Hacker News new | ask | show | jobs
by fairpx 2989 days ago
Autopilot tech in general should be open and shared among ccompanies and universities. The more brains we can put on this the better imho. If these things happen too often, Im afraid regulators will start making laws that prevent the use of this tech, which I believe has the potential to really make life better for all... one day.
4 comments

I have always wondered why each company is developing their own software and algorithms... How will NTSB certify the software? Or will they? I mean, they certify car safety features now through crash tests, roll over tests, etc... Why is Tesla, or anyone manufacturer, allowed to modify vehicle safety features through OTA updates without recertification?

I realize it adds more regulation...and I'm sure there can be a middle ground. But I foresee a scenario where an OTA update is pushed and the next morning there are car crashes everywhere.

Patents and the ability to earn money by licensing out technology. That, or hoping they are the best and get more customers through it - I think that's currently the case for Tesla, selling the only commercially available semi-autonomously driving car ATM.
All manufacturers sell semi-autonomously driving cars. At a minimum cruise control, but more advanced features like e.g. https://www.toyota.com/safety-sense/

Tesla sells a feature set that sounds slightly more advanced on paper, but in practice the benefits are pretty comparable.

> I think that's currently the case for Tesla, selling the only commercially available semi-autonomously driving car ATM.

"I think that's currently the case for Tesla, selling the only commercially available semi-autonomous [electric] car ATM."

FTFY

Cadillac has a similar system out in production.
The regulators have been really easygoing so far with automated cars. I strongly suspect its because they take the optimistic comments of the tech industry at face value that these technologies are road ready. Meanwhile, I think Tesla and Uber perceive this as more of an open alpha. This will not end well.
For someone in the medtech space, it's been interesting to see how loosely (not at all?) marketing is regulated in automotive and tech in general. Were the FDA regulating automotive, there is absolutely no way that Tesla's driving assist package could be marketed with the name "Autopilot", for one. More broadly than just marketing, it's also amazing to me that Tesla could have released a feature (I know they've revised it since) that was designed for human supervision but did not require hands on the wheel for long stretches of time. What did the risk management look like there?
I think (at least in the US), the relationship the automakers have with regulators is such that if the automakers think they can make money by selling a feature, they will be able to get the regulators to come up with a way to allow it.

Not necessarily one as simple as OTA updates whenever the manufacturer feels like it.

As far as regulations preventing use of the tech, if level 2/3 systems end up causing more crashes than human drivers, we shouldn't allow use of them.

I would guess people would just stop using it
Serious question: how will it make life better for all? I haven't given that question a lot of thought, and I haven't heard legit answers from others yet.
If autonomous systems reach the point where they are measurably safer than human drivers, there's an obvious benefit to using them.

Again it's an if, but lots of people drive long distances and don't enjoy it. An autonomous vehicle would relieve them of the driving task. It would also add a transport option for people that are unable to drive long distances.

It's likely to reduce the cost of taxi like services, as drivers are currently a significant portion of those costs (autonomous driving turns pricing into almost a pure calculation about return on investment).

I guess you might dismiss those as not being "legit" enough since they are all contingent on the systems working well.

No, I'm with you here. Just thinking through the other side of some of these. Lost jobs. Not sure "safety" is a big enough issue here. On a larger scale, I'm significantly concerned that society (thru tech) is becoming increasingly optimized for comfort and safety. That bothers me. It removes so much of what is good about life. I don't want guaranteed life until 100. I don't want comfort all the time; that stunts growth. There is a lot of grey area here, of course. But this has been on my mind for a while now.
Losing truck driving jobs isn't removing what is good about life. Neither is too much safety. The people whose lives are being saved by technology aren't worrying about having too much safety. Only those who are privileged and don't have to worry about anything, realistically.
Living in a city where people park on the side of the road, I can't wait for autonomous cars to completely transform the urban space.

Currently, there are two lanes of parked cars, plus two sidewalks and (usually) two lanes for traffic.

Autonomous driving would allow cars to park themselves off-site, such as underground garages a block or two away. That alone could double the space available for pedestrians, or other uses of public space such as outside cafes, green spaces etc.

The ability to quickly call a vehicle, adaptable to your current needs (minivan for family on Saturday, open two-seater for a summer date) could also start a major move away from car ownership, reducing the total number.

Improvements in traffic management (shorter distances between cars, coordinated starting/stopping, interleaving intersections) also have the potential to allow us to reduce 4-lane roads to just two. We might even find ways to make one-way streets work for most residential area. All this again frees up public space.

All valid points, thanks
If done well (long way to go still) there will be less traffic accidents, less traffic jams, etc.
Fewer dead people sounds good.
I don't know about that. Let's talk unemotionally about it. People die. Not all of them live until natural death. That's ok. That's how life works. There is plenty of good that comes from this. Is a goal of keeping every person alive until X# (90? 100?) a good one? I'm not sure it is. Simply saying "people dying is bad" isn't a very reasoned thought if you take away the assumption that everyone agrees with this.
Sure, fewer people dying is a bit undifferentiated if you want to go for ethical discourse. I would probably be okay with a few extra dead people if that increased the overall happiness of the remaining people. I'm not sure that having people die in traffic is a good way to achieve that.
I get that, but this product development aimed at prolonging life, comfort, and safety through technology should not be taken lightly without this discussion right here (ethical discourse, as you call it). I can't divorce the two in my mind.