Good. It seems exceedingly unlikely that this is a road-ready technology, and we don't need yet another software-brained megalomaniac releasing an insufficiently-developed, unproven product into the physical world, where "bugs" will actually just straight kill you.
George Hotz cancels his Tesla Autopilot-like ‘comma one’
451–460 of 498 posts
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#452Earlier quoted context omitted.
The marriage license argument is textbook strawman. That is not regulation of an industry, and you know it.
I agree, but you proposed a stringent definition for 'Rule of Law' that seemed to eliminate the possibility of government regulators possessing independent decision-making authority over citizens and businesses alike. I mean, how do even mundane things like zoning codes and work permits work in this universe where there are no regulators and gov't bureaucrats? Everything must go through a court? Why are you so sure a…
> how do even mundane things [...] work in this universe where there are no regulators and gov't bureaucrats?
I talked about regulators. You added bureaucrats on your own.
Perhaps NHTSA is a great organization. I know nothing about them specifically. I'm talking about general principles.
> Why are you so sure a group of 12 random people will come to decisions more effectively than a bureaucrat in a domain that requires specialized knowledge?
I'm saying they'll be more impartial. Don't know about effective.
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#453Earlier quoted context omitted.
If 10,000 people were killed on the way to making driverless cars safe, and this wild west approach ended up getting them into mass market use a few years earlier, it would save the lives of thousands of people overall. Is this a bad thing? Is it not justified by utilitarianism?
> If 10,000 people were killed on the way to making driverless cars safe, and this wild west approach ended up getting them into mass market use a few years earlier, it would save the lives of thousands of people overall. This is a perfect example of the biggest and IMO most compelling critique of utilitarianism -- namely, that it's too susceptible to bullshitting. People decide what they want to do, then completely…
So what do you propose as a guiding principle in place of utilitarianism?
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#454Earlier quoted context omitted.
It's not a great name, but I'd give him the benefit of the doubt and assume it's "all things AI". Think classification language, like "Car, Self-Driving". They could be "Car, AI" and "Assistant, AI". Less charitably, it's step one toward Tres Commas.
Given that he referenced the HBO show in his Techcrunch presentation, I'd say the reference to the three comma club is probably the most likely. Also, one comma would be $1000, which is what the thing was supposed to cost.
I don't know if it was ever intended as a consumer facing name, but even without the "coma" puns, it's too wonky for customers, do you think?
Besides, isn't one comma 99k worth of room? It's marketing lingo aimed at the hubris of venture capitalists.
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#455Earlier quoted context omitted.
I wish it were possible to buy a Tesla without any autopilot features except basic cruise control, because I know how damnably difficult the edge cases are for a self-driving car, and I'm perfectly capable of driving a car manually, thank you very much. Tesla's cars are so good I'd almost be willing to pay more for one without autopilot. Almost.
While driving, you have to do a double-pull on the lever to activate the autopilot, a single pull just activates the basic cruise control. So you can still buy one and just never activate it. Alternatively, earlier editions of Model S shipped without autopilot, which was then sold as an aftermarket upgrade as late as this spring (when autopilot was an optional $2,500 package on a new Model S).
Fair enough re: the older Teslas. As to the new ones, even if I never activate autopilot I'm still driving around in a massive autopilot data collection engine for Tesla Corp, which bugs me.
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#456Earlier quoted context omitted.
I have a 100% perfect safety record over probably hundreds of thousands of miles of driving a class B truck with my knees instead of my hands, using cruise control, on straight empty highways in Nevada, Idaho, and Utah, with perfect weather and good road conditions. There's nothing revolutionary about my ability to perfectly drive with my knees...I just very carefully selected the conditions where I was willing to do…
You may have that perfect safety record, but other people don't. People still crash every day in good conditions. That's thousands of lives that could be saved with our imperfect, sunny weather on the highway only self driving car. You underestimate how bad human drivers are even in perfect conditions.
You are completely missing the point, so I'll try explaining via a Reductio ad absurdum hypothetical.
Lets say that Semi-Autonomous Cars are currently tested on about 80% of the tasks that humans currently face in the real world, and that through some feat of engineering and design we don't have to worry about the ridiculously messy transitions between Autonomous Mode and Human Mode.
And let's say that for those 80% of tasks, the Semi-Autonomous Cars have a 0.05% accidents/100k miles incident rate.
And let's also say that Humans on average have a 1% accidents/100k miles incident rate.
