I think that what people think they want from GenAI and what GenAI will be reasonably able to do in a repeatable way are still quite distant.

The results are probabilistic (as in: there is no comprehension) and very confident and that’s enough to misinform very badly.

I think the performance of the computing is still on the up: these new chips are incredible, and although faster hardware generally breeds slower software (so it always feels the same) there is potential for amazing speed.

But at this point we have so much: instant writing, audio and video calls; a much-modernised browser, payments, fitness and health (I just got an Apple Watch S8 after 6 years without one and it does so much more), maps, cameras… I’m not sure the next Big Thing is something we can see from this side of it being announced, from anyone. Google isn’t exactly in the ring either, and Microsoft on the software side seems to be going for GenAI.

I think we are on a plateau, everywhere. Maybe there’ll be a few years where the same stuff just gets faster and cheaper.

…how would that ever happen? The UK’s been out for almost 3 years, they know.

It’s not a contract with the devil, it’ll be software-controlled and resettable.