Welcoming our first riders in San Francisco • Hacker News Board

Re: Welcoming our first riders in San Francisco

#291

Post by daveslash » Tue, Aug 24, 2021, 5:58 PM UTC

noneeeed wrote: post #79

As someone who lives in Bath, an old european city with roads (and other drivers) that can give experienced human drivers a nervous breakdown I'm rather more interested in how they do in SF than their current deployment in Phoenix. From what I've heard repeatedly, SF sounds much more irregular and messy than Phoenix, so it should be something of a stepping stone to making them usable more widely if they can crack it.…

Re>> "that can give experienced human drivers a nervous breakdown"

A couple of years ago I took a wrong turn and got lost in the Long Beach harbor - and had neither GPS nor phone. Holy Cow man.... holy cow... (admittedly, low quality comment)

Re: Welcoming our first riders in San Francisco

#292

Post by Bukhmanizer » Tue, Aug 24, 2021, 5:58 PM UTC

ra7 wrote: post #144

shadowgovt wrote: post #121

Earlier quoted context omitted.

Level 5 isn't feasible as much for legal reasons as technical ones. I don't think any company wants to sign off on the notion that their software will handle all classes of problem, even ones they have no data for at all.

Except Tesla, who are going to be "L5 by the end of the year" every year!

I’m pretty sure Tesla’s whole strategy is to overpromise so much it’s as much a legal liability to not have L5 than to have L5.

Re: Welcoming our first riders in San Francisco

#293

Post by _ea1k » Tue, Aug 24, 2021, 5:58 PM UTC

nradov wrote: post #92

ra7 wrote: post #19

It's fantastic to see Waymo's progress. SF is a real nightmare to drive. If they can nail it there, that's 2/2 for busy, urban street driving (SF) and "boring" suburban driving (Chandler, AZ). They've been quietly very confident of their tech, but this is a real test. Tangentially, I've also noticed that Waymo has picked up pace ever since the recent leadership changes. They are publishing more blog posts, offering m…

SF is hardly a nightmare to drive. Try Boston during a winter storm.

Honestly, I think Boston is harder even without a winter storm.

Re: Welcoming our first riders in San Francisco

#294

Post by Johnny555 » Tue, Aug 24, 2021, 5:59 PM UTC

MontyCarloHall wrote: post #59

“All rides in the program will have an autonomous specialist on board for now” This tells me that we’re still a long way from full level 4 (and certainly level 5) autonomy in a busy city like San Francisco. The edge cases requiring immediate human attention are still too frequent for the human safety driver to be remote, as is the case in Phoenix. Also, just a reminder that Waymo in Phoenix is nowhere close to being…

The edge cases requiring immediate human attention are still too frequent for the human safety driver to be remote, as is the case in Phoenix.

I think you can only determine that if you know how many times the human attendant takes over.

Just having a human behind the wheel doesn't tell you much, I don't see how to get full self-driving without an intermediate step of human supervision.

Re: Welcoming our first riders in San Francisco

#295

Post by Lammy » Tue, Aug 24, 2021, 5:59 PM UTC

cascom wrote: post #278

Lammy wrote: post #67

Earlier quoted context omitted.

> SF is a real nightmare to drive. 'cause they decided not to build most of the roads: https://www.cahighways.org/maps/1955trafficways.jpg

Supposedly more road construction doesn’t alleviate traffic, in only induces more demand (which is moderated by high traffic levels) Source (great read if your interested in the subject): https://www.amazon.com/Traffic-Drive-What-Says-About/dp/0307...

Yes, inducing some demand is the point. People have to live somewhere, work somewhere, and until recently generally had to commute between the two. When this happens in your circulatory system it's called a stroke :p

Re: Welcoming our first riders in San Francisco

#296

Post by _ea1k » Tue, Aug 24, 2021, 6:00 PM UTC

lifekaizen wrote: post #156

rpmisms wrote: post #39

Earlier quoted context omitted.

Waymo and Tesla have very different approaches to autonomy. I think both are valid avenues, but it's fascinating to watch them develop in parallel. You don't have to make it into a spiteful contest.

Definitely both valid business approaches, Tesla more interesting from a startup / leverage approach. As a consumer, I'm less excited about Tesla putting out beta software in situations where mistakes can be deadly https://www.nytimes.com/2021/07/05/business/tesla-autopilot-...

