Hacker News new | past | comments | ask | show | jobs | submit login

Being semantically right does not help against technically colliding. I can hook up an avian autopilot to a car to keep my bearing. While this is semantically a car with an autopilot, the contraption is useless on a road.

When I talk about autopilots, I'm talking about a system that allows me to disengage from steering. Literally by employing a self-steerer. The Tesla "Autopilot" does not permit this because I still have to closely monitor the trajectory at every moment. As such it does not fulfil the main expectation I have of an autopilot. What are your expectations of an autopilot?




I'm sorry, I'm terribly confused by your logic.

> hook up an [aircraft] autopilot to a car

This sounds histrionic

> semantically a car with an autopilot

??

> Literally by employing a self-steerer. The Tesla Autopilot does not permit this

This is not how Tesla autopilot works.

If you are genuinely asking my opinion, I would very much like to see this technology in the driver seat of more vehicles on the road as soon as possible. Where distracted drivers kill 9 people PER DAY in the US [0], if an autopilot system (speaking about the current one available today) is anything less than that, then it is well worth it.

[0] https://www.cdc.gov/motorvehiclesafety/distracted_driving/in...


Autopilots are reliable in aviation because they are simple. An autopilot for road steering cannot be that simple. The term transfers badly.

It is a false dilemma to say that we need autopilots to avoid road deaths. Because assistive tech is already successfully being used for exactly this.

We want self-driving cars to avoid the drudgery of driving. But the current batch of implementations needs very controlled conditions (Waymo), or close human supervision (Tesla).




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: