EVNow
Well-Known Member
Licensing FSD won't be easy. OEMs need to redesign their cars to include cameras the way they are in Teslas.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Elon: FSD Beta tweets
- Thread starter Terminator857
- Start date
Yea this is more of a PR move than a real chance of it happening. Also at this point (L2) there is little incentive. If Tesla could offer a solid L3 system then there may be some takers. If it makes L4 ("end of A year"
) and is the only consumer system available then Tesla would really have something to offer.
willow_hiller
Well-Known Member
To be fair, we were all saying the same thing about Tesla releasing their charging standard as NACS. Now look at what Ford's doing.
EVNow
Well-Known Member
I think there is a big difference. Super charging network is a unique infrastructure that is simply not possible to recreate in a hurry. And it works great. Infact I don't know whether it is really all that helpful to Tesla unless Ford is paying a lot (which would help expand the network).
FSD is a totally different thing. Currently it doesn't even work well as ADAS. OEMs don't even want to take up MobilEye which is designed keeping in mind OEMs.
I would guess licensee. But in event of lawsuit both could be liable I expect. In reality it's all likely something that would be discussed as part of a license contract discussion.
EVNow
Well-Known Member
Yes - OEMs are very used to this. Think of hundreds of sub-assemblies they use and what happens when something goes wrong (like the air bags, LG batteries or Firestone tires).
willow_hiller
Well-Known Member
The rate of FSD Beta tweets has decreased, but the rate of firmware releases are increasing:
Except the rate of improvement is literally NOT increasing. Lane selection is the worst it’s ever been, creeping is scary as heck because FSD simply can’t see good enough with the stupid camera placement.
EVNow
Well-Known Member
One regression doesn't mean there are no other improvements.
Indeed. I am surprised that in this day and age the general public is not more familiar with the sawtooth nature of improvements in early life software.
When software is stable and nothing major is changing then you should expect each release to be 99% improvement over the previous, but when it undergoes a significant or fundamental change you nearly always get some unexpected regression and it takes some time to achieve parity with the previous version.
The reason for voluntarily incurring this regression is when you think the underlying change opens up greater potential for the future than the code it replaced.
In the case of FSD they’ve just changed the entire lane planning module which has improved performance in areas it used to struggle with. The areas where it’s taken a step backwards will catch back up in a release or two.
I never looked up further on these, but does it also do any kind of auto-lane changing when needed?
Mapped highways but under 40mph means only under moderate/heavy traffic?
EVNow
Well-Known Member
It's basically just traffic jam assist. A PR stunt.
My take is a little more nuanced. It sounds like a real feature that will be of genuine use to some drivers, so I don’t think the technology itself is a stunt, but writing it up as if Mercedes have scored some massive victory over Tesla at a game that Tesla isn’t even trying to play is somewhere between lazy journalism and a desperate stunt (depending on how much of it is coming from Mercedes and how much is from the second hand writeups).
The interesting question is how much headroom Mercedes’ platform has to develop further. Tesla’s system (by which I mean all the back-office data collection, testing, and training systems, as well as the in-car platform) has the capability it needs to drive on all roads. How far can Mercedes take theirs?
EVNow
Well-Known Member
There is a 26 page long thread on all this.
Will Mercedes jump to level 3 before Tesla? Looks like it.
I'm not a software engineer or AI expert by any stretch, but I have to wonder if Tesla Vision will ever deliver what we've been promised. Current iteration of autonomous driving beta is still quite wonky. Phantom braking persists, and for me that means leaving autopilot off most of the time. My...
teslamotorsclub.com
Totally agree. But it is hands off and eyes off. There is no Tesla nag to touch and no Tesla keep looking forward warnings. This alone is quit impressive; even with the MAJOR limitation. Now if it could only be scaled up to double that speed and then unmapped streets.
EVNow
Well-Known Member
If only the OD was very different...
Wouldn't that mean they would have solved FSD .... that nobody has done till now.
Of course. That makes sense. The problem is that 11.4 specifically mentions improvements to lane selection. When something is said to be better, but yet it gets worse, that’s a bit disappointing. Really does not give much faith at all in FSD progress if stuff that is mentioned as getting better gets worse.
willow_hiller
Well-Known Member
Surprised this thread missed this. In 2016 Elon theorized the ability to summon your car from NY to LA (in 2 years, but who's keeping track). And now when someone asked Whole Mars about an LA to NY trip he replied "Soon."
Similar threads
