Comment by bedstefar

And even if we did manage to limit/cull hardware development, what would keep people from continuing to find ever more efficient models runnable on contemporary hardware?