Welcome to Thunderdome

In other news today, NVIDIA halted production on their RTX line for the same reason.

Yay AI!

Here's an interesting take on current AI news, and I only post it because I know there are kernels of truth in here, specifically that LLM is being discovered to NOT be on the path to AGI.

Not only is OpenAI working on adding ads to their experience, but they're also now toying with the idea of taking a cut of any creations people make using their product.

That'd be like Microsoft asking for a cut of Steven King's latest book just because he used Word.

BTW, Anthropic owns Claude. Duh. (aimed at me)

Leisher wrote: Fri Feb 06, 2026 12:05 am Sign up to work for AIs.

I think I'll wait until it's mandatory.

Good plan. Enjoy your retirement for a good year...

TheCatt wrote: Sat Jan 10, 2026 3:49 pm There really is no RAM: https://www.cnbc.com/2026/01/10/micron- ... msung.html

https://www.wsj.com/tech/micron-is-spen ... yURL_share

Micron will spend $200 billion to expand production. Starting to come online mid-2027

Tesla is apparently also working on producing RAM, but like Micron, it's years out. I'm pretty sure I read someone else was going the same, but I don't remember who.

Also, Intel is getting into the GPU industry. Not for consumer stuff right away, but hopefully they'll be able to shift quickly when the bubble bursts and we'll have three real GPU companies (if AMD doesn't implode).

Easy to spot for now. I think the only shake up will be scouts being needed again.

All current AIs are basically exactly the same.

AI Tilly Norwood releases song. "Song"

When we can virtually fuck her, as Dennis Miller once said, "It'll make crack look like Sanka."

META bought Moltbook because...?

Leisher wrote: Thu Mar 12, 2026 11:26 am META bought Moltbook because...?

Probably more of an acquihire, get the founders type of play. But with the amount of $$ Meta has, why not buy a few lottery tickets.

Listening to yesterday's Tech Brew...

They were covering a story about the future of AI moving to space for unlimited power needs (solar) and Jensen Huang was quoted as saying something along the lines of, "We've got to figure out how to use space to cool it."

I thought, "Try opening a window."

Does he not understand space?

I would assume the problem with space, while very cold, has very little actual mass. So convection is out of the question. You can basically radiate heat, right? Whereas most heat transfer here is through physical contact (air/water cooling).

TheCatt wrote: Wed Mar 18, 2026 2:49 pm Whereas most heat transfer here is through physical contact

And it's wonderful...

The easiest part is going to be shielding the critical bits from the sun. That'll be done by the solar panels.

As for the cooling, I wasn't kidding when I said open a window. You are correct that the heat wouldn't know where to go in space. However, doesn't that assume a static position? Say your satellite is in constant orbit with the solar panels always facing the sun. Panels on the sides of the satellite are, literally, open to space. As the satellite moves, would the heat not be left behind, as it isn't tethered to the computer equipment, and then get instantly cooled by space? Point being, you wouldn't be venting the heat, you'd simply be endlessly moving away from it.

Also, I didn't research it, but I know the ISS somehow vents its heat into space, so the tech is already there and we're just being nerds for fun.

Leisher wrote: Wed Mar 18, 2026 3:40 pm would the heat not be left behind,

There's no way for it to be left behind? There's no matter in space. It has to radiate out heat, so I think they usually use a bunch of thermal fins. Maybe, I dunno.

Leisher wrote: Wed Mar 18, 2026 3:40 pm and then get instantly cooled by space?

I don't think space can cool you, cuz there's no matter to cool you. Like cold water cools you because it touches you. In vacuum, no touchy.

Man, where's Thib when you need him?

TheCatt wrote: Wed Mar 18, 2026 3:46 pm There's no way for it to be left behind? There's no matter in space. It has to radiate out heat, so I think they usually use a bunch of thermal fins. Maybe, I dunno.

I don't either, but I equate it to a liquid. If a liquid is not in contact with something in space and that something moves away from it, how would it follow it? Wouldn't it eventually drift out the open panels on the opposite side of the direction the satellite is moving?

I did see some "metal whiskers" in a NASA document, and that might be the thermal fins?

Could you use pressure to move water like a cooling system in space? Because such systems DO touch components, perhaps that would do the trick?

Leisher wrote: Wed Mar 18, 2026 5:13 pm If a liquid is not in contact with something in space and that something moves away from it, how would it follow it?

But what's the liquid? Heat is not a liquid. It's a property of the thing. Thus, it's part of the thing. Moves with the thing.

Leisher wrote: Wed Mar 18, 2026 5:13 pm Could you use pressure to move water like a cooling system in space? Because such systems DO touch components, perhaps that would do the trick?

Absolutely, within a given system (like the ISS or such)... but then if there's excess waste heat, it would still need to leave the thing.

OK, I googled it.

Cooling things in space is achieved primarily through thermal radiation—emitting heat as infrared light—because there is no air for convection or conduction. Spacecraft use large radiator panels, heat pipes, and fluid loops (often with ammonia) to move heat away from electronics and humans, releasing it into the cold, dark void.

So they use circulatory systems to get it away from some things, then use large panels to radiate it into space.

A black surface in space would get VERY hot because nothing is there to move the heat off. We freeze because we are mostly water and the moisture leaving us and going into the vacuum of space takes our heat with it. A metal object doesn't "sweat" and there's no airflow for traditional heat sinks to work.

They says it's harder to stay cool in space than it is to stay warm.