> If Autopilot poses a significant risk to other road users, then an investigation is warranted.
Except there's no evidence that it does. I drive a Tesla. I've seen AP brake hard in circumstances where a distracted driver would have failed to notice. The system works. It does not work perfectly. But all it has to do to make that "if" clause false is work better than drivers do, and that's a very low bar.
Tesla themselves publish data on this, and cars get in fewer accidents with AP engaged. (So then the unskewers jump in and argue about how that's not apples and oranges, and we all end up in circles of nuance and no one changes their opinion. But it doesn't matter what opinions are because facts are all that matter here, and what facts exist say AP is safer than driving.)
At the risk of beating a dead horse, the issue is that AP engaged/disengaged isn't a random test, it's a user choice. Presumably people leave AP on when the outside conditions are inherently safer per mile (including, for example, on a highway). If Tesla randomly made AP unavailable some days to some drivers and ran a controlled test, that would be compelling.
You are absolutely beating a dead horse, sorry. This argument gets made repeatedly every time that report gets mentioned and... frankly I'm sick of going through it every time. Nitpicking against a result you don't like without presenting contrary data is just useless pontification. It proves nothing, and has no value but supporting your own priors vs. inconvenient evidence. Find some data.
I mean... there are two million of these cars on the street now, and all but a tiny handful of early Roadsters and S's have full featured autopilot. If there was a significant safety signal in all that data somewhere it would be very clear against the noise background now, and it isn't.
My point wasn't that I had contradicting data. My point was that the data Tesla publishes isn't valuable. It's a different statement. The premise could be right or wrong.
> My point was that the data Tesla publishes isn't valuable.
That's not true at all, and you know it. Consider the extremes: if autopilot got in zero accidents over these however many million miles of travel, you'd have to admit it was safe to use, right? If it got in 10x as many accidents as the median vehicle, we'd all agree it was a deathtrap. There's value in those numbers. Well, the truth is somewhere in between, it gets in 4x fewer accidents (I think) than the median car. That still sounds good to me.
You're just nitpicking. You're taking a genuine truth (that the variables aren't independent because "miles travelled with AP" aren't randomly sampled) and trying to use it to throw out the whole data set. Probably because you don't like what it says.
But you can't do that. The antidote to incomplete or confounded analysis is better analysis. And no one has that. And I argue that if there was a signal there that says what you want ("Tesla AP is unsafe") that it would be visible given the number of cars on the road. And it's not.
Stop nitpicking and get data, basically. The time for nitpicking is when you don't have data.
Yes, a very strong signal (everytime AP is turned on the Tesla explodes or everytime AP is turned on the Tesla drives perfectly from parking spot at origin to parking spot at desitnation, and there has not been a crash in 10 billion miles) can overcome the noise in Tesla's published data. However, any signal there is not that strong and is so weak it is drowned in noise.
Meanwhile, I do have this one datum about how OTA updates caused a couple of accidents (see OP's story.) You cannot just dismiss it because it doesn't agree with a larger trend, if that larger trend is undetectable in the data.
You're just making stuff up. That's not a weak signal at all, and you know it. The argument you should be making is that it's a confounded signal. But that requires a level of sophistication that you aren't supplying (nor is anyone else). Which is exactly why I mentioned way upthread that these digressions were so tiresome.
There's literally no data in the world that will satisfy you, just admit it. Yours is an argument from priors only, to the extent that one anecdote ("A Tesla hit me!") is enough to outweigh a report like that.
Tesla themselves publish data on this, and cars get in fewer accidents with AP engaged
This is false, or at very best misleading, considering that Tesla's Autopilot is responsible for more than 10 fatalities. Or in other words, Tesla's AP has killed more people than the entire rest of the domestic auto industry's smart cruise control features. Combined. (And Tesla's "safety" data conspicuously fails to mention any of these fatalities.)
And that doesn't even include all the times that AP ran into stationary emergency vehicles, or the many slow-moving accidents in parking lots, or the many near-misses that were avoided because the other driver was able to avoid the suicidal Tesla in time.
Come on, the "entire rest of the domestic auto industry's smart cruise control features. Combined." constitutes what, 20k vehicles or something like that? It's just in the last few months that we've seen lanekeeping features start to land in volume models, and even then they tend to be extremely heavily geofenced. It's very hard to kill someone when no one has your product and you won't let them use it.
The entire rest of the domestic automobile industry sold millions of cars with smart cruise control features similar to Autopilot over the past few years, so the misleading one would be...you?
Unless you're now claiming that FSD is just a version of smart cruise control?
No, using vanilla cruise control, almost nobody dies every day. It is in fact rare enough that when it happens it gets reported in the news.
The fact that Tesla fans can't grasp the concept that Tesla is (a) not perfect and (b) deliberately endangering people's lives for profit is unfortunately not shocking any more, but it doesn't mean the rest of us have to put up with that schadenfreude.
>> If Autopilot poses a significant risk to other road users, then an investigation is warranted.
> Except there's no evidence that it does.
It sounds like you two are talking past one another.
I don't believe op is presenting evidence that it does. Instead, they are saying that there is at least one antidote (TFA) and that someone (presumably a government agency) should investigate it.
First, a very valid reply to there being "at least one anecdote (<-- sp)" is that there is a bunch of actual data on exactly the subject at hand that shows the opposite of the stated hypothesis. You don't throw out or ignore existing data just because one guy on HN got rear-ended!
Second: I think you're being too charitable. I think it's abundantly clear that grandparent believes that there is a "significant risk to other road users", and that this single data point confirms it. And clearly it doesn't.
Except there's no evidence that it does. I drive a Tesla. I've seen AP brake hard in circumstances where a distracted driver would have failed to notice. The system works. It does not work perfectly. But all it has to do to make that "if" clause false is work better than drivers do, and that's a very low bar.
Tesla themselves publish data on this, and cars get in fewer accidents with AP engaged. (So then the unskewers jump in and argue about how that's not apples and oranges, and we all end up in circles of nuance and no one changes their opinion. But it doesn't matter what opinions are because facts are all that matter here, and what facts exist say AP is safer than driving.)