I'm not sure if any of this actually saves you money. If anything, it may drive the algorithm to pair you with less than stellar drivers. In a customer retention phase, which Uber is in right now, the algorithm should aim to pair good clients with good drivers and poor clients with poor drivers, just to protect the repeat business.

> I like how people in comments are keen to change the world, but I - more realisticly - only focus on gaming the system so I can actually save myself couple bucks right away.

This type of attitude sometimes makes the world a worse place. If everyone has this attitude, the system can break down.

> I rarely rate drivers positively

Drivers live or die by their rating. Refusing to give good reviews harms the drivers who are already barely scraping by. This is a very good example of how you can cause real world harm with by trying to game the system for yourself.

I don't know if this saves you money, but if it is, you are certainly working pretty hard for it.

Wouldn’t / shouldn’t this heavily penalize you? Like, what I want to optimize for is that when there’s little Uber capacity and many requests, that I get prioritized and get put as close to the front of the queue as possible. Which is why I tend to stick to just one app (Uber), and try to have a high rating.

But idk if it’s doing anything. At least back when Uber had tiers (platinum / diamond) it was more obvious the perks you get with more use (like with airlines), but now that they dropped those I’m feeling less inclined to keep using just Uber vs a mix of all apps and picking cheapest.

> I set pick up and destination, exit the app, open another rides app, wait few minutes for uber to notify me that the price went down.

Just tried this, didn't get any notification.

Thank you for these tips.

I do similar to you. I wonder if not rating is limiting your gains relative to giving (gasps) four or three star ratings.

> Probably false, like the urban legend that Uber charges more when a device's battery is low [0].

That article certainly doesn’t prove that was an urban legend. Uber built an entire system to serve fake data to government officials

https://www.bbc.co.uk/news/resources/idt-f2971465-73d2-4932-...

https://www.nytimes.com/2017/03/03/technology/uber-greyball-...

They were almost kicked out of the App Store for building a system to deceive Apple’s employees, too:

https://financialpost.com/technology/personal-tech/uber-was-...

Given the other ways they tried to intimidate journalists, anything like this which is conducted with the journalists’ own phones or near places they frequent is inadequate to say whether this actually happened — especially after it was going viral and all they’d have to do is flip the switch and lie about it until the attention dissipated.

The fact that its unfalsifiable is the problem. They can (and why wouldn't they) run all sorts of shenanigans with their pricing algorithms.

Pricing algorithms should be not be private. Free markets depend on transparent pricing.

As usual, the real story is buried beneath garbage anecdotes like this. The multiple elephants in the room include

- positional pricing, certain geocodes command what is called pricing elasticity, aka. people going to the hospital dont care about the price. Time based pricing elasticity is also exploited, for example, people leaving the concert at 12am likely dont have other options

- driver bonuses and rider coupons optimizing for marginal rides. The riders and drivers at the margins of pricing are typically switchers, which mean that if you are loyal to any particular app, you will likely never get see these coupons or bonuses

I believe data from all New York City trips from companies like Lyft and Uber since 2019 [1] can be downloaded from Taxi & Limousine Commission [2]. This data appears to include information such as the fare, date and time and TLC taxi zones for the pickup and dropoff [3].

I wonder if there is enough granularity in this data to make a determination if there's a large discrepancy/variation in fare in certain taxi zones at the same time with Uber, compared with Lyft or other competitors.

[1] https://www.nyc.gov/assets/tlc/downloads/pdf/trip_record_use...

[2] https://www.nyc.gov/site/tlc/about/tlc-trip-record-data.page

[3] https://www.nyc.gov/assets/tlc/downloads/pdf/data_dictionary...

Not for battery, but Uber definitely charges riders different prices. I've done the experiments (accounting for order, rider rating, discounts, etc) and seen differences as big as +50%.

Uber support also confirmed it, after a lot of question dodging (emphasis mine):

---

Thank you for reaching out to us.

