Practical Typography is really good.
A short history of science blogging.
The evolution of popular music: USA 1960–2010.
AI stuff
The Economist’s May 9th cover story is on the long-term future of AI: short bit, long bit. The longer piece basically just reviews the state of AI and then says that there’s no existential threat in the near term. But of course almost everyone writing about AI risk agrees with that. Sigh.
6-minute video documentary about industrial robots replacing workers in China.
Bostrom’s TED talk on machine superintelligence.
PBS YouTube series It’s Okay to Be Smart gets AI risk basically right, though it overstates the probability of hard takeoff.
Sam Harris says more (wait ~20s for it to load) about the future of AI, on The Joe Rogan Experience. I think he significantly overstates how quickly AGI could be built (10 years is pretty inconceivable to me), and his “20,000 years of intellectual progress in a week” metaphor is misleading (because lots of intellectual progress requires relatively slow experimental interaction with the world). But I think he’s right about much else in the discussion.
I look forward to looking at the Nick Bostrum Ted talk on Sunday when I have free time.(At least he is discribed as a philosopher.) But I had to lol at the idea that a computer might develop values. That would take a miracle.
I can see why robots taking over basic manual labor jobs is a threat but the Nick Bostrom and PBS videos seem like Y2K hysteria.