News for hackers. Stuff that matters.
Re:Live Counter of Uber's Net Losses ($27B and Counting) by ghaff on Monday October 17, 2022 @11:17AM (#33234690)
I'm genuinely curious how those who claimed full self-driving would be a thing by today broke out between:

1.) Techno-optimists including those who saw the extremely rapid progress in ML across various domains after very little progress for decades. If we've made so much progress in 5 years, surely we can achieve almost anything in ten more even if we're not quite sure how!

2.) People whose self-interest was wrapped up in it happening so they just didn't want to probe too deeply.

3.) Those who knew it was probably flim flam, but hey whatever parts the rubes from their money

4.) Those who just figured that so many people who apparently believed in it can't possibly be wrong. (And there were a bunch of, especially, younger people who really wanted this world where they didn't need to own a car.)

Re:Live Counter of Uber's Net Losses ($27B and Counting) by rsynnott on Monday October 17, 2022 @11:41AM (#33235119)
I think it was mostly (1) and (4), with (1) feeding (4). There was a lot of "this time it's different" about the current AI bubble, though I think it's fading a bit, now.
Re:Live Counter of Uber's Net Losses ($27B and Counting) by ghaff on Monday October 17, 2022 @11:54AM (#33235327)
The non-cynic in me tends to agree. There has been remarkable progress. I was just in someone's Tesla "FSD" beta a couple of weeks ago and it was pretty impressive. It's also a long way from being able to be turned loose on arbitrary roads in arbitrary weather without a human in the loop. But it can be easy to dismiss all the work it can take to solve the "last 10%" of many problems. And, of course, there's a general mindset in tech to be positive about overcoming obstacles as opposed to being "it will never work" about some technology.

I do think today there's a better appreciation of just how challenging it is for AI to deal with an unconstrained physical world.