What would a decentralised Uber look like?
By @edent
· ReDeCentralize uber · 9 comments · 700 words · read ~3,194 times.
Uber are undoubtedly a company engaged in extremely dodgy activity. But, on the other had, they're ridiculously convenient.
A few months ago, we landed in a foreign country, opened the same Uber app as we used back home, and booked a cab. It just worked. I didn't need to register for a different version. I didn't need to create a new account. I didn't need to add a new credit card. That's the sort of seamless experience which can only come from a centralised service.
But, hey, we're all moving to a ReDeCentralised Federated Future. Nae king! Nae quin! Nae laird! Nae master! We willna' be fooled again! So let's think about how a decentralised Uber would work.
Broadcast
In the days before mobile phones, we could have used something like CB radio. Broadcast a message to everyone in a few miles saying you want to travel from point A to point B.
Cabbies, or their agents, could reply to you with their availability and prices.
But that puts a lot of work on you to manage all the replies. And you have no idea of the driver rating. And you'll be broadcasting your location to all and sundry.
And, sadly, modern handsets have no concept of a broadcasting in a P2P manner.
Perhaps a little centralisation?
In most parts of the world, taxis are regulated. You don't want rapists, thieves, and murderers driving you home late at night. You want a predictable fare calculated by an honest meter.
So let's say you have an app which knows your location, and send your request to the taxi-regulation-boards which cover your area. They can then broadcast out the request to all drivers, they confirm the fare, and then send it back to you.
I guess that could work. Think of all the taxi-regulators as little ActivityPub servers, and the cab drivers getting a stream of requests on their federated timeline. Your phone could receive "accepts" from multiple drivers and you could pick the one you like.
But how would your phone know which local taxi services are available? You would need a centralised list of them!
FOAF and Signed Claims
One of the supposed advantages of Uber is that it allows participants to rate each other. Passengers don't want to get in a car with an arsehole. Drivers don't want to pick up people who litter.
How would that work without a central entity? Could there be a solution with Verified Credentials? Have all participants write cryptographically signed ratings to a *spits* Blockchain?
User Needs
Of course, before sketching out such a system, it probably helps to sketch out some vague thoughts on what users actually want from a taxi system - and how decentralisation would help.
Firstly, who are the users?
As a driver, I don't want to have multiple phones and apps stuck to my dashboard.
As a user, I only want to use a reputable driver. But who determines reputation? What options do I have if something goes wrong?
Basically - what are the advantages of decentralisation in the taxi space? Users want competition on price and service - which they don't get if Uber are a monopoly. Users also don't want to deal with dangerous drivers - which is easier to enforce if there is only a single platform.
Drivers don't want their livelihood ruined by unjust reviews - how could that be enforced with a decentralised platform?
We haven't even got to payments yet!
Basically, it's not clear to me that is a compelling user need for decentralisation. Don't get me wrong, I like it as a concept. But I see it as exchanging one set of wicked problems for another.
Does this already exist?
Well, ish. There are various different attempts to build federated delivery services and the like.
But I am very interested in being proved wrong! Is there a decentralised takeaway service I should investigate? Are there federated hairdressers? Does anyone offer an open protocol for domestic cleaners?
How do you decentralise emergency alerts?
@Edent Rideshare apps are just better experiences all around for riders. I try to use Lyft instead of Uber in most cities. In NYC I try to use Driver.coop which is a rideshare cooperative corporation that is owned by the drivers. I really hope that model takes off in more places, Uber is always last resort for me. drivers.coop
@Edent I've been pondering how this would work in the scientific review space too... I guess you'd have a few "primary" instances run by institutions that everyone recognises as good measures of quality. Either way, I reckon it could all work!
@Edent parking apps need to be federated firstThey are awful right now
@Edent there was briefly a "taxi" app that essentially aggregated many of the local private hire companies. It worked something like a combination of an insurance comparison site and Uber: say where you want to go and it'll give you a list of quotes for the fare; pick whichever you want, and the app handles booking, journey tracking, and payment - just like Uber.Sadly, it closed a few months after I discovered it (unrelated events, I'm sure).
What would a decentralised Uber look like: shkspr.mobi/blog/2022/12/w… Comments: news.ycombinator.com/item?id=341857…
@damien @Edent In my town all the main taxi companies use icabbi (which works quite well, at least from a passenger POV) but due to being competitors have separate branded apps which go to exactly the same place! BITD the taxis did use VHF radio schemes (private business radio licence from DTI/Ofcom), but 10-12 years ago I noticed the radio traffic from these was disappearing and the drivers themselves told me they had all gone to using GNSS enabled apps..
@vfrmedia @Edent Karhoo is the name of the one I was trying to recall earlier: https://www.failory.com/cemetery/karhooAt least around Manchester, the majority of private hire firms around that time didn't have their own apps, or they were very poor. What Happened to Karhoo, the B2B Cab Marketplace?
@damien @Edent Karhoo appears to have been bought by Groupe Renault although now sold as an add-on service to other transportation companies.. I suspect the reality is the profit margins for a taxi service are tight and there's nothing that tech can do to improve this (in my town there are only two main taxi companies are they charge pretty much exactly the same prices and are both suffering a shortage of drivers due to fallout from Brexit and Covid..)
DinoNerd says:
Well, one option would be to make it look like plain old taxi service, but with a common front end.
The authority that issues taxi licenses (medallions) cares about not having outrageously bad drivers. (Not just poor driving, but also egregiously bad behaviour.) The various taxi companies care about not having egregious drivers driving for them. If the car is owned by the taxi company, the company cares even more.
You don't get surge pricing, and you don't get J Rando trivially able to take up driving for pay, in the car they normally use for commuting - but that just might be a feature rather than a bug.
Hmm - I wrote this and then saw I was saying much the same as several other commenters, and one even pointed out this was already being done, though his example might not work well for someone visiting from far away.
At the risk of being redundant, I'll post this anyway.
More comments on Mastodon.