As someone who has seen and driven their "pro" pilot competitor, I am not bothered by that. The non-beta competitors are often worse and sometimes more dangerous (sigh).

Re: Welcoming our first riders in San Francisco

#297

Post by chrisseaton » Tue, Aug 24, 2021, 6:00 PM UTC

keewee7 wrote: post #289

chrisseaton wrote: post #54

Why does Waymo use a niche British luxury car for their platform? Must be very expensive and what are the benefits?

Is Jaguar still a niche brand? Most taxis in Denmark and Germany are Mercedes-Benz so it's not uncommon for the taxi industry to use luxury brands.

Niche in that they only sell a hundred thousand or so a year, yes. If Waymo are planning to seriously scale up they'll be using every Jaguar they're making!

Re: Welcoming our first riders in San Francisco

#298

Post by dragonwriter » Tue, Aug 24, 2021, 6:01 PM UTC

hammock wrote: post #148

MontyCarloHall wrote: post #59

“All rides in the program will have an autonomous specialist on board for now” This tells me that we’re still a long way from full level 4 (and certainly level 5) autonomy in a busy city like San Francisco. The edge cases requiring immediate human attention are still too frequent for the human safety driver to be remote, as is the case in Phoenix. Also, just a reminder that Waymo in Phoenix is nowhere close to being…

>“All rides in the program will have an autonomous specialist on board for now” This tells me that we’re still a long way Did you expect something different? I can't really see a boardroom writing a roadmap that goes straight from test rides (no passengers) with a backup driver onboard to actual rides with passengers - no backup driver onboard with no in-between steps.

Waymo is already doing test rides with neither passengers no backup drivers in CA, so they wouldn't be jumping from no passengers plus safety drivers to paid passengers without safety drivers if they did offered paid, full-driverless rides.

Re: Welcoming our first riders in San Francisco

#299

Post by nicoffeine » Tue, Aug 24, 2021, 6:01 PM UTC

actusual wrote: post #190

kspacewalk2 wrote: post #175

Earlier quoted context omitted.

Stacking shelves in a warehouse is one task. Driving is not one task. There are too many corner cases for a modern-day AI system to perform as well as a median driver in, say, 95% of environments and settings in North America and Europe. I think the argument is that such a system might as well be AGI.

The idea that an AI must have the ability to learn how to do anything in order to learn how to drive seems like an extremely pessimistic and misguided goalpost. That is also not how iterative development works.

I think ML is fantastic, and combined with LiDAR, inter-vehicle mesh networking, and geofenced areas where humans take over, we could quickly arrive at mostly automated driving without trying to reinvent the human brain. We should also be more focused on enforcing established legal limits to newly manufactured cars. Just preventing someone from exceeding the speed limit or driving the wrong way would start saving lives immediately. It would also allow traffic flow to be optimized, and eventually prioritize emergency traffic or allow metro areas to be evacuated efficiently for things like natural disasters.

It would be great to see the dawn of AGI, but I don't think it will ever happen with classical computation. GPT-3 spits out nonsense with the input of the largest and easiest to parse portion of reality, and I have not seen any ML approach replicate the abilities of something as simple as bacteria. ML requires constant validation from human operators, so the same is going to hold true for ML powered vehicle navigation.

Re: Welcoming our first riders in San Francisco

#300

Post by dzdt » Tue, Aug 24, 2021, 6:02 PM UTC

MattGaiser wrote: post #49

gfodor wrote: post #27

Kind of interesting to consider how this adjusts priors on these outcomes: - Self driving is not possible - Self driving is not possible anytime soon - Self driving is possible, but requires LIDAR - Self driving is possible, and can be done with normal cameras I've engaged with a lot of people who presume self-driving reduces onto AGI, therefore it will not be achieved anytime soon if ever. I wonder which of their as…

I assume that they just don't solve the really wacky edge cases which might require AGI. Many people assume that a self driving car must be 100% successful. To me it is sufficient to beat the top 20%/bottom 80% of drivers. So the car will crash if a seagull flies in front of the camera and there is a deer just beyond it. The human would probably crash too.

I think self driving can work with AGI to match a horse. It doesn't need to understand with a depth and clarity of a human.

That is still a pretty high ask for current software.