True, we don't know what technology will come tommorow and technology moves the fastest it can.

So predictions will always be off to the pessimistic side than the optimistic.

All we know we could invent a new form for cpu's, which actually happen with atom thick transitores which would allow us to stack it.

Agi and Asi could come way faster than predictions, the problem is that we just don't know what hurdle we have to climb to get there before it happens.

The predictions right now seems pretty realistic with agi being achieved around 2030 to 2040, and asi being achieved soonly after around maybe 2050, after that predictions are off.