Ok, but besides today's models mostly not being trained to find those inefficiencies, who says that finding them is actually the most profitable thing to do with a good brain? I know Juergen Schmidhuber likes it, but why should capital markets be the test of intelligence?

▲

jqpabc123 3 days ago | parent [-]

but why should capital markets be the test of intelligence?

Why do AI companies rely on capital markets to keep them afloat?

Why? Because this is where all the money is. And because the survival of AI companies ultimately depends on the ability to earn more than they spend.

▲

eli_gottlieb 3 days ago | parent [-]

> Why? Because this is where all the money is.

Well no, it's not. If finding inefficiencies in capital markets was the most profitable thing to do, the capital markets would never invest in anything but themselves. Recursion hits bottom somewhere and investors, at some point, want to invest in a real business, not just in a recursive investment vehicle.

	▲	jqpabc123 2 days ago \| parent [-]
		the capital markets would never invest in anything but themselves. The "capital markets" are not closed. The "market" itself doesn't decide what to invest in -- individuals and corporations do. I would expect an individual/corporation armed with real "super intelligence" to have a distinct advantage in this regard --- wouldn't you? What better way to prove the value of AI than applying it to earn mountains of money in the open market --- instead of setting mountains of the stuff on fire.