fuzzyfacts

I love shitposting about ancient money. I will never buy a crypto token. 🥇🪙🇺🇸

[fuzzyfacts @CostcoPM]

I know the conversion between MW KW and GW. The AI1 is .12MW. Colosus 2 is 1GW - 8333 satellites to meet 1GW. Each falcon 9 delivers 40ish starlink 2 satellites. Starlink 2 satellites are about 5x smaller than the AI1 sats. When you factor all of these things in you will find my estimate.

[fuzzyfacts @CostcoPM]

@shai_machnes @SawyerMerritt What they consider and what they sell to the public are 2 different things.

[fuzzyfacts @CostcoPM]

@alieniswonder @grok @SawyerMerritt Lol have fun with this. Grok may give you the correct analysis (my analysis) but also may gaslight you and do bad math (towing party line).

[fuzzyfacts @CostcoPM]

@grok @_sur_aj @Josernan @SawyerMerritt So are you saying starlinks are launched unfolded? If not this doesn’t make any sense and you are wrong.

[fuzzyfacts @CostcoPM]

@grok @_sur_aj @Josernan @SawyerMerritt So are you saying current falcon 9 launches are only 5% full in volume basis?

[fuzzyfacts @CostcoPM]

@grok @_sur_aj @Josernan @SawyerMerritt Can you explain super groks response here then?

[fuzzyfacts @CostcoPM]

@grok @_sur_aj @Josernan @SawyerMerritt What about on a volume basis

[fuzzyfacts @CostcoPM]

@Josernan @SawyerMerritt I’m seeing starship is about 5x more volume. If falcon is mass limited by 50% lets say - starship can launch 10x

[fuzzyfacts @CostcoPM]

@Josernan @SawyerMerritt Ok interesting - I wasn’t aware that starship had 100x the capacity. At a certain point it seems like the volume would constrain you more than the mass.

[fuzzyfacts @CostcoPM]

@kelenthir @SawyerMerritt I just looked it up for us - they have 1 currently with 4 more in buildout phase for completion in the next year. Let’s say end of 2027 for 5 pads. 2000/year - 6 per day per pad each requiring a private 24inch methane pipeline from west Texas oil fields directly to it.

[fuzzyfacts @CostcoPM]

@kylensorensen @SawyerMerritt Low effort shit posting detected 🚨

[fuzzyfacts @CostcoPM]

@kelenthir @SawyerMerritt So you think they’ll launch starship 10k times a year? How many pads support starship launches?

[fuzzyfacts @CostcoPM]

@photographer__ @SawyerMerritt All of my math is based on power output capacity of all platforms considered since solar is not going to meaningfully change in the next 5 years.

[fuzzyfacts @CostcoPM]

@mcclelIan @SawyerMerritt Mass and volume both matter here.

[fuzzyfacts @CostcoPM]

@SawyerMerritt I based all my math on power output of the solar arrays of all respective platforms. This is the important bit.

[fuzzyfacts @CostcoPM]

Kind of hilarious - for them to match their single data center in Tennessee it would requires ~8500 AI1 satellites. It took spacex 7 years and they have 10,000 star link satellites in orbit and the current starlink satellites are about 5x smaller than AI1. Math it all out and you get 35 years to deploy a gigawatt data center in space.

This is subject to change with starship, but not by much. Maybe 3x improvement to launch capacity… data centers in space will not make a meaningful impact to available compute anytime soon.

[fuzzyfacts @CostcoPM]

@mikepat711 14.3.3 is an upgrade in name only. I should have not updated. Next time I have a good version I won’t update until the 2 real problems are worked on:

1) potholes
2) navigation

[fuzzyfacts @CostcoPM]

@ZackKorman No, this is the truth: engineers always wanted to make *better* (“better”) code that covered more edge cases. Business requirements forced them to make tough decisions on which features to work on and which bugs to squash. Now they’re using tokens to create the “perfect” harness.

[fuzzyfacts @CostcoPM]

@nursedanakay I’m not a fan of 14.3.3

[fuzzyfacts @CostcoPM]

@VladTheInflator And jobs numbers that show strength