Both comparisons are relevant. The lack of a competitor comparison means we don’t know if anyone else is doing better.
On the quotes, I suspect that you accidentally damaged the markup in some way. I periodically see quotes formatted like plain text with some of the markup.
If you quote the entire post, the board software will remove the quote. Has done that for quite a while. I’ve learned to make some small edit in the quote — drop a final period is common for me.
Their general instructions tell you not to sign posts, too. Phooey on that!
—Peter
Doesn’t matter. What others are doing doesn’t get Tesla to lvl 4 or 5 any faster or slower.
There could be ZERO other competitors and it would still take Tesla just as long to get to lvl 4.
How many of those are safety critical interventions?
I also drive FSD daily, not nearly as long of a commute (10 miles).
I haven’t had a real safety intervention in many months, but I intervene almost daily to be more polite, to take a better route or to provide feedback to Tesla on things that should be improved.
Mike
Which is only one of the questions people would like answers to. In particular, if someone else is getting there twice as fast, that has investment implications.
That is a different sort of problem your style of intervention. Musk wants you off the steering wheel and surfing the net. His real profit motive is gathering up the internet time while traveling.
I get it will take beyond lvl 4.
Back to who is getting to lvl 4 soonest?
Is Waymo not there already with their self driving taxi?
Investment implications was not the point of the study.
Relevant for this board - not for the study.
Which was my point…
I doubt it. I think we are at 15 mph
I don’t believe that there is a remote driver in the L4 specs. It also reminds me of a while back when some automaker, German?, got approval for L3, but with so many qualifications on when it could be used that it seemed meaningless.
There can be. The definition of Level 4 systems allows for it. Level 3 systems allow the car to drive itself - the user is not driving - but require the user to be ready to resume driving functions in the event the system demands it. A Level 3 requires a user to be ready to intervene in order to be safe. Level 4 does not require the user to ever drive - but it also does not mean that the vehicle will be able to complete the trip without user intervention. Just that the car will never need the user to intervene in order to be safely operated.
NOTE 2: Level 4 ADS features may be designed to operate the vehicle throughout complete trips (see 3.7.3), or they may be designed to operate the vehicle during only part of a given trip (see 3.7.2), For example, in order to complete a given trip, a user of a vehicle equipped with a Level 4 ADS feature designed to operate the vehicle during highspeed freeway conditions will need to perform the DDT when the freeway ends in order to complete his or her intended trip; the ADS, however, will automatically perform the DDT fallback and achieve a minimal risk condition if the user fails to take over when the freeway ends (e.g., because s/he is sleeping). Unlike at Level 3, the Level 4 feature user is not a DDT fallback-ready user while the ADS is engaged (see Example 2 below), and thus is not expected to respond to a request to intervene in order to perform the fallback. Nevertheless, in the case that a Level 4 sub-trip feature reaches its ODD limit, the ADS may issue an alert to the passenger that s/he should resume driving in order to complete their trip. (Note that in this latter case, the alert in question is not a request to intervene, because it does not signal the need for fallback performance.)
But, that is in reference to a passenger/driver in the vehicle and for switching from doing L4 for part of the trip to doing conventional driving for the other part. What Waymo is doing is using a remote human to sort of patch over a weak spot.
Sure, but it’s the same thing. The defining difference between L4 and lower levels is that you don’t need to have a human in the driver’s seat (either literally or remotely) to serve as a fallback to the AV system. It doesn’t mean that you can’t have humans stepping in to give the car direction if the car has encountered a situation that requires it to pause the trip.
I’m not out here claiming Waymo is all that and a bag of chips. They may also be quite far from having a system that can provide meaningful transportation services without much human involvement at all. But to be Level 4 (or 5), you have to have a car that can safely operate without a human driver being available to intervene upon demand. Waymo has cars that do function as Level 4 vehicles. They require some degree of remote operate infrastructure to complete some of their trips. Not to operate safely, but to navigate past situations that have caused the car to pause the trip. There’s no indication that Tesla’s anywhere near that.
Same thing conceptually … or same thing according to the spec?
I have used Waymo. It is very good. Quite impressive.There are a few show stopper problems.
Waymo was the leader about 15 years ago with Google big daddy pumping in Billions but have not been able to execute and monetize this for a long time. It was a lab project under Google’s umbrella. It has just recently began expanding.
The spec doesn’t address remote drivers, but it does delineate the difference between Level 4 and lesser systems as being the requirement that a driver must be in a position to take over driving the car upon demand by the AV system when it disengages. Level 4 systems don’t have a driver in that position - Level 3 and under systems do.
Waymo doesn’t have that, so it’s a Level 4 system. Whether the human driver is waking up from a nap in the back seat to get behind the wheel, or leisurely logging in remotely for guidance after the car comes to a minimal-risk stop - either way, there is no driver operating the vehicle or prepared to immediately intervene while the AV system is running. That’s the mark of a Level 4 system - the absence of an intervening driver, not the location of a human that might step in if the Level 4 system has reached a point where it can’t continue the trip.
@WicksAndSticks posted this link on the paid side:
https://x.com/raines1220/status/1839686199847313452
Worth reading down the thread a bit too. Puts a bit of a different view on things.
Not that different a view, though.
Again, the key difference between FSD and Waymo is that Waymo doesn’t need to have the human take over during disengagements to drive safely. FSD can’t emulate that trick by calling for help if, for example, it can’t figure out what speed to drive. To be a Level 4 system, and not have someone in the driver’s seat, the car needs to be able to figure out those types of driving functions.
The author is correct that non-critical intervention rates wouldn’t be apples to apples comparisons between Tesla and Waymo, because Waymo can (and by all accounts does) have situations where the car has to drive itself into a minimal-risk situation and call for help. But those types of interventions - the Level 4 type of interventions - cannot require the immediate intervention of a driver to avoid a crash.
Tesla doesn’t release intervention data. The only data we have on FSD safety-critical interventions comes from crowd-sourced efforts. None of that data suggests that Tesla is anywhere close to being able to operate a system without a driver in the seat. At best, crowd-source data puts FSD 12.5 at a few hundred miles between critical interventions, which is about four orders of magnitude below what’s necessary to run a driverless system.
