(no subject)

[elainegrey]

No comment is made in this article on the irony of the organization "Man Group" fostering an AI

At least, i didn't find it on my skim.

H/T Matt Levine's Money Stuff newsletter, which is filled with adorable snark about blockchain and silicon valley.

Comments

[weofodthignen]

Interesting. I didn't know about the 2010 Flash Crash (that doesn't look like a very good Wikipedia article).

I note that they disincentivize applying strategies that humans already use, so why are they surprised the AI does surprising things?

I am thrilled there might be less employment for MBAs in future. Computer scientists aren't much better but I don't think they've damaged academia as much.

M

[elainegrey]

I'm not sure, though, that i want MBAs running any more loose than they already are.

[amaebi]

I have read the article with interest*, but I don't understand the irony. Is it Man v. Machine, maybe?


* They don't specifically say, but I gather their approach is a price history approach and not a value approach. Hello, bubbles!**

** Let me note how much I want a Tobin tax on stock market transactions.

[elainegrey]

Ah, first, i am using "irony" in a more colloquial than literary sense, but yes, The Man group developing artificial intelligence.

It's hard to say, with the whole bit about black boxness. Patterns could reveal hidden values. The Bayesian statistics underlying a great deal of the pattern recognition are particularly good in picking up correlations such as large disaster being a signal that demand in a particular type plastic is about to go up.... well, that's a correlation that's a little more obvious, but it might be easy for humans to be connect the first few hops in a supply chain, but an AI could pick up a signal of a correlation in the a supply network that would be hard for a human to track down.

It's a pity gamblers aren't more patient... heh, that reminds me of Long Bets, and the featured bet is, perhaps, interesting to you: http://longbets.org/362/#adjudication_terms