vihart
patreon
Changing my Mind about AI, Universal Basic Income, and the Value of Data
🕑 Added 2019-05-30 22:27:23 +0000 UTCHello good humans!
I've been in heads-down work mode for the past two months writing this piece: https://theartofresearch.org/ai-ubi-and-data/
It's about AI and the value of data, and why the pairing of AI and Universal Basic Income is attractive on the surface but has some deep problems (particularly regarding general respect for humanity). I also wanted to explain the motivation behind some of the work I've been doing with M Eifler and Jaron Lanier.
The piece seems to have kept growing and growing, needing just a little more research on this or that (60 footnotes!), and I'll admit I've been working more hours than I should have recently. I've also been ignoring anything else involving typing (emails, posting here) partly because I kept allocating all my time to the piece (in constant "just one more push and it'll be done" mode) and partly because all the typing means no hands left for other typing tasks. I gotta give them a break now.
I'm curious what you think. With such a long piece covering such big topics there's sure to be things I missed or got wrong, but hopefully it's at least useful and interesting.
I'll be back with more updates before long! Thanks for sticking around.
Vi
Comments
Hi, wanted to revisit this piece but it seems like the art of research page is down. Can you maybe repost it on your website?
Super interesting article, full of new-to-me ideas that seem like they’ll be very important! Also, I’ve been hearing about Jaron Lanier for a while and I feel like this was a great introduction. The one thing that was a bit frustrating to me when reading this article was the assumption that we’ll never created artificial general intelligence. I think it’s a perfectly fine assumption to make, but I would have liked to have seen it briefly addressed, instead of just stating that AI will never be able to, for example, read human language, perceive subtleties in meaning, infer or look up slang usage, and judge whether the content is appropriate. These tasks are only possible with AGI, which is probably far away, and maybe impossible, but I think it’s a big enough unknown that it’s worth at least saying, “I’m assuming for the sake of this article that we don’t develop AGI anytime soon.” Anyway, sorry for nitpicking, and thanks for writing this.
