As I said in another thread, this cybercab/robot event was glitzy big-budget stuff but basically content-free, like being on the set of a sci-fi movie. If it was supposed to look impressive, it succeeded. If it was supposed to inspire confidence that anything of substance will be delivered any time soon, it failed to do anything remotely like that. I’m only surprised there wasn’t a little set showing what a colony on Mars will look like when it’s established next year (in the immortal words of Bullwinkle the Moose, “this time for sure!”).
I expect the first Cybercabs will be geofenced, just like Waymo. They’ll work on a few hand-selected routes that are known to work. FSD is already good enough for that. And the licensing will be much easier in that case.
What they showed has almost nothing to do with FSD itself. That’s a software problem that will be solved on all cars simultaneously. It doesn’t have any features that make it more suitable for FSD than other models.
The important aspect is that they have what looks like a final design for a $30k EV. The self-driving aspect is only important inasmuch as they saved a few bucks on the steering wheel. You generally don’t put together 20 prototypes unless you’re pretty confident in the design.
Incidentally, it looks almost exactly as it did in the Walter Isaacson biography. So not much actual surprise. One bit that came up is whether they plan on having a model with a steering wheel. Franz and the others eventually convinced Musk that they need to at least have the possibility of a steering wheel–if for no other reason than that certification in Europe will take a long time, and they’ll want to sell cars in the meantime. The book alleges that they did convince Musk of this. But things change…
And when they sell that car for 30K at a profit, let me know.
ETA: Probably a hijack, so maybe take it elsewhere if they ever do that.
If they don’t have a working model, then they don’t have a $30K car. The projected price has as much validity as Musk’s promises of delivery dates
What do you mean? They have 20 working models. They’re still prototypes, obviously. And they only self-drove around the parking lot. But they’re functional vehicles.
And while I shouldn’t have to point out the obvious, I did not say they have a $30k EV. I said it looks like they have a final design for a $30k EV.
I have no idea why you trust Musk when he says it will be $30K. I don’t believe him.
I’ve no idea what you think I’m claiming. They showed a car that is intended to be their $30k EV, self-driving only. They had 20 instances actually driving around. And we also know from the Isaacson book that it’s been a serious project for a long time.
What it actually costs when it ships is a completely independent question that I have no hard position on. Though the Model 3 did actually ship a $35k version in small quantities–living up to their promise, though only barely so.
Getting back to more concrete FSD news…
Just tried it out. Not so great. Had the first real safety intervention that I’ve had in a while. The left-turn lane problem I mentioned a while back is even worse! It didn’t just move into the lane at the last minute; it actually tried to turn from the lane going straight. Really not great.
And again it missed the turn for my condo. That’s a regression with 12.5.x in general, though.
On the upside, it is smoother at low speed. I wouldn’t take that trade, though.
You are claiming they have a design for a $30K car, but I guess I have no idea what that claim means if you say the cost when it ships is completely independent of that.
The only valuable information they could provide is what the car will cost to consumers. I guess we agree that we have no idea what that number will be.
I’m curious – can you regress to an older version that you liked better? I mean, is it like the Android model where you take an update and you’re stuck with it, or like the model of a real computer where you can control software versions?
the mechanical south african?
… Or his promises on price …
Are you sure that was 12.5.6? I don’t think that’s been released to the general public. I think you must have tried 12.5.4.1 which is where most people are now and have gotten in the last week.
That said, 12.5.6 has the supposed big highway improvements so what you experienced may not change from 12.5.4.1. The 12.5.6 release should be the last one for FSD 12.x. Per Elon FSD v13 will hace 7x fewer interventions than v12 and will be released in October (lol).
You are stuck with the new version. No going back.
so, its smart to sit on the fence an monitor the interwebs for feedback from early adopters before jumping the gun
Some people do that, just like with the iPhone. Most don’t bother. They internally test first. Then they release it to employees and very long time customers and bloggers/influencers who like to be on the cutting edge. The bloggers will give very comprehensive reports. This group tests the hell out of it including taking the car to known tricky areas. On occasion it won’t get past this point.
Next it rolls out over a week or so to people like me who have their software settings on “Advanced”. We’ll have it for a while before it goes to people without Advanced by which time it’s been thoroughly vetted. You can still decline to install it even at this point.
All software updates have an associated FSD version. Some software updates will keep the same associated version and are for the UI or adding other features like a new available music streaming service. A quick google indicates that only 19% of Tesla owners are FSD subscribers. Some of them just get it for one or two months a year when they are doing a big road trip.
-haj, Tesla expert since March
More to the point… who cares? Tesla showed off last night that they can re-body an existing platform, something that all car companies do all the damn time. It’s a Cybertruck inspired Model 3 coupe with a made up price drop and an autonomous promise that nobody believes, because Tesla has no credibility anymore. I’ve been reading automotive publications for 30 years and any other company that rolled out so little in a major press event would have been thoroughly lambasted. Even though people have all caught on to the BS, Tesla still somehow gets special treatment.
I’m also not sure how anyone thinks these FSD predictions even work. Like, if you’ve got engineers working towards a common goal, even that’s hard to give timelines for, but we’ve got decades of experience trying. What are any of these AI companies doing anymore, feeding larger and larger training data sets into larger and larger neural networks and then documenting what comes out? None of this is predictive, we’re at the mercy of the technology now. Neither Elon nor anyone else has any clue what could possibly happen in the next 2 years that would make this strategy suddenly start working. Maybe it will, maybe it won’t, maybe it’ll be 2 years, maybe it’ll be 30 days, maybe it’ll be 30 years.
Only 3 years away from the first decade now.
I agree and find that Tesla, just like Apple, have (unvoluntarily) moved from leading edge to trailing edge, and try to (ever lesser results) cover this fact up with outlandish claims and hollywood-shows … (in the case of apple its AI-“ready” (which is NOT ready), in the case of tesla its the robo-taxi) …
just give today’s news outlets some juicy photos and visuals and it will show up in the news-cycle.
but yeah … lots of vapor, very little substance.
That presentation was so ridiculous that they are getting heavily roasted even on the fan boy fora. Stock is down 7.5%. No one thinks that those autonomous taxis being deployed any time soon are even close to a reality.