Hacker News Reader

Comment by sneak | Hacker News Reader

I think it’s pretty obvious: sell the machines that replace humans as AI improves, and be ready to run that AI at massive scale.

First off is robotaxis and semis (99% of highway driving and 80-90% of city driving driver jobs have been very effectively replaced right now by machine, look at Waymo if you don’t believe Elon), and then Optimus for stuff that needs hands, like warehouse workers or other low-skill “move heavy things from here to there” jobs.

My six year old Tesla 3 just drove me from Vegas to LA, door to door, without me having to give it any control inputs. Not a lot of people seem to be paying attention to how good it is now.

I imagine Optimus will be a really big deal once they get it working as well as Teslas drive.

View on HN

I'm quite aware that the self driving capabilities of a Tesla is good, has been for a while.

So were is the revolution now?

Even the internal Tesla paper determined that the self driving market is not that big. There are clear issues no one seems to really like to hear:

People don't mind driving; A lot of people have to drive at the same time (which is not economic to catch this spike and then have all these cars with low utilization the rest of the day); We know the taxi/uber market.

Either it is so good that i literlay can do whatever i want while its driving and not getting car sick and move around, OR i'm sitting in a car seat anyway. I hate being not the driver, its a lot more boring than driving.

> First off is robotaxis and semis (99% of highway driving and 80-90% of city driving driver jobs have been very effectively replaced right now by machine, look at Waymo if you don’t believe Elon)

Waymo hasn’t achieved that, though, right? They’re deployed in minimal markets, and I would be surprised if they had even replaced a majority of taxi rides where they are deployed.

Your trip still requires you to be at the wheel and conscious because the technology can’t be guaranteed to not try to murder you. How many miles do you go between driving accidents? Really asking: how many miles do you go between intervening with self driving?

The lack of complete market takeover is a sign the technology just isn’t there yet.

Worth noting that Mercedes and BMW were the first to offer real L3 under specific conditions (as in: you could legally do something else, and the company would be responsible in case of accident), but they dropped it this year due to low demand and high cost. Some Chinese manufacturers offer it still.

Waymo has very little market penetration where I live. I've only seen a handful of them driving around. When I went to visit San Francisco last month, I never realized how prevalent they are.

I'd say that only some places have seen city drivers replaced by Waymo to such an extent. As for self-driving semis, the market penetration is near zero. Maybe the tech is good enough but no one is buying it.

Elon/Tesla haven't delivered on unsupervised self driving cars, how is Tesla/SpaceX going to deliver on robots replacing humans?

Your claim that robotaxis and semis have replaced humans is not credible. Show me all the truck driver job losses due to self driving trucks.

Once the Tesla semi is shipping in greater numbers, I expect a lot of trucker jobs to disappear. Not all, and not even most, but still, tens or hundreds of thousands. (IIRC there are 3 or 4 million truckers in the usa.)

Reply on news.ycombinator.com