The prices of the trips and promotions are unique to users, that is correct. If you have in the future any trip that you feel like it's too expensive, please open a contact, attach the correct trip and we will gladly review it. There is sadly nothing you can do personally to change the fare rate of the trips before any discount is applied.

We wish you a nice day.

> I checked the price of an Uber with credits in my account against Lyft to the airport just now, and Uber was slightly cheaper.

I appreciate that you checked and that you acknowledged the limitation of it as an anecdote, but it doesn't actually give us much information.

Companies are always doing experiments behind feature flags on subsets of the population. Some of those experiments would surprise the public. At any given time, the set of experiments that would surprise the public are only ever seen by a tiny slice of users.

People outside those companies will never hear of most experiments. And often only a few people inside the company will be aware of them.

Something like this is more like an existence proof (does it ever happen?) rather than a proof of universality (does it always happen to everyone, or even to most people?)

> Claims like this go viral because they're practically unfalsifiable

How in the world is this unfalsifiable? Can't you just have two Uber accounts and cross check their prices?

Probably. Still, a pricing algorithm may be based on all sorts of data points that get fed into an ML model. How sure are you that battery life or presence of credits don't make it in there, even accidentally?

> or by multiple phones looking at the same route implying increased demand (the latter search might display a higher price)

This post sounded like it's saying that the later search is the one with the lower price.

I mean it's unfalsifiable because the marching forward of time makes asking a system to tell you a price is totally black boxed. But it would be nice to see a _bit_ more effort placed.

> A proper experiment would involve dozens of phones under different scenarios making searches in a random order, then trying to correlate the variables with the prices. But for whatever reason, nobody ever does those experiments.

Because of how much Uber spent doing counter-intelligence against every locality in the planet, I feel like even that might not be good enough. I could see them "knowing" what's going on!

It reminds me of legends: "YouTube listens to you and then recommends videos based on your offline chats"

Seems very falsifiable, as indeed the author tested. Get two people to get the same ride. Check pricing. Repeat.

Interesting - you're not logged into Maps?

The rideshare tab in Maps gives me time/distance/traffic estimates and no prices. Whether I'm logged in, or Incognito. No prices unless I tap "Open App ->"

There is no rideshare tab in the desktop version, only tried it on my Android.

If you can do this repeatedly, say 15 times, and record the numbers each time (say just put them in a spreadsheet), you can actually pretty easily run a "statistical test" to prove your theory at 95% confidence (which is what research journals often do).

You obviously don't need to justify your observations to anybody else, but if you did find a conclusive result it would make a good blog post.

[P.S. chat gpt could help with the statistical test part]

> People are assuming that some evil mustache-twirling PM/engineer decided to charge users more if they have credits in their account.

> The ML model, after much historical analysis, figured out that it can charge users more if they have credits available. The relevant engineers treated the ML model as a black box, ran an A/B test, saw a 1.3% increase in revenue, and decided to roll it out.

This is the same thing. The fact that they might be turning a blind eye to why the ML model optimised the way it did doesn't absolve them from responsibility.

You might be right, but I'm failing to see how this is less evil mustache-twirling. A PM/engineer decided "I bet we can create an ML model that will find suckers we can charge more," is basically the same as "I bet people with credits will pay more." In either case, they're making a decision to find people they can screw over. In the latter case, it's based on intuition. In the former, it's based on statistics. However, the motivation remains the same.

If people are thinking this is evil, it's the motivation and the result that matter, not the means through which they achieve their goal.

And if they used ML as a black box, it could act in problematic ways. Maybe the ML figured out that it could charge Black riders more. They didn't look into it and just rolled it out. Things like that can happen when you don't look into stuff.

And I'd say that it's questionably legal. If a store charged you more money because you were using a gift card, I think we'd all agree that the store had unjustly taken value from you. Here, Uber is allegedly charging people with credits a higher amount - essentially stealing gift card value.

