NTSB chair calls on Elon Musk to change design of Tesla Autopilot | Hacker News Reader

NTSB chair calls on Elon Musk to change design of Tesla Autopilot

techcrunch.com

3 pointsedward12 comments

In a perverse way, I wonder if there is a first-mover disadvantage for Tesla. Wonder if someone could draft in the wake of their approach, and come out without so many scars.

Waymo is essentially the first mover and has no scars at all because they realized that any consumer system that requires constant driver attention and allows override is unsafe.

You can tell someone to pay attention and hold on to the wheel while the software is controlling the car, but we know humans don't do that.

Waymo is sure about their thing, but sometimes corporate America looks more like high stakes poker than a scientific lab or a scientific publication.

Sometimes bluffing and faking it till you make it...that makes you win the race.

> Sometimes bluffing and faking it till you make it...that makes you win the race.

I think that works in a lot of cases, yes. I would need to be convinced that SDC is one of them.

If you think of the state of AI (and anyone who's used Google Assistant or kept up with IBM Watson knows that it essentially doesn't exist), SDC is such a massive breakthrough that it's hard to imagine someone faking it until they make it. It's like saying someone is going to fake their way to practical fusion reactors (exaggerating, but you know what I mean).

Tesla with an inferior product will have much more money to throw at the problem vis-a-vis Waymo .

Also everybody wants to work at Tesla , not so much Waymo.

Unless you think there is a correlation between the IQ of people who are turned off by Musk and Tesla hype antics (and "faking it till you make it" kind of process) , then there is reason to believe that Tesla is in a better position to attack the problem and not Waymo.

> Unless you think there is a correlation between the IQ of people who are turned off by Musk and Tesla hype antics (and "faking it till you make it" kind of process) , then there is reason to believe that Tesla is in a better position to attack the problem and not Waymo.

I certainly think that the smartest and most ethical engineers in the SDC space are turned off by Musk and Tesla.

I'm sure there are plenty of unethical engineers, like Anthony Levandowski, who will do whatever Tesla needs, though. After all, Facebook still employees many of the best engineers in software.

Let's hope this is the case.

In my experience it's very rare to find a person that has a strong overlap of book smarts and street smarts

These people idolize Musk, just look around in here.

> Tesla with an inferior product will have much more money to throw at the problem vis-a-vis Waymo .

>Also everybody wants to work at Tesla , not so much Waymo.

Highly doubt both of these. Tesla has a bigger budget total than Waymo but Tesla's FSD team is probably smaller than Waymo. Plus they have much more access to compute resources and AI expertise, Alphabet's R&D spend on AI is probably close to Tesla's entire revenue

But Google doesn't have just Waymo in their portfolio.

Besides Musk is willing to push the company on the brink of bankruptcy and is able to convince Wall St. to allow him to do so.

Larry & Sergey won't ever push Google on the brink of bankruptcy just to make Waymo a success.

That's a good point. My counter to this (not saying I endorse this) is that Tesla has redefined what people's expectations of a SDC are, both in terms of safety and capability (neither in a positive sense IMO). So "self-driving" has become more of a fluid concept than maybe you/I would envision.

Technology probably, but product I would say no. I prefer the cautious option, but the general perception (dangerous word to use in this context) of an autonomy product is driven by Tesla, not Waymo.

Significant improvements have been made that make NTSB recommendations null. Frankly, Homendy appears to be several years behind Tesla's advancements in Autopilot and FSD. The NTSB may be adding more risk with their recommendations due to likely human circumvention and willful ignorance of data.

Reply on news.ycombinator.com