For example, Waymo claims 700k fully autonomous rides in 2023, which was more than ~0 Tesla rides that were fully autonomous in 2023.
To define fully autonomous, I'll use the simplest definition which is that no human operated any direct controls inside the vehicle like the steering wheel or pedals. (UPDATE: And no human in the car is actively monitoring the driving being ready to intervene.) A ride must be of a non-trivial distance (e.g. crossing the parking lot does not count).
I will delay resolution until both companies report results or until the result is extremely clear, or until it's January 2027, where I'll make an informed guess.
Update 2024-13-12 (PST): Rides in the Las Vegas Loop will not be counted towards Tesla's total, as they do not represent general self-driving capability on public roads. (AI summary of creator comment)
1,000
3.00@JamesGrugett that’s lame. It seems like such a low bar to say that Tesla doing 1 million rides in a curved tub for 1 mile with no traffic or lights beats waymo doing 999k rides on city streets.
@JDTurk Is it "lame," or is it a really great way of raising incremental difficulty and finding synergies between companies? If you're smart, then you will realize that there's a lot of machine learning that can be applied to predicting what passengers are going to do just before and just after they board or disembark a robo-taxi. The Las Vegas tunnels would be a great place to gather that data, without the complication of traffic!
@StCredZero Yeah, it's not lame for Tesla to do this, it would just be lame for us NO holders to lose on a technicality, which this could potentially be.
To be clear, I'm betting NO because Tesla has not actually figured out level 4 autonomy yet and I think it will take well into 2025 if not longer for them to serve even their first fully autonomous ride. Whereas Waymo is serving >150k per week and growing. I think it seems like Tesla is close based on how impressive FSD is, but it's all about the number of 9's. If there's even one time in a month or a year where you personally had to intervene to prevent FSD from crashing, that's a far cry from the reliability level needed to have no one in the driver's seat.
Unless maybe it were limited to something very controlled and predictable like the Las Vegas loop? Idk, people complain that Waymo is limited in that they only serve rides where they've thoroughly mapped the routes. But as we know from Google Street View, that doesn't make Waymo's approach unscalable.
I see potential ambiguity in that definition of fully autonomous. Suppose that Tesla self-drives totally fine, except that it's oblivious to pedestrians. As long as a human is in the driver's seat ready to hit the brakes, you could technically have thousands of rides in which no human intervened. The human was only there "just in case" but it's a pretty huge "just in case" and doesn't count as autonomy.
So I propose that "rides with no human in the driver's seat" is the right proxy for this question. Waymo's 700k number sounds like it refers to rides by public users and with no human backup driver. (I'm several of those 700k -- it's amazing!)
Alternatively we could say level 4 autonomy as certified by someone other than Tesla or Waymo? I don't know what "certified" would mean but maybe consensus of some form suffices. I believe current consensus is Waymo is level 4 and Tesla is level 2.
In conclusion, I think by any reasonable definition of "fully autonomous" Tesla is not on track to catch up to Waymo, and am betting accordingly. But I'll feel more confident if we clarify that a ride with a human in the driver's seat whose attention is needed to make it safe doesn't count, even for rides where the human didn't happen to need to intervene.
@dreev Cool. I think it might be a bit too restrictive to say no one can be in the driver's seat.
I think we could add that the person in the car is not actively monitoring the situation -- seems reasonable. But I also think this is mostly unnecessary, since if active monitoring is necessary, the car would most likely require you to put your hands on the wheel every so often, like it does now.
There's a legal classification of self driving technology. Waymo is operating level 4 self-driving technology, Tesla is currently operating level 2; Mercedes-Benz has a level 3 car operating. Am I correct that the resolution criteria for this is that Tesla needs operate more level 3 rides than Waymo operates level 4 rides (since Waymo does not operate at level 3)? Level 3 is the lowest level that fits the criteria "no human in the car is actively monitoring the driving being ready to intervene"