IIRC article well enough, pros were that they were able to talk to anyone at company at the same time, had all data available constantly and were great at drawing normal from data so in standard scenarios they could easily lead the company.
They hallucinated sometimes, but research found that not more than actual CEO’s lol.
What they failed at was exceptional, bad situations. Too little data resulted in choices that ended with the simulated board firing them.
IIRC article well enough, pros were that they were able to talk to anyone at company at the same time, had all data available constantly and were great at drawing normal from data so in standard scenarios they could easily lead the company.
They hallucinated sometimes, but research found that not more than actual CEO’s lol.
What they failed at was exceptional, bad situations. Too little data resulted in choices that ended with the simulated board firing them.
This is exactly what I mean.
They just actually can do these kinds of tasks as well as people.
… and they’d be less expensive to run, in those roles, than the humans that currently have those salaries and stock options.
Unrealistic, boards haven’t fired or assasinated Elon yet, this is blatant anti-clanker bias.