There's a lot of true believers who think Tesla+Musk will crack self driving and/or humanoid robots any day now.
Rollout both technologies at scale , and try to guess with one will cause more harm giving th fact there will be users in both cars trying to put legs on a steering wheel :
A stupid tech that will not even try to do safe things
Or software that is let’s say 4x less safe vs avg human but still very capable of doing maneuvering without hitting obvious walls etc ?
You get used with the system to work correctly and then when you expect less it does the unthinkable and the whole world blames you for not supervising a beta software product on the road on day 300 with the same rigour you did on day one.
I can see a very direct correlation with LLM systems. Claude has been working great for me until one day when it git reset the entire repo and I’ve lost two days work because it couldn’t revert a file it corrupted . This happened because I just supervised it just like you would supervise a FSD car with “bypass” mode. Fortunately it didn’t kill anyone , just two days of work lost. If there was the risk of someone being killed I would never allow a bypass /fsd/supervise mode regardless of how unlikely this is to happen.
Teslas has sensors , eye trackers etc is it possible to shoot yourself in the leg, sure. But not in any different way vs human doing irrational things in the car, make up, arguing , love etc.
Human-being is an irrational create that should not drive except for fun in isolated environment. Tesla or Waymo or anyone else.... It is good to remove human from the road, the faster the better.
I’m all for this but not to replace dumb people with dumb software. I think the FSD should be treated more like the airplane safety. We have the opportunity to do this right not just what’s the cheapest way we can get away with it.
When I'm driving FSD If i want to drink, eat, etc, instead of doing weird one hand tricks every driver did, i just turn FSD and let it drive. When I'm tired , I'm doing the same. Again , attention control works really good, it doesn't let you sit on the phone etc. unlike many other cars with less advanced features. You can't be on FSD + Phone but you can easily be on the phone + lane control in other car.
Phone is by far the biggest real killer of people, and no body is trying to create a campaign against phone mounts, etc.
If you're driving a vehicle in the former category, you'll be on the hook for reckless driving if you aren't fully supervising the vehicle.
I'm pretty sure the original commenter was supervising the driving, though.
How many have you tested in your day to day life?
Personally that’s way more useful for me even if they didn’t let you turn it on at highway speeds.
Which if anything looks worse for Tesla long term. If luxury car owners aren’t willing to pay 200$/month for self driving then trying to up charge people buying used model 3 and Y’s after canceling the S and X looks dubious. Which means that 100$/month subscription likely loses them money vs an 8k purchase.
They are looking to reintroduce it with a much higher top of 81MPH which might help, but agin my issue isn’t with the particular system but the underlying assumption of how much people value attention free driving.
That's kind of like saying 'look this algorithm is awesome' if we feed it all the data in the optimal order.
See this related article and discussion: https://news.ycombinator.com/item?id=47051546
No, you do not "become scary good" every single week the past 10 years and yet still not be able to drive coast to coast all by itself (which Elon promised it would do a decade ago)
You are just human and bad at evaluating it. You might even be experiencing literal statistical noise.
And the results speak for themselves.
https://www.gurufocus.com/news/8623960/tesla-tsla-robotaxi-c...
> Mr Keegan said he was “pretty confident” that in “the next five to 10 years” driverless vehicles would “make a major contribution in terms of sustainable transport” on Dublin’s streets.
As always, people were overoptimistic back then, too. There are currently no driverless vehicles in Dublin at all, with none expected anytime soon unless you count the metro system (strictly speaking driverless, but clearly not what he was talking about).
Tesla’s own Vehicle Safety Report claims that the average US driver experiences a minor collision every 229,000 miles, meaning the robotaxi fleet is crashing four times more often even by the company’s own benchmark.
https://www.automotiveworld.com/news/tesla-robotaxis-reporti...
The car cannot be drunk or high. It can't be sleepy. It can't be distracted. It can't be worse at driving than any of the other cars. It can't get road rage. So why is it running into a stationary object at 17mph?
Worse, it's very very easy to take a human that crashes a lot and say "You actually shouldn't be on the road anymore" or at least their insurance becomes expensive. The system in all of these cars is identical. If one is hitting parked objects at 17mph, they would almost all do it.
It's cheaper, that's all it is.
Scalability is usually about O(n²) vs O(n log n) or something, not a smaller constant that's significant but not a game changer.
They'll probably get there faster too because the decisions the camera+lidar model makes can be used to automatically evaluate the camera-only model.
It might a nice feature for your car to have. But most people aren't paying for it, the conversion rate is very low.
So they are not making money from taxis and not making much money from software sales.
So does it matter that for you personally it drives you around sometimes?
Even if you price in a 4x increase in FSD buy conversion ratio, you can't explain the stock price.
And I say this as a former Tesla investor who assumed that conversion ratio would be better then it is. But for that reason (and many others) I couldn't justify the valuation and dropped the stock.
I do wonder if folks who say Tesla’s FSD works well and safely are simply lacking a self-preservation instinct.
And of course, Musk's insistence that they don't need other sensor types like lidar or radar definitely looks like it's getting in the way.