@DanFessler That's for one datacenter. That has no maintenance crew and compute requires better hardware that's prone to failure so you have to keep replacing those satellites just as often as Starlink, maybe more. Assuming 5k network you'd need to launch 2k a year just as "maintenance"
@sobieraj_agata To po ki wuj w XTB jak można na realnym futures/opcje? Przecież na takim rozmiarze spread zjada Ci więcej niż F&C nawet jakbyś chciała to na MNQ zbierać. Dać Ci kontakt do kogoś z EdgeClear?
@PTrubey 3. So while you don't pay for land or electricity you NEVER stop paying for "compute" satellites. I wonder how fast and how expensive that gets. Because so far this looks like Cybertruck. Nice idea that might entirely fail in practice once the costs are known.
@PTrubey 2. To "operate" data centers in Space you'll need to build, control and keep replacing hundreds of them all the time. You don't get to just replace CPU - you have to rebuild and launch whole thing.
@PTrubey 1. One satellite -> one rack. So you need ~800 satellites for each "data center" in already crowded LEO. You can't do maintenance and compute CPUs are expensive. So now you'll have to replace hundreds of expensive satellites every year whole year.
@SzymonMachalic1 Ja mam serio pytanie. O chu chodzi z jazdą na CFD? Bo przecież można tak samo było stracić na ETFie na złoto czy kupując złoto spot. Czy może czegoś nie wiem i XTB manipuluje CFDkiem, a złoto realne już na 10k?
@zntrading The biggest expense for the W-day was premium steak dinner for my wife and I, best man and maid of honor. Most people have priorities upside down in their lives :)
@benEzra314159@peterrhague@PTrubey 3. Modern AI dcent will house about 800-1000 120kW racks. Given SpaceX launch numbers it's actually doable to get one dcentr equivalent in space every 1-2 years. Is this prone to chain reaction catastrophic failure (Kessler)? Yes. Will they find buyers? Maybe.
@benEzra314159@peterrhague@PTrubey 2. I'm not surprised SpaceX shows realistic design pre-IPO ... they'd have to have negative IQ to do otherwise ... but it's like Roadster, Roof, Hyperloop, Cybertaxi, Cybertruck, FSD ... lots of ideas, pumps the stocks, fails to ultimately deliver
@benEzra314159@peterrhague@PTrubey 1. Which was precisely my point. Hague had his BS "reactor in space" ... well, a CPU needs 100m2 ... and not surprisingly that's THE DESIGN that has been shown. Of course we know for sure this will happen. Just like FSD, colony on Mars, fleet of Tesla taxis and so on, right?
@stockdatamarket Sorry - my point was that "some" are mix strategies - like replicating BCOM but others leverage underlying + futures and sometimes those futures are just extra leverage on underlying. The latter is relatively recent addition that makes math way more dicey
@stockdatamarket Who wants to eat -3% on a year where benchmark did +3%. What's the base failure of a junior trader? Trading PnL. What does passive do? Force people to trade PnL
@stockdatamarket But it's also psychological. Retail is weak hands but institutional players have far less ability to eat risk these days because era of tracking error and passive flows squeezed that to near-0
@stockdatamarket 2. Add on top of that stacked and all the other concepts of alpha in modern investing products that are all "finance math hacks" based on a lot of assumptions prone to breaking in worst possible moments ;)
@DadGamerOne@ZacksJerryRig 2. In space the question is operating temp. Past certain temp you pick lightest solution that is good in moving heat to radiators. Pumped loop is best solution for anything operating at power that needs cooling.