Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I remember one time my wife and I took the Super Shuttle home from SFO to Mountain View. It took about two hours since we were the last stop and in fact the shuttle went north a little first before turning south. That was the first and last time I used their service.


I love Lyft but I took a shared lyft for the first time and the driver did three circles around a busy block trying to find a GPS dot (that never moved) that was obviously on the other side of the street, and after an extra 10min jammed an extra person in the car when we hardly fit, despite our protests. That was the last time I took a shared ride.


> jammed an extra person in the car when we hardly fit, despite our protests

Are you talking about more people than seatbelts? Or just having three people across the back seat?

If you choose pooled options ride-sharing companies go by the nominal capacity of the car, so four passengers plus the driver in a standard five-seater car.


He probably means that one or more of the two, soon three, people in the back were rather large.


Yes and it was a subcompact sedan while both people had large backpacks on their lap


I am curious how hard a problem it is for the system to figure out what side of the road the driver should be on. Could they put a little arrow or something in the UX so I could point to which side of the street I'm on? Ideally before I order the car so the dispatch algo can prioritize cars that are more likely to approach from the correct side.

If the street is busy it can be hard as a pedestrian to safely get to the other side of the street. I've also had drivers just go the wrong way like I was standing between Gough and and Octavia on Market maybe 10-15meters from Gough. The driver came down Gough and turned left on Market at which point I knew it would be > 10 minutes before they could make all the correct turns to get back to pick me up. I cancel the ride and was surprised the system didn't direct them better.


It was a 3 lane one direction road and each time he went to the right hand side because the instead of trying the left handside. And instead of waiting for traffic to clear and thinking she might be on the otherside of the building he went to the back...twice.


Looks like civilian GPS accuracy is 4m or so. I could imagine that dropping someone's dot in the middle of a street, and then a driving having to guess which side of the street the error is coming from.


...or know that even numbers are on one side of the street and odds on the other as happens (I'd guess) everywhere.


Some places will put both evens and odds on the same side of the street if there's only one side, and then use more numbers if there is subsequently development on the other side of the street. Some places will assign a new house number whenever a new building is added, no matter where it is located. Some places will number in clockwise (or counterclockwise) direction.


That’s good for you and for Lyft, if you’re going to protest an extra person because you barely all fit, you’re not a good candidate for shared rides.


I was more bothered by the incentives which require you to spend forever waiting for a driver to figure out how to read a map, locate people, while also being unsafely crammed into a sedan. All because the driver wants to make a few more dollars. Which he could have earned more via tip and a positive review. But mostly is the massive time variability as they could accept multiple rides on the way and reroute each time.

I’m sure it works fine most of the time but I’ve never been a patient/people person who feels it’s worth my trouble to save a few dollars.

My father had a stint driving cars and he was appalled by the story since it was common to a) treat customers with respect (including their time) and b) expertly know how to navigate around a city.

If we’re lowering the denominator of quality of service with casual drivers then I’d rather offset that risk by reducing such responsibilities and variables on the drivers.


In this case, the driver is supposed to mark the other passenger as a no show after 60 or 90 seconds. There's a timer that counts down on the driver app. Many drivers will wait longer, however.


This kind of service could probably be impproved with route optimisation and booking using an app. A certain delivery company is already using route optimisation for parcel deliveries so that their employees don't have to think about it, just follow a set course. I don't know if it's enough, though. People would rather take ride hailing services out of convenience anyway. Maybe it could work for smaller vans 6-9 seater with 2-3 stops max instead of cramming up into a shared ride.


I used to use a shared van service fairly regularly. After the second or third time my ~1 hour airport drive turned into a 2-3 hour late night marathon, I stopped doing it.




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

Search: