I think that psychologists call that “projection.”

I mean, launching hardware is an expensive endeavor all around. Which, I think we agree, the author of that funny image does not understand.

it’s an issue way 100x higher than that: ISS flies at ~400km and regularly has to reboost due to drag couple of times per year or it will fall out of the sky.

Power gen at high beta orbits (again, couple of times per year) is *significantly* reduced because of drag.

Like, it’s a huge deal.

- Also, radiation! (God, I hate rads) Anything that flies has to be rad hardened in new and exciting ways that no consumer tech has to be. All of that takes R&D to get right.

Starlink (yikes) works because it uses many cheap, expendable, redundant units and space is hard on everything.

- His analysis doesn’t consider bespoke aerospace HW efficiency - neither does the original post so fair is fair. But any such “data center” wouldn’t use commercial grade stuff. It would be a much smaller box full of ”Apple Neural Engines” or equivalent, optimized accordingly.

Notes:

- ISS uses anhydrous ammonia (vs H2O) in its ATCS. Doesn’t freeze, great viscosity, similar thermal capacity, etc. Potential issue in hab modules but fine in practice over 30(?) yrs
- Notes on geo-sync orbit & latency, rocket costs are spot on. LEO vs Geo costs are non-trivial!

The kids used to call that kind of thing “irrational exuberance.”

This is delightful

Grim stuff

It was a Very Good Book™

What did they do when you called them out on that?