Gregory's Server
@shanecelis Counterpoint, somewhat: increases in energy demand trigger new energy infrastructure investment, which lowers the barrier for extra renewable capacity too. It's important to remember that the carbon cost of these things is determined by our current energy generation mix, and that this can - and is - being changing under our feet for the better.
Of course, I don't intend to defend the absurdity of training enormous AI models for the sole purpose of wilful copyright infringement.
@shanecelis I think I explained myself poorly here.
If your hypothetical grid supports 1 TW with constant demand, of which 50% is renewable, then the desire to eradicate the 50% non-renewable is low. The up-front cost of building infra is high, maybe too high to leave the local minimum status quo.
But if you've projected demand for 2 TW next year, it's time to get building for an extra TW: and while you're at it, why not the extra capacity needed to ditch the 0.5 TW of non-renewable power too?