I am not saying this will happen - in fact if I were a betting man I'd say we don't see wide release by end of year (defined as anyone who has paid for FSD can enable the beta).
It also probably goes without saying, but any kind of robust Level 3 that doesn't require constant driver supervision is NOT going to happen this year. Therefore the question becomes: what milestone/design changes would be required to make FSD Beta as safe and easy to use as possible for the masses?
Here are my thoughts:
That's my list so far. What else would you add? And do you think there's any way the above is possible by end of year? I think it is, because most the items on my list are about UI polish, limiting functionality (parking lots), and user awareness.
It also probably goes without saying, but any kind of robust Level 3 that doesn't require constant driver supervision is NOT going to happen this year. Therefore the question becomes: what milestone/design changes would be required to make FSD Beta as safe and easy to use as possible for the masses?
Here are my thoughts:
- Single stack. Leaving aside the improvements that this might bring to highway driving, I think it's necessary just for a consistent user experience. The folks on this forum understand 'two stacks', but the average driver would be confused and frustrated by the inconsistent UX and feature sets.
- More polished visualizations. FSD visualizations have come a long way, but they are still rough. Jiggling/morphing cars are still a problem. And the street markings (curb lines, painted street lines, etc.) are jumpy and garish. One particular glaring cosmetic issue is when you pull up in heavy traffic and the UI changes to the overhead view, with the curb markings jumping in and out of existence chaotically on the screen.
- Keep the driver attention monitoring and 5 strikes. However replace the permanent eviction from the beta with some kind of timeout period (say a month or two).
- Consider removing the steering wheel nags and rely entirely on driver attention monitoring. I think the nags actually distract the driver and result in unintended AP disengagements. I suspect the only reason they still have them is to keep the regulators off their back.
- Big improvements to the (presumably still) hard-coded driving and routing logic. They should focus a couple of releases on fixing obvious path planning issues, such as late (or early) lane changes, lane changes back and forth for no apparent reason, turn signals clicking on and off repeatedly while waiting at traffic lights, etc.
- Geofence out parking lots and limit FSD Beta to public roads. FSD Beta will 'try' to drive through parking lots, but it is often incredibly erratic. On more than one occasion I've had FSD Beta think the speed limit in a parking lot is 35mph. They should prevent FSD Beta in parking lots until they've put the time into training/coding for this specific problem space (and at the same time replace Smart Summon and add Parking Seek or whatever it's called).
- PUBLIC EDUCATION ABOUT THE LIMITS OF FSD BETA. Elon has done Tesla no favors with his hyperbole about both current and future state of FSD. The average person you meet on the street is often surprised that Tesla drivers can't sleep in their cars today ('I thought they were already self-driving'?). This is particularly dangerous when one of these people buys or rents a Tesla and turns on FSD Beta.
That's my list so far. What else would you add? And do you think there's any way the above is possible by end of year? I think it is, because most the items on my list are about UI polish, limiting functionality (parking lots), and user awareness.
