1,999 karma · joined January 30, 2011
Previously: - YouTube: - Shorts: founding engineer and infrastructure lead - Stories: infrastructure engineer - Uploads: Android TL - Vidmaker (acquired): founder/engineer (browser-based video editor) - Sony Vegas: engineer (desktop video editor) - WI state senate: legislative staffer - Various: LAMP stack hacker back when P was for Perl
One objective of the project might be simply to provide credible negotiating leverage when dealing with existing suppliers like NVidia. You don't have to deploy at scale for that to work, but you do have to look like you could if pushed hard enough.
Surely this is something AI tools will (eventually?) be able to help with?
Which is exactly why we previously had such a dense network: to provide redundancy in the face of inevitable outages.
That too, of course, is too low to live comfortably on your own in SF.
I've used GPT-5.5 and Opus both for FPGA design with good results. We built a lot of tooling around it to help the models, but even without that they're definitely capable of designing digital logic.
Also airliners usually just become cargo planes for quite a long time before retirement. Eg. there's a bunch of DC-3s still being commercially operated. Jet engine noise regs killed a bunch of early jets, but older prop aircraft are still going strong.
We've seen this movie before with crypto export bans in the 90s. The rest of the world caught up and then surpassed the US very quickly - and that was without the enormous financial incentives of AI.
If every university were subject to similar constraints, the average "quality" of research proposals would go down (everybody would have less time to spend on it) but since the pool of research dollars is assumed constant everyone would still get roughly their same slice - just with less overhead.