While still a valid callout, these “new chips made out of non-silicon are BETTER in EVERY WAY!” experiments happen about once every 6 months. As it turns out, 10% less energy use isn’t compelling enough to convince a trillion dollar industry to completely retool from the ground up, and even without the profit motive that’s unlikely to change much.
Two things why this is different. China controls bismuth. So this is a strategic resource they control and puts them squarely outside of current western restrictions on silicon. They are also currently tooling to get their 3nm forges for silicon off the ground. So they are already undertaking the expense. A bonus third thing: China will develop this if they think it is in their long-term best interest in spite of the cost. It’s a cost for them to spend now. If it truly gives them an edge here, then it will be even more painful when the west has to do all that retooling to catch up. Patience is a weapon.
Just to be clear: “China wouldn’t dare because it’s too costly” is a way the west has consistently underestimated them for 40 years now. This is not something a planned economy worries as much about. They can think more long term. For the capitalist, if you can’t profit today, there is no tomorrow.
To be fair, it says in the article that this new transistor architecture can be fabricated with existing industrial platforms, so it doesn’t seem like it would require a whole lot of retooling.
As it turns out, 10% less energy use isn’t compelling enough to convince a trillion dollar industry to completely retool from the ground up, and even without the profit motive that’s unlikely to change much.
is likely to play out differently outside of a US-style declining capitalist economy combined with China’s current situation.
Complicated logistics isn’t a problem when we’re essentially in an arms race but with computing tech. If the CPC see developing their own semiconductor technology as politically or militarily essential, they will do it, at whatever cost.
The main issue is time. The technology is more useful the sooner it can be implemented.
I’m no expert, but if it works I can definitely see a use for 40% faster speeds in things like AI server farms and military applications. It’s not like they need to roll it out to every phone and dishwasher in the world.
10% less energy is also a bigger deal than it sounds. Heat dissipation is basically the bottleneck of modern processors. That’s why we’ve shifted to more cores instead.
“But turning laboratory breakthroughs into commercial chips typically takes years — sometimes decades”
2 years ago: China is 10 years behind on semiconductor technology!!!
1 year ago: China is 2 years behind on semiconductor technology!!!
6 months ago: China is 1 year behind on semiconductor technology!!!
Now: China’s new transistor invention will take 10 years to commercialize!!!
While still a valid callout, these “new chips made out of non-silicon are BETTER in EVERY WAY!” experiments happen about once every 6 months. As it turns out, 10% less energy use isn’t compelling enough to convince a trillion dollar industry to completely retool from the ground up, and even without the profit motive that’s unlikely to change much.
Two things why this is different. China controls bismuth. So this is a strategic resource they control and puts them squarely outside of current western restrictions on silicon. They are also currently tooling to get their 3nm forges for silicon off the ground. So they are already undertaking the expense. A bonus third thing: China will develop this if they think it is in their long-term best interest in spite of the cost. It’s a cost for them to spend now. If it truly gives them an edge here, then it will be even more painful when the west has to do all that retooling to catch up. Patience is a weapon.
Just to be clear: “China wouldn’t dare because it’s too costly” is a way the west has consistently underestimated them for 40 years now. This is not something a planned economy worries as much about. They can think more long term. For the capitalist, if you can’t profit today, there is no tomorrow.
To be fair, it says in the article that this new transistor architecture can be fabricated with existing industrial platforms, so it doesn’t seem like it would require a whole lot of retooling.
also the game theory logic of
is likely to play out differently outside of a US-style declining capitalist economy combined with China’s current situation.
I don’t think jaomarromis is saying the profit motive is the biggest factor in making the logistics complicated.
Complicated logistics isn’t a problem when we’re essentially in an arms race but with computing tech. If the CPC see developing their own semiconductor technology as politically or militarily essential, they will do it, at whatever cost.
The main issue is time. The technology is more useful the sooner it can be implemented.
I’m no expert, but if it works I can definitely see a use for 40% faster speeds in things like AI server farms and military applications. It’s not like they need to roll it out to every phone and dishwasher in the world.
10% less energy is also a bigger deal than it sounds. Heat dissipation is basically the bottleneck of modern processors. That’s why we’ve shifted to more cores instead.
Speak for yourself, my dishwasher can only play Doom 2016. No rest until it can crush Eternal at 144hz high graphics