceX of the 2030s — the classic "10 year overnight success" story.
Full disclosure: - I've looked into the physics and technology quite closely. - I have close friends who are working on this professionally. - The relevant technology was invented at MIT and I've been able to ask people smarter than me lots of questions about it.
I have one steak dinner bet riding on this and I'm happy to have another if you're up for it? Suggested bet: First net-positive energy output (defined as "could be used to make a power plant" — no NIF-style creative accounting) happens before 2030.
Companies are suddenly pre-purchasing GW-decades, not MWh. That will drive the tens of $bn of R&D funding we'll need to develop all those technologies. It just wasn't going to happen otherwise: US power needs have been flat for more than a decade, and we were going to get more "business as usual."
We have *a lot* of awesome new tech in the pipeline: - supercritical CO2 turbines instead of steam (efficiency++) - passively-safe Gen 4 nuclear reactors (can't melt down!) - fusion power (GW-scale base load without long-term nuclear waste) If the AI boom provides the $bn to fund them, so be it!
This. The pitch at the dawn of the nuclear age was "abundant electricity too cheap to meter." I'm down for getting us back to that. I was chatting with some folks in the industry last week, and apparently AI is *massively* helping fund new nuclear and large-scale renewables development.
Part of the reason I'm not that concerned with AI's energy usage is because ~96% of the electricity the USA's putting in this year is renewables and nuclear. All our coal power pants combined have a 180 GW capacity, this year alone we're putting up 60 GW of renewables www.eia.gov/energyexplai...
Discovered a new #rustlanggithub.com/obi1kenobi/c...
Removing a pub field from a union type is clearly a major breaking change. But how about adding a new field? I chatted with @joshlf about this, and the current plan is like this: Any public API 1 f...
Ok listen. All I know in all my years of life is that this is the peak moment for my sense of humor. My whole life has led to this moment. Nothing will top this. It is wholly perfect.
