Live data from Hacker News

Pedro Domingos on the Arms Race in Artificial Intelligence

spiegel.de

11–12 of 12 posts

Re: Pedro Domingos on the Arms Race in Artificial Intelligence

#11

The following part of his answer deserves much more discussion. (If you disagree, but please say so why, rather than just downvoting.) We are allowing a very powerful system called capitalism to drive large multinational corporations that increasingly use simplistic objective functions to satisfy greed and often just greed alone. (Greed has its place, no doubt, but it should not be the overpowering force driving the…

Ok, I'll bite.

I think the issue with the quote from Domingos and your substitution is that regulation is a very blunt instrument.

He says he would like Facebook's algorithm to maximise not only engagement (which isn't how these systems actually work, but ok), but also the how truthful it is. Now you're left with the question of how to measure how truthful something is, and we don't know how to do that, so usually you come up with some metric that you do know, maybe you decide that authoritative news sources are truthful and things by your friends that talking about non-controversial things are ok, and maybe you throw in a few more heuristics and you get some ugly proxy for truth, how much weight do you put on it vs engagement?

The same question could be asked of how some regulation on capitalistic greed should work; though I would argue that the regulation we have on greed is taxes, where we can let companies optimize their profits, but then tax the profits (either at the company level, or at the individual level when the profits becomes real(ized) gains), and then we provide a whole host of institutions that are the ones we want.

I would say the challenge with regulation is not saying whether we should have any or not, but figuring out what good regulation would look like. Maybe the real regulation we need for Facebook is what we did to Microsoft (the other network effects behemoth) and force them to be interoperable with other platforms so that users could migrate without having to migrate all of their friends.

Re: Pedro Domingos on the Arms Race in Artificial Intelligence

#12

I don't quite understand the sudden optimism around the feasibility of strong AI. The only major advancement in the last two decades seems to be neural nets in pattern recognition. Before deep learning, pattern recognition wasn't even considered proper AI. The other advancements seem to be just based on better compute and not a new revolutionary idea. I hope people also get to hear from researchers on the other side…

Strong agreement w/ you. The Jordan article (https://medium.com/@mijordan3/artificial-intelligence-the-re...) is a good one to read.

On top of that, we should be having this discussion w/ or w/o AI -- letting algorithms of any kind make critical decisions in society is fraught with dangers large and small, and something we should do with eyes wide open and great deliberation (something seemingly challenging in today's political reality).

Post reply on HN