AI winter is well on its way (2018)

blog.piekniewski.info

12 pointsdarkhorse136 comments

The formula is so consistent: cherry-pick some disappointments, interpret any pivot as retreat, claim the hype is unsustainable, predict imminent collapse, position yourself as the sober realist. Rinse and repeat every 6-12 months, adjusting the timeline when the previous prediction fails to materialize.

His prediction of self-driving cars being a fiasco does not need any adjustment, and it stands after 7 years.

Living in LA where there are plenty of Waymo cars constantly driving around, I strongly disagree with that.

Yes, 400 of them.

https://en.wikipedia.org/wiki/Waymo#Robotaxis

LA has more than 6,000,000 cars.

That is how strong you can disagree with me. 400 out of 6,000,000. And I'm being kind by considering just LA.

What counts is not the total numbers of cars today but the growth rate of this number.

Ten doublings is 1000x

Waymo was founded in 2004. That's a very slow growth rate.

So far, the prediction made in 2018 stands. Self-driving car companies lose money, and are a fiasco.

Maybe that will change in the future, but that's the future. We'll talk about that then.

Reply on news.ycombinator.com