> Dragon is using CotS, non rad-hardened CPUs. And it's rated to carry humans to space.

Those are not independent facts. They put the hardware inside, behind the radiation shielding they use to keep the astronauts safe. It's why regular old IBM laptops work on the Space Station too. That kind of shielding is going to blow your mass budget if you use it on these satellites.

SpaceX, which prefers COTS components when it can use them, still went with AMD Versal chips for Starlink. Because that kind of high performance, small process node hardware doesn't last long in space otherwise (phone SoC-based cubesats in LEO never lasted more than a year, and often only a month or so).

▲

NitpickLawyer 6 hours ago | parent [-]

> They put the hardware inside,

Which is exactly how you'd do a hypothetical dc in space. Come on, you're arguing for the sake of arguing. CotS works. This is not an issue.

> That kind of shielding is going to blow your mass budget

SpX is already leading in upmass by a large margin. Starship improves mass to orbit. Again, this is a "solved" issue.

There are other problems in building space DCs. Rad hardening is not one of them. AI training is so fault tolerant already that this was never an issue.

▲

notrealyme123 3 hours ago | parent [-]

> AI training is so fault tolerant already that this was never an issue.

Such nonsense.

	▲	NitpickLawyer 2 hours ago \| parent [-]
		Between fp nondeterminism, fp arithmetic, async gradient updates, cuda nondeterminism, random network issues, random nodes failing and so on, bitflip is the last of your concerns. SGD is very robust on noise. That's why it works with such noisy data, pipelines, compute and so on. Come on! This thread is having people find the most weird hills to die on, while being completely off base.