-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Poll: Will we see the FSD button in June?
- Thread starter DanCar
- Start date
Or to be even cheekier, he could initiate the FSD subscription and wait on the FSD beta product for another year. Nothing like attracting more suckers for vaporware. As a four and a half year EAP and FSD owner, I have no idea why I ever purchased FSD when EAP has 99% of the combined functionality, but fortunately I didn’t lease my car and intend to keep my car for much longer than four years.
diplomat33
Average guy who loves autonomous vehicles
Terminator857
Active Member
To be fair, the poll didn't actually say June of 2021. So there's still a 1-in-12 chance of the answer being yes, as was always the case.
Silicon Desert
Well-Known Member
Yes, but I wonder if Elon will later come along and say something like... "People, I was talking about days on ~ 10 days on Mercury, not Earth !" and if that is the case, he could still deliver on time because if I remember correctly from science class, a day on Mercury is equivalent to something like 2 months on earth. Let's see, 10 Mercury days.... hmmmm, sounds like he still has about 16 more earth months to deliver
There were too many June Bugs
Cheburashka
Active Member
Wow, 8.2 was released that long ago and no signs of progress since.
diplomat33
Average guy who loves autonomous vehicles
Well, I think going from V8.2 to V9 is the switch to pure vision. That would take time.
But I think the main reason for the seemingly lack of progress is probably just the difficulty in doing "L2 City". Karpathy mentioned at CVPR that FSD beta struggles with more complex driving like what you might see in busy SF. There are lots of situations in city driving that require split second intervention. We've seen videos where FSD Beta seemingly starts to do an unprotected left turn and doesn't see cross traffic. The driver has to be very attentive. Tesla is probably trying to make FSD beta more reliable before they can safely release it wide.
Maybe geofence it? I know that’s a dirty word at Tesla. But large parts of the country look nothing like downtown San Fran.
They already geofence NoA and Smart Summon.
diplomat33
Average guy who loves autonomous vehicles
I agree. There is a reason why AV companies geofence. L5 is a huge problem. It just makes sense to break it down into smaller chunks. If Tesla geofenced FSD Beta, they could release it sooner in areas where it does work. Without geofencing, everybody has to wait until FSD works everywhere. It will take a long time before FSD actually works well enough everywhere, and in all conditions, that you can remove driver supervision everywhere. Maybe they could get FSD Beta to work well enough in a specific area to remove driver supervision and do geofenced robotaxis in certain areas sooner?
diplomat33
Average guy who loves autonomous vehicles
"Didn’t expect it to be so hard, but the difficulty is obvious in retrospect."
Seriously?! He did not realize how difficult FSD is? Literally, everybody working on FSD has been saying it is a very difficult problem. IMO. this makes Elon look really clueless.
FOOLSLFDRVNG
Member
Excluding the area with possibly the highest Tesla-per-person ratio around would be some pretty bad optics, though.
The Bay Area, yes, but nobody drives in downtown SF if they can help it.
Elon clearly thought FSD would be much, much easier than it is, based on his perpetual optimism regarding timelines. From the start he thought it was a year or two away.
Elon gets the job done. Just never on time. And in the case of FSD, maybe not while the early buyers of FSD still own their cars. Maybe not while people buying cars today still own their cars.
No kidding that they underestimated city complexity. Given freeway FSD issues that still persist, should we be even slightly surprised? In addition to the bizarre sudden braking on freeways it amazes me that they still haven’t solved freeway lane merges such that our cars don’t cut off other cars by immediately centering in the merge lane. Freeways are surely many times easier than streets so I expect to be taking over many times on every city FSD drive for at least the next three or four years — and probably longer. Tesla surely isn’t the only company that exaggerates but it’s ridiculous how much FSD was oversold starting in Fall 2016. Four and a half years later and we aren’t close to what they sold us would be completed in a year or so.
diplomat33
Average guy who loves autonomous vehicles
Other companies were off on their predictions of when FSD would be solved, for sure. But they did not sell "FSD" to customers. The big difference is that Tesla is the only company that actually sold "FSD" to customers, claiming that the AP2 hardware was "FSD capable" and it was just a matter of finishing and validating the software, when they did not have FSD.
Similar threads
- Poll