I'm not sure "oh, it was just our ML deciding to steal your money," is a great defense. Ok, but what was the ML supposed to do. Well, it was supposed to figure out who we could take more money from for the same product.

The act of feeding the model with data about users' credits necessarily means that they know it's using that as a factor in pricing. What difference does it make if there's an ML model involved?

At some point, the not giving a shit because it maximizes profit is all you need to be a bit judgy.

If they were doing something relatively frictionless that obviously increased efficiency? People wouldn't care.

Mindlessly doing what the model says, even though it's awful for the customer, to marginally increase profit some of the time by charging some customers more? People get pissed off...

That isn't that different than choosing to charge more for accounts with credits. Make "having credits" available to an ML model and then let the ML model "ethics wash" the decision.

Thats almost the same thing isnt it? The motivation is to charge as much as possible to each user, how it's implemented isn't really relevant

Is it less evil if an unsupervised process does it?

It already is.

Yes, ML models are for accountability laundering. More news at 11.

There are many people who believe vigorously in dynamic pricing and it providing the "best service", and don't need to be paid to believe so.

The difficulty arises because the counter-argument of "surge pricing can kick in way faster than new cars show up on the road" is, in its details, subtle. And "this is just grabbing money for no increased service" goes against many people's idea that people should be able to charge arbitrary amounts for services if they feel like it. "People shouldn't be bidding against each other like this for transportation" is yet another can of worms.

This is a space where there are so many sticking points that many people will disagree with. A place where you can really understand gaps between people's axioms, depending on what part you focus on. And it's a place where Uber sits comfortably in the conclusion of the simple arguments, and the opposition to its tactics lies in the conclusion of very long, nuanced, contextful arguments (that still, at one point, require acknowledging that it's not good for society as a whole for a service provider to have unlimited pricing power, which is a big ask!).

How would you word a regulation against this sort of thing to apply against the low man on the totem pole, the lowly SWE? Might it be better if it was regulated as unfair and deceptive by a strong FTC with a nice bounty for SWEs?

Instead of something merely sketchy and probably not real, what about something clearly illegal and proven real like for example the VW emissions thing?

Or does being sketchy / probably immoral vs being actually illegal make a difference? Maybe something around twisting yourself in knots to justify something being corrosive in ways that knowingly doing wrong isn't?

Might come down to two variables:

- do they believe they're doing a job that is otherwise net positive to society or their users ?

- how hard would it be to successfully argue against this feature getting implemented ?

For instance I'd they think Uber is saving humanity but this feature would cost their whole career if they went hard against it, the remorse would be negligible.

For others it could be the last straw breaking the camel's back.

In general I think we're turning the blind eye to many shitty practices if we assume it's a small issue in the grand scheme of things. I mean, we wouldn't have so many dark patterns and spam emails from almost every companies we interact with if it wasn't the case.

When was the last time you went to tell your marketing department to dial it down ?

I suspect it's less apathetic than that; more "sweet, this'll get me a bonus!"

They made me an offer I couldn’t refuse: health insurance.

I think all the people that have moral qualms have already refused to build such things. It doesn’t matter, because there’s a whole line of people in it only for the money that can take their place.

Instead of regulations, I think this is the kind thing that should be driving a push to unionize tech. For the most part we don’t need to worry about dangerous job sites or low pay, but we do have to worry about unethical business practices.

Specifically, if we had widespread “codetermination”, which gives board seats to union members, the people that build things would have a say in what they’re being asked to build.

Not a dev, but someone who went through a relatively shitty process implemented by a SaaS company at my first job.

The general idea was enshitification once new subscribers dried up. We auto renewed everyone that we could into a drastically different, and more expensive, pay structure then forced them to call if they wanted cancel.

I knew it was scummy at the time (circa 2012), and now wish I had at least SAID something. As a fresh college grad still, I needed the job and stability.

In an ideal world, we would empower workers to stand up for what is right, but every step of the system is designed to take that away from us. Having a moral choice is a luxury in today's workforce.

