Hacker News new | ask | show | jobs
by lacampbell 3355 days ago
Automated driving on an industrial scale is still decades into the future. And even when it comes, arguing that it will be a huge, earth-shaking change seems far fetched.

Like I said, it's the general trend (we could talk about automations on the same scale as your human driver example that have already happened), but the alarmism is not justified.
2 comments
FussyZeus 3354 days ago
I agree it's decades off but I disagree that the alarmism is not justified, simply because we are not set up as a society to at all handle the side effects of the last good paying job that you can attain without really any high cost/long term training (truck driving) going away, and that's not even taking into account the multitude of businesses that were built on the backs of truck drivers; diners, motels, truck stops, service centers, practically any business that operates right next to a highway depends to a certain degree on truckers.

Plus truckers themselves are risky as workers, they have issues with exhaustion, the turnover is insane, they have accidents and they steal things, I'm not trying to demonize the profession don't misunderstand, I'm just saying that if there is one job that companies would LOVE to get people out of, this is it.

Also I really doubt it's decades. Maybe two of them, maybe but I'd say it's closer to 12-15 years depending how the DoT keeps pace, but with so much money and so many interest groups involved in making it happen, I think it's something we really need to start putting serious thought into.

link
willhslade 3350 days ago
It's really not that good paying of a job. True, gross costs are over 100-200K, but the wage net of maintenance, insurance and food is under 50K.

If I'm wrong, I'd love to be corrected.

link
darioush 3355 days ago
Decades is a long time. The internet is only 2.5 decades old and now you can click and have anything you want delivered to your doorstep. Your phone can tell you how to drive to avoid traffic. Your car already runs more software in it over a few minutes than the largest research universities could run in weeks. Image recognition has gone from "scanning digits on checks" to "outperforming humans on classifying images" (at least on some tasks) just in the past decade.

Which aspect of it do you think needs decades?

link