That, and a $3 component is a *really* expensive one. $3 instead of $0.50 on 10,000,000 devices is $25,000,000.

Do you have a citation for that prediction?

It really wouldnâ(TM)t. Suppose your sats have 5m square in area - I choose this number because thatâ(TM)s what will fit in a stack in a starship. Now give them fold out solar panels, 4 on each side. Thatâ(TM)s 200 m^2 of panels. Solar panels in space generate about 1kW per square metre, so thatâ(TM)s 200kW of power per satellite. A normal nuclear plant will generate somewhere between 500MW and a GW. So to generate the same power you need 2500-5000 satellites. You can launch around 60 of those on a starship, so weâ(TM)re talking 40-80 launches. Based on SpaceXâ(TM)s predictions thatâ(TM)s about $800m-$1.6bn in launches. The intention is that these satellites go into sun synchronous orbit, at between 500 and 2000km altitude. At 2000km, you can stay there basically indefinitely. At 500km, you need some thrust, but youâ(TM)ll stay up there for 10 years passively. So in reality what weâ(TM)re dealing with here is the lifespan of the satelliteâ(TM)s usefulness, not its orbit that we need to worry about. GPUs get faster fast, but the advances are slower. Itâ(TM)s probably fair to assume that the satellites may last 5 years, maybe even a decade given the 5 years lead time here. Given that, letâ(TM)s assume that we need to launch new sats 5 times to match the lifespan of our nuclear plant. So conservatively, $8bn in launches costs.

How much does a nuclear plant cost? Well, vogtleâ(TM)s new units were predicted to cost $14bn. Theyâ(TM)ve so far cost $37bn. And thatâ(TM)s with a bunch of infrastructure all around them already built.

Thatâ(TM)s a *lot* of spare cash to build and operate the satellites right there.

Well, the conservatives have been informing us for sheâ(TM)d that the invisible hand of the market will magically solve it. Because there arenâ(TM)t people doing those jobs, theyâ(TM)ll end up paid more, and thus become more desirable.

Right, itâ(TM)s not a labour shortage, itâ(TM)s a wage shortage.

And since when is a turbopump âoeSpaceX technologyâ, as if no one had done it before.

Yes, but also, shielding something that comes into the atmosphere at 11km/s is very different from shielding something that comes in at 8km/s. There's a reason that the shuttle used a reusable heat shield, and this uses an ablative one - because the shuttle's reusable shield is no where near as capable as this one.

I have a Toto toilet - you non heated seat weenies are missing out! And thatâ(TM)s before you account for how accurately it can clamp your balls in place.

They have broken ground. SPARC is well under construction. ARC is going to break ground soon. Which is why I said 5-10, not 5.

Unlikely because its magnets suck by modern standards and so it canâ(TM)t reach very high pressures. SPARC/ARC might do it though. And yes, we absolutely will have commercial fusion in 20 years. Likely more like 5-10.

Fusion needs three things to work:

- A âoehighâ density of plasma (high is relative, in reality itâ(TM)s close to a vacuum)
- A high temperature (not relative, literally the hottest things in the solar system)
- A long time at those conditions.

The higher you make the product of those three the more fusion you get and the more energy you get.

At the same time though, you need to stop your fusing stuff from touching the outside of your reactor, because itâ(TM)s incredibly high temperature and radioactivity will destroy the reactor. It will also rapidly cool the reaction, shutting it down. That means you need to be really good at containing the hot plasma in the middle of the reactor.

These reactors do this by using magnets to squeeze the plasma into the middle (increasing the density), and various forms of heating to get to the desired temperatures. However, when you squeeze the plasma, it wriggles around like an eel trying to escape. At some point when you try to grip it too hard, the eel just slips straight out and shoots across the reactor into the walls. That limits the size of the product described at the start, and so limits the amount of power you can get out.

These scientists have discovered that if youâ(TM)re very careful about how you position the eel when you start squeezing you can squeeze a lot harder before it starts wriggling about like crazy.

Also, where do I sit as a fat person? Itâ(TM)s generally considered pretty impolite for me to take an economy seat and then overflow into neighbours space, so I get business seats to have enough space for a more ⦠uhh⦠generously proportioned human.

Where do I sit as someone who has severe restless leg thatâ(TM)s badly exacerbated by sitting in the exact same position for hours, and results in me repeatedly kicking the person in front of me?

Where do I sit as an old person who canâ(TM)t stagger down the incredibly narrow aisle for the length of the aircraft and then get over a bunch of fixed handrails?

Where do I sit as someone whose pelvis is in a cast that canâ(TM)t fit into an economy seat?

Thereâ(TM)s a ton of reasons for booking business class that arenâ(TM)t just that work are paying for the ticket.

And yet, we have concrete evidence that isreal continued to expand into palestinian territory during this time, fueling Hamas, and encouraging the terrorism. Stop trying to make bullshit claims that Isreal is somehow clean in this situation.

Ah yes, killing about 100 civilians (a disgusting attack, that should never have happened), deserves a genocide of literal millions, thatâ(TM)s a proportionate and reasonable response.

The crazy thing is that the price increases have already made Appleâ(TM)s infamously ludicrous upgrade pricing seem reasonable. Apple want £400 for 32GB. A couple of months ago, that was 4 times over the going rate. Now itâ(TM)s pretty much normal. The cheepest DDR5 6000 on newegg is £330, and higher bandwidth stuff costs of the order of £5-600. (And thatâ(TM)s ignoring that Appleâ(TM)s on chip RAM is anywhere between 50% faster and 300% faster than high end DDR5 depending on the chip).