Elon Musk quote: SpaceX will lean in big on the Moon

[thespacecow]

Well this is certainly a big pivot from early this year when he said "The moon is a distraction". If you have been following him on X in the last few days, you can see him being increasingly interested in the Moon, goes well beyond the usual handwaving about "Moon Base Alpha", for example: https://x.com/elonmusk/status/1985013074382991407

Quantum computing is best done in the permanently shadowed craters on the Moon

There's also this one, which kind of telegraphed what his interest is: https://x.com/elonmusk/status/1984868748378157312

Starship could deliver 100GW/year to high Earth orbit within 4 to 5 years if we can solve the other parts of the equation.

100TW/year is possible from a lunar base producing solar-powered AI satellites locally and accelerating them to escape velocity with a mass driver.

Today we finally get some clarity, although it also raised more questions. First is the post in the title: "SpaceX will lean in big on the Moon", which is a reply to "I’m pretty sure that in a few years, you’ll be able to book a trip to the Moon or Mars with SpaceX, right here on 𝕏 app.". Unlike the "distraction" post earlier this year, this one has no ambiguity.

Then he finally explained his reasoning:

Of course not. Duffy understands nothing about how anything works.

The importance of the Moon for scaling AI did not occur to me until we did the scaling math on Earth to figure out what it would take to launch >1TW/year of orbital AI.

Now the question is, what will SpaceX do about it? Note given this is related to scaling AI compute, it has some overlap with the other thread SpaceX's new business: data centers in space. I see that thread more about near term data center play based on Starlink launched from Earth, while this thread focus on what SpaceX will do on the Moon, which would obviously based on Starship.

[Kazioo]

I think the "moon is a distraction" was either meant as a reply to a different post or a misunderstanding, because OP said nothing about the moon.

My hypothesis is this statement was a reference to the rebuttal of a very popular concept of using moon as a "gas station" to get to Mars "easier", because it ignores near future logistics and real economics.

[thespacecow]

Obviously there's a lot SpaceX can do with the Moon once they get Starship working as intended, but the question always comes down to funding: Where would the money come from?

A few possibilities:
1. Self-fund: They could use Starlink revenue, sort of invest in the future. Similar to how they used profit from launch to invest in Starlink, although I dare say this investment would be more risky than Starlink.
2. VC investment: They could raise more money based on their valuation. Elon will need to convince investors that build AI compute on the Moon would be actually profitable, not an easy task.
3. Debt financing: Very common for building infrastructure. This is how a lot of AI data centers are being funded right now, use GPUs as collateral. For the Moon, maybe they can use lunar real estate and infrastructure as collateral?
4. NASA funding: Not a lot but a steady stream. However they'll need to fight Congress to get it, also not an easy task.

It would be interesting to see how this plays out.

[DanClemmensen]

Building massive AI capability using the Moon potentially conflicts with SpaceX/Musk's "make life interplanetary" goal.

The explicit reason to go to Mars is as a sanctuary for humanity in the event of destruction of humanity or of human civilization on Earth due to any of several existential threats. The threat Musk mentions most often is runaway AI: SpaceX is in a race to create a self-sustaining human population on Mars before AI kills human civilization on Earth. But if AI can be produced more quickly using the Moon, then the AI risk is suddenly larger.

Maybe Musk thinks he can guide an AGI that is under his control, so his AI will not be a threat. The AI safety community has been working on this for thirty years and has not really found a solution, but maybe it's worth a try.

[meekGee]

Building massive AI capability using the Moon potentially conflicts with SpaceX/Musk's "make life interplanetary" goal.

The explicit reason to go to Mars is as a sanctuary for humanity in the event of destruction of humanity or of human civilization on Earth due to any of several existential threats. The threat Musk mentions most often is runaway AI: SpaceX is in a race to create a self-sustaining human population on Mars before AI kills human civilization on Earth. But if AI can be produced more quickly using the Moon, then the AI risk is suddenly larger.

Maybe Musk thinks he can guide an AGI that is under his control, so his AI will not be a threat. The AI safety community has been working on this for thirty years and has not really found a solution, but maybe it's worth a try.

You can say the same about Starlink.

Mars is still the goal, as part of "interplanetary". The moon isn't.

