Is there a class of problems where Databricks Spark might be the best perf/cost solution?

What about finding bugs, zero-day exploits?

Can this be used for interrogation technique?

Can the same argument/concerns be applied to all text generation AI?

Maybe it is a decentralised Ponzi scheme

a ponzi scheme requires that $ from ealier investors is paid out to later ones. bitcoin pays nothing out.

> a ponzi scheme requires that $ from ealier investors is paid out to later ones.

(the other way around)

A BTC is worth nothing unless you can sell it for something (such as USD). It is the later "investors" who buy BTC for sky high prices that allow earlier "investors" and miners to cash out.

Interesting counterpoint! How would we allay the repartee that buy-in is mining hardware and the payout is coin? It's not exactly cash-to-cash, but cash-to-hardware-to-coin-to-cash where the endpoints are still the same. Or is it arguable worse and more risky than a Ponzi scheme in that regard?

that is more analogous to selling mining equipment during the gold rush