I could also imagine that there is lot of gaslighting involved to feel good about themselves when making these features. For example, they will talk about in terms of prioritizing Ubers for customers because they need it more urgently hence the the premium. So, (1) you can comfortably talk about developing these features with each other (2) even though you know its BS you don't care because you know its gonna lead to good perf review (increased revenue by X%) => bonus.

Kind of how Apple justifies crazy App Store policies/pricing for keeping "customers safe".

No one even had to code them necessarily. Imagine building a purchase prediction engine. You could easily accidentally slide credits in as a feature. In the same way that you have to be very careful in ML to not accidentally include either the answer or a close proxy for the answer in your training data and have the model learn that it should just pick the proxy for the answer.

Regardless, I'm also not clear what I think about government regulating the price of a service like Uber. My first thought is probably more harm than good. We mostly don't regulate the price of airline tickets. I don't see how Uber is any different.

Dynamic pricing is supposed to be based on the overall demand of the service at that instant. In this case, it is “price discrimination” based on the user, rather than the overall demand.

I noticed this exact behavior between my wife and I. I avoid Uber like the plague but unfortunately they have destroyed the alternatives. We need a right to anonymity for services. Knowing who I am isn't a requirement for the actual service but they want it so they can make additional money off of you. There should always be an option 'remain anonymous' that requires the vendor to immediately delete any information on you and use nothing identifiable about you in their service. If they are found to break that then there should be severe penalties that are paid directly to the person impacted and all lawyer fees go to the company with no option for 'arbitration'.

Anecdotally, I find if I go to Lyft, then go to Uber, and back to Lyft 30s later, I often get an updated decreased price.

Nitpicking on your one claim that "taxi prices were transparent" - in some countries, maybe. In others, taxi drivers still have very creative ways of scamming their customers, with or without meter. This was actually the main reason I recall that Uber and others used to promote their services, clear upfront prices.

During their growth years when they arguably did the shadiest stuff they also had record-breaking quarterly losses in the billions while their counterpart DiDi in China was doing well enough to acquire Uber China (who according to Travis Kalanick was losing over $1b a year).

My credit card gives me monthly Uber credits.

Anecdotally, can confirm.

And the trick of course is to first get the deal (which lasts for 1-2 weeks typically), and then add credits.

That's not how Spotify payouts work.

Full self-driving will kill it

I don't think it's necessarily the case that dynamic pricing shouldn't exist at all (eg, cooperative load shedding on the electrical grid).

However I should, as someone who just left a concert, have a notion as to how much it's in my best interests to either rideshare with my friend and deal with two stops, or walk up the street to a restaurant and kill an hour or two while we wait for the chaos to dissipate. If the price difference is going to pay for a drink or an appetizer, then maybe I voluntarily timeshift.

And we don't seem to be near that bar at the moment.

IMO, totally banning dynamic pricing is too far. Wouldn't it solve this problem just as effectively if dynamic pricing were still allowed, but the algorithm had to be fully open?

there's nothing wrong with dynamic pricing, but there is a lot wrong with price discrimination, bait and switch, hidden fees, and outright fraud. (and we know that uber has historically engaged in many highly dishonest practices)

here's my favorite rideshare gouge, I'm sure others have had the same experience.

"currenly 20 minute wait, upgrade to uber+$ for 3 minute availability?"

YES!

"OK booked, your ride at the higher rate is on the way, currently 20 minutes away!"

now, it is possible that "offer" message was shown to a bunch of people at the same time and somebody snagged it in front of me leaving only the other car 20 mins away, i.e. the computer was "honest", but I shouldn't be paying the higher rate. clever programmer types can come up with other scenarios, the point is, we need guarantees that the market making is honest, or we need regulation.

A great way to fight dynamic pricing is to force companies to compete with the government.

Uber is going to have a lot of fun pricing a market where the bus is 10x cheaper and faster than the Uber thanks to dedicated lanes.