@intellectronica HK looks like a very restricted experiment to me?
"Baidu Apollo International Ltd was approved to conduct trials for 10 autonomous vehicle in North Lantau, according to a statement released by Hong Kong's Transport Department on Friday.
The license will be valid from Dec. 9, 2024 to Dec. 8, 2029, with only one self-driving car operating on specified road sections at a time in its first phase, according to the statement"
@OllieBase with a backup driver too:
"During the trial, a backup operator stationed on board will take over control of the vehicle if necessary, it added."
@benshindel It might be kinda cheap to count Mesa as distinct from Phoenix but there's also Austin, Texas.
Clarifying questions before I bet:
The John Carmack & Jeff Atwood wager is about level 5 autonomy and this is about level 4, right? Waymo is level 4, meaning no human in the driver's seat nor remotely in control when the car is driving but the car may have restrictions on where it can go and in what conditions and may require human input to make confusing decisions. (I think the human only gives that input when the car is stopped. So the human is not in the loop in real time when the car is in motion.) In any case, I predict that level 4 will be kind of ubiquitous before any city has level 5. As of late 2024, level 5 doesn't exist at all.
Does it count as "generally available" if there's a wait list?
What Chinese cities currently count?
Same story not paywalled, on Waymo's blog: https://waymo.com/blog/2024/09/waymo-and-uber-expand-partnership/
So just in the US we're about to have
Phoenix
San Francisco
Los Angeles
Austin
Atlanta
But Austin and Atlanta won't be available to the public until "early 2025" and this market is about 2025 January 1.
@ScottBlanchard if you read this piece, you'll notice it links to another Waymo release. While Zoox and others may be extensively using remote control, Waymo (like Apollo Go) is not:
https://waymo.com/blog/2024/05/fleet-response/
Having a remote helpline that inquires to the city which lanes are meant to be blocked off by cones is not a "remote human driver". If human Uber drivers were using a service like this, it would be hilarious to describe it as such.
@benshindel Mate - As I said, I no longer have a dog in this fight. I did read the blog post and I don't think you are grasping why they are contorting the language the way they are. Again, I exited the market already. I absolutely respect your right to your deluded understanding of what these companies are passing off as "fully autonomous". I see a ton of motivated reasoning in my line of work and have learned to ignore it. Please consider me as having no interest in this market and how it resolves. Thank you and vaya con dios!
@ScottBlanchard I also currently have no shares, so I’m not sure what my motivation would be. Your motivation to hold onto the last vestiges of human autonomy seems pretty understandable though.
@ScottBlanchard more so than Waymo? They also use remote drivers to get out of tricky situations, but I think they clearly count for the resolution of this question.
@HenriThunberg Well - your resolution says “no test, backup or remote human driver”, so if services using remote drivers count, perhaps you shouldn’t have written that they don’t when writing the resolution rules.
@HenriThunberg In any event, I exited the question so I no longer have a dog in the fight.
@ScottBlanchard not my question, but I'm exiting my YES based on agreeing with you that resolution criteria are tricky and ambiguous.
@HenriThunberg My apologies - I assumed that as you replied to me it was your question. My mistake entirely and once again I am relearning the lesson of the dangers of making assumptions.
@benshindel Your understanding is wrong. As with Waymo and I suspect every ‘autonomous’ car out there, there is extensive human intervention. The companies don’t highlight (or even acknowledge) this normally, but Waymo had to release a ton of data in response to the CA investigation and the amount of human intervention is shocking. The Chinese companies have also admitted that there is some remote intervention, but I see no hard data on how much. Given Waymo’s data, though, and the relatively lower cost of human monitors in China along with the higher stakes of failure, my baseline assumption is that the human component is even larger in China.
@intellectronica you probably want to weigh in here, for the sake of the market?
I think your spirit is that adding new Waymo cities count, but if Scott's info above stands uncontested then it's bery hard to know what to bet on for people:)
Perhaps closing market and relaxing the criteria for autonomous a bit?
@ScottBlanchard Your assumption is mistaken. Apollo Go is fully autonomous. The presence of human monitors (as with Waymo) is not relevant to this question. Monitors are not “remote drivers”!