Hacker News Reader

Comment by yellow_lead | Hacker News Reader

I'm not sure I would accept $200/mo. to drive this thing on the highway or roads with speed limits above ~30mph.

Conversely, we’ve bought several Teslas and have covered tens of thousands of miles on Autopilot, usually at 70mph (sometimes faster if the speed limit allows for it [Texas]). No issues.

In fairness, survivorship bias would apply to your anecdote - quite literally.

There are over a million Teslas on the road (and close to that amount being sold annually at current run rate) and only 7 Autopilot related deaths. Regulation and oversight is important (crucial even considering life safety), but the naysayers are crying that the sky is falling based on the statistics.

Ignore my anecdote, the data speaks for itself. GM ignition switch deficiencies GM covered up killed orders of magnitude more people (124).

https://www.tesladeaths.com/

GM recalled almost 30 million cars with those ignition switches. You're 1.5x more likely to die using Autopilot than one of them.

I still like those odds! As the saying goes, “I’d rather be dead in California than alive in Arizona.” I’ll take my chances in a Tesla over a legacy automaker.

More power to you but how do you feel about endangering others around you?

I believe the risk is reasonable compared to your average human being. Software doesn’t get tired or distracted, and I would not use Tesla self driving features if it would put others at an unreasonable level of risk.

>> Software doesn’t get tired or distracted

It just has a lot of corner cases, untested even. This blind trust in SW and by extension, technology, is really exasperating.

Really curious, how would you quantify the reasonabl risk ? Should the risk be based on number of fatalities or on the number of potential fatalities that were averted ?

The data that you point to is about fatal crashes that made the news. What about all the cases that the SW malfunctioned and had a potential to be fatal ? Only tesla has that data or perhaps even they may not.

Bought a Tesla last year and have driven it on two 4k mile road trips, often on autopilot for large portions of rural America. I've had multiple incidents of autopilot doing crazy things, particularly in areas where (I assume) not as many Tesla drivers go (Wyoming, Dakotas). My car abruptly tried to drive itself off the road, pushed me towards construction cones and concrete barrier, have come very close to weaving semis, phantom braking, etc.

99% of the time it's great, and road trips on autopilot are so much more relaxing than in my previous ICE car. But that 1% is eventually gonna kill me via car crash or via heart attack.

I do not recommend using driver assistance features you do not feel comfortable and confident in supervising. I appreciate the data point regardless. I have not experienced these issues out West or the Southwest US.

I took my Tesla on a road trip from AZ to the Texas panhandle recently. On the 2000 mile trip, autopilot disengaged four times. Once during sunrise on a curvy section of highway with very distinct shadows. Once during sunset, driving into the sun. Twice more when close to semis. Had I not been paying attention, each could have resulted in an accident.

I really only feel comfortable on autopilot during rush hour bumper to bumper traffic.

Only when Tesla can be sued by my insurance company if the car gets into a crash while on self-driving, will I trust that the appropriate due diligence was applied.

That 99% goes down using it in northen winter climate, or rural areas.

Reply on news.ycombinator.com