Hacker News new | past | comments | ask | show | jobs | submit login

> If they're too close, the heat they radiate away will end up on Earth again.

So put a mirror on the Earth side of it?

> If they're too far away, latency & maintenance will become an issue.

There are many compute tasks where latency is irrelevant. To take a recent example, AI model training. It does not matter if the compute farm is a few light minutes away when the computation itself is going to take days to months.

Maintenance is performed locally. It's not as if you're going to have Earth and then a single solitary server farm on the far side of the Sun. By the time this becomes relevant to planetary energy there are multiple space stations with permanent staff.




> Maintenance is performed locally. […] By the time this becomes relevant to planetary energy there are multiple space stations with permanent staff.

It seems whether or not you can keep maintenance staff close-by would depend on the temperatures of those server farms. Yes, the regime in which this could work might be fairly large but remember that we're talking about exponential growth of energy production here and the whole reason behind moving the power plants (and server farms) to deep space was that they were emitting enough heat to affect planet-level thermodynamics.


But that still doesn't affect energy density. If you want twice as many server farms then you build twice as many space stations in twice the volume of space and the energy density remains constant.


> But that still doesn't affect energy density.

But it does? I think you're confusing energy density with power density. The former is an integral over time and would be monotically increasing with time since nuclear fusion would allow us to basically pull energy out of of thin (ok, maybe not so thin) air.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: