It’s not a delayed walk signal. Watch the video very carefully. The light turns red before the crosswalk turns green. You can see the light turn yellow in my screen capture. The pedestrians start walking because the light is yellow and the Tesla doesn’t have enough room to go anyway so they assume it will stop.
The point is that you don't know if the training is successful unless you test every possible scenario. There's no way to do that rigorously, so all you can do is go after a large set of scenarios, declare it a success, and hope that it's good enough. That's fine for an Office product, but not so great for a safety critical system.
I have lots of people on ignore, might be one of those folks who posted.WOW, talk about NOT looking at the thread.
There's an odd intersection at 16:00 where the painted arrows show both lanes going straight except there's only 1 destination lane across the intersection:
That looks like a right-turn-only lane marker to me. The straight part has faded, and they didn't repaint it. That's certainly consistent with the physical layout of the roads.
but you're more likely correct. It would be confusing for many drivers new to that intersection though.Looks like https://twitter.com/texasjkim911/ is gone too now. Both were located close to Fremont/Lathrop (Stockton) or Austin, so if they weren't employees, maybe the vehicles used to be Tesla employee vehicles?
This is really funny if it's the case. Elon Musk gave thumbs up to the posts then did this happen?
Except that she goes all the back to the Roadster, has owned other Tesla's since and been testing FSDBeta for awhile so doubt she wasn't aware of the potential consequences.