What you're telling me is that the Semi-Autonomous car has a much better safety record, so give us Semi-Autonomous RIGHT NOW OR WE'RE ALL GONNA DIE!!!
But what I'm telling you is that you don't know enough to make that decision yet, because you don't know exactly how well humans do on the 80% that Semi-Autonomous cars currently handle, you merely know the average accident rate over the current 100% of scenarios. For all you know, Human drivers could have their average 1% accident rate as a result of a 0.0% accident rate for that 80% subset and a 5% accident rate on the remaining 20% that Semi-Autonomous cars can't handle. And if that were the case, then forcing us all to use Semi-Autonomous cars would actually increase the average accident rate from 1% to 1.004%.
Until you fully understand what Semi-Autonomous Cars are capable of, AND know how well human drivers do on that restricted subset, you can't definitively say that current technology is better than humans.
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#457Wow. If he really moves his project to China to take advantage of a lax regulatory environment even though he couldn't pass the scrutiny in the US, he's being very immoral and cavalier about people's lives in China. I don't understand how a startup like Cruise can get past this regulatory hurlde in the US but Hotz can't.
Not to mention pretty risky for himself. If his invention kills someone in China, he will probably end up wishing he was dealing with US legal system (and/or US prison).
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#458I find it interesting that he was on This Week in Startups only a few weeks ago calling every company he found unconvincing "losers" https://www.youtube.com/watch?v=2zy_07g2IrM Maybe they're all losers and you're a winner, but perhaps being an elite hax0r isn't all that's needed to go to market... If you can't reply to a regulator letter, how are you going to deal with supply chain or cash flow issues? Bad reviews in…
He sounds really immature and irresponsible. If you look at his code on GitHub, it's full of sloppy stuff, commented-out code, and worrisome little loose ends with comments like "should I check this value?"[1] Even his commit messages are full of sloppy stuff.[2][3] That's fine for hacking on game systems, emulators, or various consumer electronics (and he's clearly quite good at it, given his accomplishments), but would anyone really want to trust their life to code like this? How much testing has he actually done on this thing? By his own accounts, it didn't even work until less than a year ago.[4] The NHTSA is right to question him about safety if he's planning to put this on the market for anyone & everyone to use on public roads a matter of months from now. Choosing to shut down and go work in another country after receiving just one inquiry about safety is a sign that there could be serious problems. A car is not a toy.
(Oh, and in case anyone responds with "well let's see your code!", please note that I'm not the one making asking people to trust their lives to my code, making wild claims about what I've built, nor issuing challenges to the likes of Tesla.[5])
[1] https://github.com/geohot/kvm-kext/blob/master/main.cpp#L642
[2] https://github.com/geohot/kvm-kext/commit/082b7ca99cba4c3b9c...
[3] https://github.com/geohot/kvm-kext/commit/96441be079562b0dd0...
[4] http://www.bloomberg.com/features/2015-george-hotz-self-driv...
[5] https://electrek.co/2016/04/06/tesla-autopilot-comma-ai-geoh...
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#459Earlier quoted context omitted.
It's the other way around - you need a proof that they perform comparable to an average driver during expected use, and they do not have absurdly stupid corner cases during actually expected use.
Well, humans have absurdly stupid corner cases during actual use, too. It depends on how small the corner is, I think.
We've been dealing with those corner cases for thousands of years and we know them pretty well. Given that we all run on pretty much the same hardware and firmware (with minor differences), you could say humans have been thoroughly tested for few millennia, including a hundred or so years of road testing.
So yeah, I a bit of thorough testing of a completely new hardware running completely new software isn't too much to ask for.
Re: George Hotz cancels his Tesla Autopilot-like ‘comma one’
#460Earlier quoted context omitted.
To get that certification, do they test you first, or do they wait for you to screw up and then take it away from you??
We know, being humans, and that we have the ability to process lots of complex information in a way that's very difficult for computers to replicate. Hard AI doesn't actually exist (yet). We also have a 100 years of humans driving cars worldwide so we understand well what they're good at and what they're not, so laws & safety designs take all of this into consideration. Each computer system will be encountering new,…
One of the features of the common human firmware is self-preservation instinct. It lets us trust that our fellow drivers, while still prone to mistakes, won't generally make obviously suicidal errors. Can one say the same about a new ML algorithm running on some board designed half a decade ago? How exactly would one know, without a thorough audit?