Been using FSD-b since two years ago when it was being made available for a somewhat larger crowd of no-kidding Beta (some would argue Alpha) testers.
We have two Teslas; one, a M3 for the risk taker (that would be me) and a MY for the SO. She hasn’t been thrilled with the zaniness and need for strict attention that driving with FSD has required, but she does like the lane changing capabilities on highways, where the car’s many eyeballs make changing lanes much safer. So, the MY has EAP.
In the past, FSD-b has required constant attentiveness when doing city streets since, on a pretty regular basis, it would attempt to kill itself. Running red lights, not stopping at stop signs, blocking other traffic when making turns, and so on.
However, as time has passed, FSD-b has gotten better. It used to be 5 or more interventions a mile; now it’s down to 5 in, maybe, ten miles. It’s kind of random in that regard; I’ve made 20 mile trips on local roads these days without a single intervention, but that’s definitely rolling the dice. And, compared to how it used to be, its current capabilities are impressive.
But, as many have pointed out, impressive does not mean error free, and to lose that Beta designator it’s going to have to be better than a human. Which is asking quite a bit.. but humans are definitely not error free, either.
I’d much rather drive around with FSD-b available than not. One gets a definite feel for where the software will run into trouble, and, in those places Just Take Over and avoid the stress. On clearly marked roads it does a kind of relaxing job and one can simply supervise. But the learning curve to get to that understanding is a bit steep.
The Tesla crowd is currently wrangling the next major FSD release into some form of safety/stability before releasing it to the general crowd of testers. They’ve made hopeful noises that they may get the intervention rate down by a factor of 100. Which would be astonishing. That remains to be seen, of course.
So, it’s still a Beta. A much better Beta than it used to be. But not quite ready for Ma and Pa Sixpack. If you’re interested in helping the effort along and having the shiny nifty stuff, sure, buy FSD-b, but plan on some time learning its vagaries.
And the whole blame thing is, to some extent, a research project. Tesla does think that they’ll deliver an ADAS for sure and are clearly aiming at a Robotaxi. But, when plowing virgin forest like this, nothing is guaranteed.
Last edited:
