The 1m rides mark was announced May 10th and 50k trips per week was announced May 16th. So they must be getting close to 100k rides per week if my math is right.
Are we seeing the hockey stick?
They've been consistently candid, and indicated the next vehicle is the one they'll scale on, custom taxi by Chinese automaker Zeekr. I figured it'd be another couple years, maybe build out LA, scale SF to highways, and then maybe Zeekr. But they recently started testing it, way ahead of what I thought they would...[2]
EDIT: -3?! Anyone have feedback? Genuinely surprised. :)
[1] May 10th to July 26th = 77 days 1,000,000 rides / 77 days = 12987 rides/day 12987 rides / days * 7 days/week = 90909 rides/week
[2] https://techcrunch.com/2024/07/23/the-waymo-zeekr-robotaxi-h... Steps: May 10th to June 1st = 21 days June 1 to June 30th = 30 days July 1st to July 26th = 26 days
1m rides in 67 days = 130k
Waymo's scale is constrained by many factors: (1) physical cars (requires capital), (2) hi-def mapping, (3) operations (depots to charge, clean and maintain cars, "fleet response" - https://waymo.com/blog/2024/05/fleet-response/), and (4) regulatory environment (permits on a city/county/state basis).
I suppose, if only for the sake of stimulating conversation, it is worth ignoring the substantive stuff to transmute "hockey stick" into "it is software, 0 and constant marginal cost"
I just don't really know what to say back, because I obviously don't think the self-driving car is software. You know that. I know that.