Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror

Comment Re:Stupid investors are the problem, not AI (Score 1) 42

On the other hand, any of the LLMs are pretty decent at making web forms. Good enough you might be tempted to make your own instead of paying Peoplesoft or Salesforce a monthly fee to do it for you. Or maybe a few friends get together and make something customized for your specific niche and charge a fraction of what they do.

Comment Re:Completely Predictable (Score 1) 44

The reason they use inefficient fuel on the boosters and not on the upper stages is *because* of the rocket equation.

The difference between hydrogen and methane isn't an order of magnitude for a moon rocket, but it is somewhere in the neighbourhood of 3x the launch mass. Quite a bit more back when NASA was actually making the choice too. You'd have to actually run the numbers but I suspect it is the difference between a direct to the moon launch being just on the edge of practical possibility and requiring orbital refueling.

That's the point of the poster you replied to. You can get there with methane, but it's a much bigger operation with a lot more infrastructure required. Personally I think that's the way to go: space infrastructure is what we should be doing. But NASA was required to use existing shuttle parts, and you guys have gotta beat the Chinese or something.

Comment Re:Ketamine (Score 1) 183

Lol. You made a claim that is verifiably false and I pointed it out. But yeah, I'm being obtuse for not just going along with your fantasy.

WTF would you want to talk to the president of Deutsche Bank? I don't ahve a senator or a congressperson, but if I made an appointment with my elected national representative, yes, they would see me. Pity yours wouldn't. Maybe that's your real problem?

Comment Re:Change the Paradigm (Score 1) 91

If it comes true. I doubt it will. The reason we maintain libraries is because complexity grows nonlinearly with the size of a system. Keeping the systems small with well-defined interfaces manages that problem.

AI isn't magic. A computer might have greater capacity for tracking down complex behaviour than a human does but it isn't infinite. And the current systems, just like humans, do a lot better when they have good libraries to stick together than they do if you ask them for a big bare metal monolith.

Comment Re:1M satellites? (Score 1) 183

Are we talking tiny satellites here and you can pack hundreds into one launch?

Yes. When they say "space datacenter" you might picture something the size of a regular datacenter with giant solar panels on it. Turns out those solar panels would have to be REALLY giant, and the radiators even more so. So giant that the structure required to hold it all together in orbit would be impractical. You've also got to keep the thing up there and probably have a staff to keep it running. So you launch lots of little satellites instead.

Starlink satellites seem to have about 5 kW of solar power on board, most of which is used for propulsion and communications. An Nvidia DGX is 8 GPUs in a box and uses about 10 kW so you're talking about something about three times the size of a Starlink. Facebook is building a 2 GW datacenter with ~1 million GPUs. If you were to put that in space you'd be talking about a few hundred thousand super-starlinks. Allow for some improvements in efficiency but also big increases in capacity and a million seems a little small.

Future Starship is supposed to be able to launch about 100 Starlinks at once. Maybe make that 30 super Starlinks. So a million is ~ ten launches a day for ten years.

They wouldn't all be up there at the same time though. When one broke you'd deorbit it. When it became obsolete or ran out of fuel you'd deorbit it. Starlinks are designed for about five years in orbit but at current rates your video cards would probably wear out faster.

Is it more practical than Sam Altman's terawatts of nukes? Maybe. Is it more practical than Facebook's (similarly sized) terrestrial datacentre? Lol no.

Slashdot Top Deals

Ignorance is bliss. -- Thomas Gray Fortune updates the great quotes, #42: BLISS is ignorance.

Working...