4 ms·

Vehicles per-operator is the important metric, imho. If one operator can supervise a dozen vehicles at a time with only occasional manual intervention then it's

by Pxtl 2y ago

Vehicles per-operator is the important metric, imho. If one operator can supervise a dozen vehicles at a time with only occasional manual intervention then it's still a massive labor-saving tool.

doctorpangloss 2y ago
All the evidence suggests that the number is >1 operators per vehicle.
- bryanlarsen 2y ago
  >150,000 trips / week. 3000 employees. 1:1 supervision is only viable if every single employee was a supervisor, and that's obviously not true.
  - smokel 2y ago
    If those 3000 employees work full-time (40 hours per week), they could theoretically serve 150,000 trips of up to 48 minutes in length, which seems reasonable, no? Another interesting statistic here is that Waymo apparently has only 700 vehicles in its fleet [1]. That would require only 700 supervisors. [1] https://www.reuters.com/business/autos-transportation/alphabets-waymo-robotaxi-unit-doubles-its-paid-rides-three-months-2024-08-20/ https://www.reuters.com/business/autos-transportation/alphab...
    - bryanlarsen 2y ago
      Supervisors don't work 24/7 like the cars do. For 1:1 coverage you need ~4X as many people as cars. As I said, theoretically viable if all employees were supervisors. But they're not.
      - smokel 2y ago
        I tend to agree with you that it's very likely that the cars are operating autonomously most of the time. However, I don't think this kind of math helps much in actually proving (or disproving) that. We don't know how many of the employees are supervisors. For all we know, Waymo hires dozens of external mechanical turks per vehicle and uses a simple majority voting algorithm to ensure safety. Unlikely, or even ridiculous, but apart from Waymo's press releases, and the accidental legislation report [1], there is little hard evidence. [1] https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-vehicles/disengagement-reports/ https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...
- jsnell 2y ago
  Ok, I'll play. Would you like to present said evidence? Remember that your claim was very explicitly about Waymo, so I'd hope the evidence is about them rather than some other company.

by Pxtl 2y ago

Vehicles per-operator is the important metric, imho. If one operator can supervise a dozen vehicles at a time with only occasional manual intervention then it's still a massive labor-saving tool.

doctorpangloss 2y ago

All the evidence suggests that the number is >1 operators per vehicle.

bryanlarsen 2y ago
>150,000 trips / week. 3000 employees. 1:1 supervision is only viable if every single employee was a supervisor, and that's obviously not true.
- smokel 2y ago
  If those 3000 employees work full-time (40 hours per week), they could theoretically serve 150,000 trips of up to 48 minutes in length, which seems reasonable, no? Another interesting statistic here is that Waymo apparently has only 700 vehicles in its fleet [1]. That would require only 700 supervisors. [1] https://www.reuters.com/business/autos-transportation/alphabets-waymo-robotaxi-unit-doubles-its-paid-rides-three-months-2024-08-20/ https://www.reuters.com/business/autos-transportation/alphab...
  - bryanlarsen 2y ago
    Supervisors don't work 24/7 like the cars do. For 1:1 coverage you need ~4X as many people as cars. As I said, theoretically viable if all employees were supervisors. But they're not.
    - smokel 2y ago
      I tend to agree with you that it's very likely that the cars are operating autonomously most of the time. However, I don't think this kind of math helps much in actually proving (or disproving) that. We don't know how many of the employees are supervisors. For all we know, Waymo hires dozens of external mechanical turks per vehicle and uses a simple majority voting algorithm to ensure safety. Unlikely, or even ridiculous, but apart from Waymo's press releases, and the accidental legislation report [1], there is little hard evidence. [1] https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-vehicles/disengagement-reports/ https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...
jsnell 2y ago
Ok, I'll play. Would you like to present said evidence? Remember that your claim was very explicitly about Waymo, so I'd hope the evidence is about them rather than some other company.