But it appears that Musk thinks it's an important part of long-term AI and that's an important piece of everything.

[DanClemmensen]

But it appears that Musk thinks it's an important part of long-term AI and that's an important piece of everything.

I think you missed my point. Massive AI is the threat, not the solution.

[Twark_Main]

Quantum computing is best done in the permanently shadowed craters on the Moon

A unique and irreplaceable "ice core" record of the Earth-Moon system going back billions of years??  Let's build a datacenter on it just to save a bit of cooling. 

I know this is just brainstorming, but that has got to be one of the worst ideas ever.


There are plenty of mostly shadowed sites (far less scientifically interesting!) that can host such infrastructure. Literally just hang up a suspended Mylar curtain to block the remaining sunlight direction, if you really need that extra sliver of cooling performance...

[Robotbeat]

I don’t agree with Cody there. He has fallen down the “don’t touch anything” hole.

But I do think there’s no reason to use the permanent craters. You can just make your own with a sunshade. And it doesn’t need to be on the Moon.

(Which is an interesting option for ISRU, now that I think about it… you could make a could trap on Mars using MLI.)

[meekGee]

But it appears that Musk thinks it's an important part of long-term AI and that's an important piece of everything.

I think you missed my point. Massive AI is the threat, not the solution.

Had a looong conversation with the missus about exactly that just yesterday... 

If you project forward, some of the outcomes are really wild....

I don't think the threat to humanity is AI going rogue, and if it is  a Mars colony won't help.

A Mars colony is an insurance against the "normal" historical collapse of civilizations. Always limited in scope to a socio-economic sphere, and the problem now is that there is only one such.  Mars would make a second one.

Either way though, I think the timeline for a self sufficient Mars (20 years?) is short enough, and also that Mars will be more AI-integrated than Earth.

As to the fate of AI and humanity, it's a whooole nuther discussion.  Not here.

[meekGee]

Quantum computing is best done in the permanently shadowed craters on the Moon

A unique and irreplaceable "ice core" record of the Earth-Moon system going back billions of years??  Let's build a datacenter on it just to save a bit of cooling. 

I know this is just brainstorming, but that has got to be one of the worst ideas ever.


There are plenty of mostly shadowed sites (far less scientifically interesting!) that can host such infrastructure. Literally just hang up a suspended Mylar curtain to block the remaining sunlight direction, if you really need that extra sliver of cooling performance...

The cooling you can get anywhere, agreed, and I like orbit better for the benefits of zero g.

Doing it on the moon is IMO a much longer prospect, factoring using lunar material and not having to launch it at all.

I think the actionable plan will end up with orbital AI even after it is produced on the moon, which itself is a 20 year goal.

[Greg Hullender]

The idea that there is any advantage in putting AI in orbit or on the moon is so stupid, I'm wondering if Elon is just trying to make up an economic justification for having a moon colony. I'd be concerned if I thought he really believed this.

[Vultur]

A Mars colony is an insurance against the "normal" historical collapse of civilizations. Always limited in scope to a socio-economic sphere, and the problem now is that there is only one such.  Mars would make a second one.

I think this is the much more plausible argument, but it doesn't seem to be the one Elon Musk generally makes.

World civilization is now all connected (unlike in Roman or Mayan times etc), therefore a collapse would hit everywhere, therefore we want a backup for technological civilization and knowledge... That's one line of thought.

But the "light of consciousness" verbiage Musk often uses suggests he's looking at a backup for humanity (not just technological civilization). This is probably much more demanding, as the backup for technological civilization argument probably doesn't require *permanent* self-sufficiency, just the preservation of the ability to return to Earth after 10-20 years of cutoff (to reintroduce technological civilization to Earth after the immediate post -collapse crises are over). This probably doesn't require making ICs etc on Mars.

[CoolScience]

Well this is certainly a big pivot from early this year when he said "The moon is a distraction". If you have been following him on X in the last few days, you can see him being increasingly interested in the Moon, goes well beyond the usual handwaving about "Moon Base Alpha",

If you have been following him for any lenght of time, you would know that the "moon is a distraction" statement has never been his view. It is a truncated quote to deliberately misrepresent his argument against mining propellant on the moon to support a Mars transport architecture. There is no big pivot here. The only thing new is him coming up with more uses for being on the moon.

