The chip export controls are largely irrelevant. Westerners badly underestimate the Chinese and they have caught up to 7nm at scale. They also caught up to 5nm, but not at scale. The original chip ban was meant to stop China from going sub 14nm. Instead now we may have just bifurcated advanced chip capabilities.
The general argument before was “In 10 years, when the Chinese catch up to TSMC, TSMC will be 10 years ahead.” Now the only missing link in the piece for China is EUV. And now the common argument is that same line with ASML subbed in for TSMC. Somehow, I doubt this will be a long term blocker.
Best case for the Chinese chip industry, they just clone EUV. Worst case, they find an alternative. Monopolies and first movers don’t often have the most efficient solution.
There’s some motte/bailey to this argument, between different levels of effect. With AI, the crux is timelines. It’s looking like in late 2025 or early 2026 there will be gigawatt-scale training systems that cost $15-$50 billion and are capable of training a model with 100-400 times GPT-4 compute in a few months, or of running 100-400 GPT-4 scale experiments. Perhaps this doesn’t move the needle on TAI timelines, but it seems too early to tell.
The impression I have from reading Chip War is that EUV is a pretty massive hurdle which took the West well over a decade to conquer. However, I also thought that 5nm was impossible without EUV, which seems to be no longer true, so this may be too complex a topic to make meaningful predictions about without deeper expertise.
I do agree that the chip ban won’t stay relevant for long. My guess is that it maybe bought an 8 month delay, +/- 2 months? Of course, we need to maintain the ban to keep the lead we got from doing this, so the ban remains relevant until China has fully caught up in chip making capacity or until someone gets to economically and militarily decisive AI (which I expect can be achieved even before AGI, if AGI gets delayed for longer than the 36 months which I expect it in).
The chip export controls are largely irrelevant. Westerners badly underestimate the Chinese and they have caught up to 7nm at scale. They also caught up to 5nm, but not at scale. The original chip ban was meant to stop China from going sub 14nm. Instead now we may have just bifurcated advanced chip capabilities.
The general argument before was “In 10 years, when the Chinese catch up to TSMC, TSMC will be 10 years ahead.” Now the only missing link in the piece for China is EUV. And now the common argument is that same line with ASML subbed in for TSMC. Somehow, I doubt this will be a long term blocker.
Best case for the Chinese chip industry, they just clone EUV. Worst case, they find an alternative. Monopolies and first movers don’t often have the most efficient solution.
There’s some motte/bailey to this argument, between different levels of effect. With AI, the crux is timelines. It’s looking like in late 2025 or early 2026 there will be gigawatt-scale training systems that cost $15-$50 billion and are capable of training a model with 100-400 times GPT-4 compute in a few months, or of running 100-400 GPT-4 scale experiments. Perhaps this doesn’t move the needle on TAI timelines, but it seems too early to tell.
The impression I have from reading Chip War is that EUV is a pretty massive hurdle which took the West well over a decade to conquer. However, I also thought that 5nm was impossible without EUV, which seems to be no longer true, so this may be too complex a topic to make meaningful predictions about without deeper expertise.
I think the next 24 months are going to be critical. Thus, if you think the chip ban slowed China by even a few months, then it is/was quite relevant indeed.
I do agree that the chip ban won’t stay relevant for long. My guess is that it maybe bought an 8 month delay, +/- 2 months? Of course, we need to maintain the ban to keep the lead we got from doing this, so the ban remains relevant until China has fully caught up in chip making capacity or until someone gets to economically and militarily decisive AI (which I expect can be achieved even before AGI, if AGI gets delayed for longer than the 36 months which I expect it in).