A promising start, but a thousand transistors at 25 kilohertz puts it where silicon tech was 60 years ago, so they've a long, long way to go.
If it scales, they can use modern tech and know-how to accelerate their progress and they can get funding, maybe this will be viable in a decade or so.
partial_accumen@lemmy.world
on 12 Jun 17:52
nextcollapse
A promising start, but a thousand transistors at 25 kilohertz puts it where silicon tech was 60 years ago, so they’ve a long, long way to go.
If you’re talking about the desire to replace today’s modern CPUs, sure. However, in the world of electronics there are lots and lots of support electronics and ICs that run way slower than 25kHz. All of this assumes the technology can scale for cost effective manufacturing yields at this current speed. If its both expensive AND slow, it will have far fewer use cases.
The article seems to imply that the intention is to replace silicon entirely, but agreed, there might be niches where it can replace silicon even if full replacement might be unrealistic.
SharkAttak@kbin.melroy.org
on 12 Jun 22:27
collapse
I was thinking more about the availability of "molybdenum disulfide and tungsten diselenide" opposed to silicon, they don't sound exactly like Home Depot stuff.
Fair point. From what I can tell, refined tungsten is actually an order of magnitude cheaper(!) than refined silicon, but molybdenum is over two orders or magnitude more expensive. ~300USD per ton, ~2000USD per ton and ~60000USD per ton respectively.
I assume that if this got up to scale industrially, savings could be made by recycling high purity molybdenum waste, but yes, it's not going to be cheap.
kata1yst@sh.itjust.works
on 13 Jun 02:38
nextcollapse
Modern transistors aren’t just silicon though. The silicon is doped with various materials, presumably gallium, boron, arsenic, phosphorus, and cobalt, among other elements.
Exactly… the price of these new materials/CPUs isn’t in the amount of “exotic” elements, which is barely measurable on a per-unit basis, but in the production.
threaded - newest
A promising start, but a thousand transistors at 25 kilohertz puts it where silicon tech was 60 years ago, so they've a long, long way to go.
If it scales, they can use modern tech and know-how to accelerate their progress and they can get funding, maybe this will be viable in a decade or so.
If you’re talking about the desire to replace today’s modern CPUs, sure. However, in the world of electronics there are lots and lots of support electronics and ICs that run way slower than 25kHz. All of this assumes the technology can scale for cost effective manufacturing yields at this current speed. If its both expensive AND slow, it will have far fewer use cases.
The article seems to imply that the intention is to replace silicon entirely, but agreed, there might be niches where it can replace silicon even if full replacement might be unrealistic.
I was thinking more about the availability of "molybdenum disulfide and tungsten diselenide" opposed to silicon, they don't sound exactly like Home Depot stuff.
Fair point. From what I can tell, refined tungsten is actually an order of magnitude cheaper(!) than refined silicon, but molybdenum is over two orders or magnitude more expensive. ~300USD per ton, ~2000USD per ton and ~60000USD per ton respectively.
I assume that if this got up to scale industrially, savings could be made by recycling high purity molybdenum waste, but yes, it's not going to be cheap.
Modern transistors aren’t just silicon though. The silicon is doped with various materials, presumably gallium, boron, arsenic, phosphorus, and cobalt, among other elements.
en.wikipedia.org/wiki/Doping_(semiconductor)
But material costs don’t matter much in computer pricing.
Exactly… the price of these new materials/CPUs isn’t in the amount of “exotic” elements, which is barely measurable on a per-unit basis, but in the production.
CPUs are not made in a home depot.
Hah, maybe not where you live.
I feel like I’d drop it, and it would crack in half.
Linus? That you?
They say it’s 2D so I would end up setting it somewhere and lose it because I can only see the sides it doesn’t actually possess.
They just keep making phones thinner and thinner
Sophons when?
That’s very cool, but does anyone else think the title image is AI generated? Neither image nor caption seem to sit right, nor fit together.
Is the ‘caption’ actually (derived from) a prompt?
Sooo … is it a conceptual illustration, or an actual SEM image?
Pretty sure the illustration here is just being used to demonstrate what’s taking place visually for the reader!
The ‘laptop’ is s conceptual illustration. The image shown on the laptop screen is an actual SEM image.