[xvel]

The idea that there is any advantage in putting AI in orbit or on the moon is so stupid, I'm wondering if Elon is just trying to make up an economic justification for having a moon colony. I'd be concerned if I thought he really believed this.

If it's so stupid you should be able to easily prove it right? So please do it (you know some approximate calculations and such) instead of whinig.

[Robotbeat]

The idea that there is any advantage in putting AI in orbit or on the moon is so stupid, I'm wondering if Elon is just trying to make up an economic justification for having a moon colony. I'd be concerned if I thought he really believed this.

Where on Earth you planning to put 10s of Terawatts of solar panels? Note this is way bigger than just de-carbonizing the existing grid, about 500GW average in the US. At 100TW, even waste heat from fission or fusion matter.

[leovinus]

The idea that there is any advantage in putting AI in orbit or on the moon is so stupid, I'm wondering if Elon is just trying to make up an economic justification for having a moon colony. I'd be concerned if I thought he really believed this.

If it's so stupid you should be able to easily prove it right? So please do it (you know some approximate calculations and such) instead of whinig.

Can we stop the attacks? Greg happens to know a lot about the topic and you can rely on his opinion.

[Robotbeat]

Lots of us know about data centers and spaceflight. Greg is going to have to defend his position like everyone else, not just assert it.

[CoolScience]

The idea that there is any advantage in putting AI in orbit or on the moon is so stupid, I'm wondering if Elon is just trying to make up an economic justification for having a moon colony. I'd be concerned if I thought he really believed this.

If it's so stupid you should be able to easily prove it right? So please do it (you know some approximate calculations and such) instead of whinig.

Can we stop the attacks? Greg happens to know a lot about the topic and you can rely on his opinion.

You are the one making attacks. Argument from authority is simply wrong. Greg can defend his position against the counterarguments or not.

It is a problem when experts just say "trust me." It is worse when someone declares an expert on computational linguistics the authority on AI datacenters of unprecedented scale. (Hint: it hasn't been done before, so no one is sole authority on that.) Especially over the opinion of someone who leads multiple companies with significant AI products, has built AI data centers in record time, and has a company that has more active satellites than the rest of the world combined, providing global scale internet access. Neither is beyond correction, and arguments can certainly be made, but trying to shut down discussion like you just did is unacceptable.

[meekGee]

A Mars colony is an insurance against the "normal" historical collapse of civilizations. Always limited in scope to a socio-economic sphere, and the problem now is that there is only one such.  Mars would make a second one.

I think this is the much more plausible argument, but it doesn't seem to be the one Elon Musk generally makes.

World civilization is now all connected (unlike in Roman or Mayan times etc), therefore a collapse would hit everywhere, therefore we want a backup for technological civilization and knowledge... That's one line of thought.

But the "light of consciousness" verbiage Musk often uses suggests he's looking at a backup for humanity (not just technological civilization). This is probably much more demanding, as the backup for technological civilization argument probably doesn't require *permanent* self-sufficiency, just the preservation of the ability to return to Earth after 10-20 years of cutoff (to reintroduce technological civilization to Earth after the immediate post -collapse crises are over). This probably doesn't require making ICs etc on Mars.

This takes me back to yesterday's home discussion...

"Light of consciousness" is specifically non biological.

If the endgame of AI and robotics is artificial people that are equal and superior to biological ones, you don't need to have "AI kill the humans".  If the artificial ones live to be 1000 years old, they'll just basically take over, have their own civilization, and be as vulnerable to collapse as we are....

Except more so, since when we collapse  the worst that can happen is we go back to the stone age. We can still make babies though.

When an artificial civilization (which, mind you, still carries the "light of consciousness") collapses, they can't even make new ones anymore - it's a much bigger problem.

So yeah. Where were we?

[Cabbage123]

Are we over conflating the quantum computing comments with the AI comments?

Quantum computing requires very cold temperatures, just a few degrees above absolute zero. AI data centers obviously need cooling, but nowhere near to the same degree.

The talk of such craters, is presumably for Quantum, as the baseline temperatures there are just above what is needed.

So possibly future "Quantum powered AI" but not your common or garden variety.