Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Tell HN: I took autonomous taxi rides to/from/around SF downtown yesterday
14 points by SeanAnderson on Sept 24, 2023 | hide | past | favorite | 20 comments
Hello! Forgive me if this isn't the type of content for HN, but I thought an experience report might prove interesting.

I registered for Waymo One in August '21 and was given access exactly two years later. Geofencing prevented downtown access until two weeks ago. My experience was superb and I will prefer autonomous driving going forward.

Notes:

- Photos: https://imgur.com/a/0PwBPVa

- It was $27 to go downtown. Uber was $19.50 + tip. I wished it had been equally priced. I waited 3-10 minutes for pickups.

- The ride felt posh and comfortable in an electric Jaguar.

- The vehicle tolerates on-street pick-up/drop-off if the street is quiet, prefers pulling over to a curb, and will go around the block searching for a curb if the street is busy.

- The doors unlock via app. The door handles slide out from within the car and "appear" when you unlock.

- There is an LED light on the car. By default, it displays your initials ("SA") and has a customizable background with several color choices. You may adjust these within the app. This wasn't useful to me, but I could see its benefit if there were multiple autonomous vehicles around.

- The vehicle greets you as you open the door and enter. "Hello, Sean. Happy first day of fall."

- The vehicle waits a maximum of five minutes to pick you up. Once inside, you tap a button on a screen to begin the drive.

- There are screens within the car which show some basic functionality: "Play Music", "Pull Over", "Call Specialist". The screens also show you how the car is "seeing" in real-time. The display shows detected stop signs, stop lights, pedestrians, other vehicles, traffic density, etc.

- The car chose reasonable routes. They routes did not match those I've had Uber/Lyft drivers take. It seemed to prefer slower routes that offered increased safety. This added ~1 minute onto ~22 minutes of travel.

- The driving was a good balance of cautious and aggressive. Turns were tighter than I was expecting. It drove through one yellow light which I felt it had time for and it paused for one yellow light which I also felt it had time for. A driver adjacent to me gunned it through the second yellow and made it. There were a couple of moments I felt it was overly defensive - slowing down more than I would expect when overtaking a vehicle parked in a bike lane and giving "too much" right-of-way to pedestrians who were considering jaywalking.

- It handled very tight spaces well. There were two moments where only a few inches of clearance existed between vehicles travelling opposite directions. It routed around drivers opening car doors repeatedly.

- There was one moment of driving where it faltered. It was driving from Mission to Dolores Park around 4PM. There were a lot of people walking on the street and jaywalking. We stopped at a red light intersection and observed a stopped vehicle with their hazard lights on the opposite side of the intersection. When the light turned green, there were people interested in jaywalking in front of the vehicle, but they were trying to figure out how the vehicle was going to respond to the blocked intersection. The vehicle started to cautiously roll forward, then the people got impatient and cut in front of it, and this caused the vehicle to think it was doing something wrong. It paused for about 10 seconds, when it clearly could've moved forward, before re-engaging and navigating through the intersection and around the stopped vehicle blocking the lane.

- There were palpable psychological effects on people. Drivers pulled up next to the vehicle to peer inside and freak out. Children gawked and pointed. People on the streets audibly exclaimed. An elderly person was visibly displeased and shook their head. The inability to make eye contact with a driver impacted how pedestrians interacted with the vehicle.

- The crowd response significantly improved my experience and, for now, justified the extra expense.



Every time I see one of these positive reviews for autonomous taxi rides, I think about how many people were having sex in it...

https://sfstandard.com/2023/08/11/san-francisco-robotaxi-cru...

It was even predicted...

https://www.fastcompany.com/90264809/self-driving-cars-will-...


To be fair, apparently Uber/Lyfts have their share of sex happening in the back seats, so it's not anything new. You'd think the human driver would disuade people, but there's no accounting for taste, apparently. Or that's the kink and the point.

Also, hotel rooms.


Is this a significant cause of concern for you?

Waymo has cameras running inside the vehicle. It would seem illogical to assume these people weren't flagged and that the car wasn't flagged for cleaning.

Most Uber's I've been in have felt more heavily trafficked than this experience and I'm no stranger to stories of people losing their lunch inside taxis.

I felt like this was a luxury experience in comparison to Uber/Lyft rides.


Why are you assuming I'm concerned?


What? I was asking if you were concerned because you brought up an article highlighting a concern and because your use of ellipses gave me the impression of additional sentiment not expressed in words.


Sentiment yes, concern no. Better to ask questions than to assume.


He did ask! He asked "Is this a significant cause of concern for you?"


No, he assumed I was concerned.


> The crowd response significantly improved my experience

lol, I was picked up by a Waymo outside the Filmore a week ago to a crowd of boo's.


     The vehicle started to cautiously roll forward, then
     the people got impatient and cut in front of it, and
     this caused the vehicle to think it was doing something wrong.
"Think"? There's no "think" involved here; rolling forward when there are pedestrians on the road isn't legal no matter how much they're "jaywalking".


They weren't on the street when the vehicle moved. They were standing on the corner. The vehicle had a green light, but wasn't moving into the intersection because the intersection was blocked. The vehicle started to roll forward to approach the intersection, the pedestrians decided to cut across without a right-of-way light, and the vehicle stopped rolling forward when they entered the crosswalk. It then took longer than I would've expected for it to decide the coast was clear and to approach the intersection again.

Yes, I understand it is not actually thinking. I'm a software engineer. I get that tech isn't sentient. I still find it enjoyable and natural to anthropomorphize technological advancements (https://en.wikipedia.org/wiki/Tamagotchi_effect). The pedantism isn't adding insight to the conversation.


Wow! This still feels like science fiction to me. Thank you for the report! Without question on-topic, IMO.


Personally, I wouldn't pay more money for an autonomous ride than a regular one. I will go with whatever's cheapest. So to hear that Waymo was more expensive is disappointing.


I checked for a 1.2 mile ride just now (23:52 on a Sunday night); it was $9.32 for a Waymo, $12.49, discounted to $9.66 for a Lyft after Lyft pink, and $11.32 for a Cruise. I don't have Uber access after using curl against one of their endpoints using my account, so I can't say what that would have cost.

So, pricing is dependent on various opaque factors, but it's not universally more expensive. Cruise used to be free, but they finally started charging for them.


When we hear from a man that a mechanized world would not be a pleasant one in which to live, quite the contrary, it should be true. And it can be true, if the fine minds which have accomplished so much in the realms of applied science would unite with the same enthusiasm to control our creation against social misuse.

Obviously, research regarding technological unemployment is as vital today as further refinement or production of labor-saving and comfort-giving devices.


The demise of the carbon based driver has been greatly exaggerated: https://www.reddit.com/r/BoltEV/comments/16pfi2z/self_drivin...


Those are Cruise vehicles, though, not Waymo. I have not ridden in a Cruise so I don't know if the experiences are comparable. I am aware of the concerns portrayed in the media which is why I felt offering a first-hand account to be valuable.


Here is your Waymo traffic jam: https://youtu.be/7J0DvWvbH2s

I would suggest, not to prefer automated driving, solely on the basis of a single drive, as you stated in your report.


Thanks for sharing! That's unfortunate to see.

Yeah, that's a fair call out. Originally I had written, "My experience was phenomenal. I do not plan on using Uber/Lyft anymore, when possible, until/unless I have a significant negative experience with autonomous driving." My original writing came out to ~6k words which was too long for HN. In trimming it down to 4k I reduced my statement and changed my stated intent a bit. Ah well.


Great report, thanks! Details were all the ones I hoped to learn.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: