I am currently going through the algebra/calculus used to train LLM's, in order to better write ai agents in my chosen niche.
It remains to be seen exactly what future ai has in terms of eliminating white collar office jobs.
I would be very surprised if LLM's achieved anything approaching AGI in my lifetime (if ever). I would be surprised if LLM's alone got to a point where they could replace white collar office jobs in at least the next decade (but concede it is not impossible at some point). I wouldn't want to bet what might happen beyond a decade either way. I think large parts of many office jobs can be made redundant by ai agents working in conjunction with LLM's now or in the near future. But getting those agents accurate enough is not easy, and still require quite a lot of human intervention.
Any AI toaster/coffee machine/bed you might see available to buy today is just nonsense, presumably the result of companies hyping ai in order to try and return some sort of short term income from the massive capital expenditure.
In short. Presumably the idea would be to 1. only beam down what is needed, and 2. have it replace fossil fuels, which are very much responsible for the change in the planet's energy imbalance.
It would also reduce the energy cost of less efficient 'on Earth' solar arrays, which have problems like intermittency that orbital solar panels wouldn't have.
IF this is anywhere near technically feasible it seems like exploring the idea publicly like this isn't a bad thing.
BUT, after a couple decades of watching proposed miracle tech going nowhere, I can say that ultimately hopium really isn't healthy: we needed to get real a decade or two (or three or four) ago. Relying on non-functioning future tech like carbon capture/storage (or this, if it isn't actually feasible) is nothing more than justification for not making necessary changes now.