alt Hacker News> For example, GPT Sol baked into a custom chip run for $100M that runs 10x as fast and 10x as cheap should pay for itself as long as the chip is useful for long enough.
but you trade updatability, which I don't think is worth it yet.
> For example, GPT Sol baked into a custom chip run for $100M that runs 10x as fast and 10x as cheap should pay for itself as long as the chip is useful for long enough.
but you trade updatability, which I don't think is worth it yet.
Maybe! (1) Would SOL level intelligence be useful 3 years from now? 5 years? (2) would dedicated chips be the most affordable way to run this model in 3-5 years?
I suspect the answer to both of these questions is yes right now, but I agree it’s borderline.