Good summary - and the strategy based on your break is coherent and makes absolute sense. Seems like it would be very difficult to pull off - as any 'off earth' project is. We only need look at the Mars objective and self drive delays to see how hoard all of this is. It also makes significant assumptions that there won't be a technology curve jumps - eg to different types of chips, Ai beyond data hungry LLM's, and that we won't find alt energy solutions... Of course if it works - could have big upside. I'd also say that I'm not so keep on private companies controlling infrastructure.
Yep. That’s probably the biggest hole in the whole thing. If AI suddenly needs 10x less compute, or someone cracks a completely different architecture, a lot of these assumptions go out the window.
What I found interesting wasn’t whether SpaceX is right. It was that they’re spending billions based on a view of what the constraints will be in 10 years, not where the demand is today.
And, the idea of a handful of private companies owning more and more of the underlying infrastructure doesn’t exactly fill me with comfort either.
Good summary - and the strategy based on your break is coherent and makes absolute sense. Seems like it would be very difficult to pull off - as any 'off earth' project is. We only need look at the Mars objective and self drive delays to see how hoard all of this is. It also makes significant assumptions that there won't be a technology curve jumps - eg to different types of chips, Ai beyond data hungry LLM's, and that we won't find alt energy solutions... Of course if it works - could have big upside. I'd also say that I'm not so keep on private companies controlling infrastructure.
Keep up to thoughtful posts JR.
Yep. That’s probably the biggest hole in the whole thing. If AI suddenly needs 10x less compute, or someone cracks a completely different architecture, a lot of these assumptions go out the window.
What I found interesting wasn’t whether SpaceX is right. It was that they’re spending billions based on a view of what the constraints will be in 10 years, not where the demand is today.
And, the idea of a handful of private companies owning more and more of the underlying infrastructure doesn’t exactly fill me with comfort either.