Hacker News new | past | comments | ask | show | jobs | submit login
The Uber Dilemma (stratechery.com)
301 points by darwhy on Aug 14, 2017 | hide | past | web | favorite | 91 comments

Here is an educational game which describes this in a fun way. http://ncase.me/trust/

I enjoyed going through these scenarios quite a bit. Now with the context of this Uber Dilemma post, it's reinforced further.

It's worth playing every single game Nick Case has ever made.

Incredibly though provoking and makes you feel very different about the issues by actually involving you. I think there have been few pieces of media that emotionally impacted more than Nick's work. This is, what I feel like realizing the full potential of moderm media looks like.

Some links:

EDIT: I should have probably also mentioned the fact that the Author has made a site collecting these types of sites: http://explorabl.es/

WE BECOME WHAT WE BEHOLD – a truly shocking game about mass media: https://ncase.itch.io/wbwwb

Coming out Simulator: https://ncase.itch.io/coming-out-simulator-2014

The most famous one, Parable of the Polygons: http://ncase.me/polygons/

behold the beast! the magnificent 2d matrix!: http://ncase.me/matrix/

NEUROTIC NEURONS: http://ncase.me/neurons/

Wow, is there a genre or something for things like this? Seems like a pretty novel medium for expressing non-trivial idaes.

Yes, as mentioned these are commonly called explorables.

https://reddit.com/r/explorables http://explorabl.es http://explorables.cmucreatelab.org

His writing is pretty solid as well.

That you for posting! Never saw the discussion on here, or knew how many awesome things nCase has created.

So cool, thank you for posting. I shared with my nephew and he loved it. Made game theory really come alive for him.

I also became a patron to the author. My first time ever doing so on Patreon. Way to go Nicky Case and please keep up these amazing programs.

Awesome game!

I'm guessing game theorists have these kind of tools, albeit more complex ones. DO you know of any of them?

I'm sure this has been studied a lot so I'm guessing there are graphs about the different rules and outcomes and such?

The animations and laughter are absolutely fantastic. They make a relatively mundane subject far more engaging.

How accurate are these simulations compared to the real world ? I understand the reasoning behind it, and it makes sense, but it strucks as an oversimplification of really complex things.

Kind of inevitable. We're trying to model complex things in a way we can understand, and that necessarily involves simplifying it until it's... simple enough. Simple usually means coarse. If your model is coarse, does that make it invalid? No, but the coarser it is, the less accurate & predictive it is across all values of the inputs.

It's rather like metaphors. A good metaphor can absolutely aid in communicating and understanding concepts. But, inevitably, if you start probing too hard on the details or take the metaphor too literally and it starts to break down.

That's really neat, thanks for the link!

Thanks for sharing this! Loved playing it.

“There was no way we were ever going to take money from Sequoia, given what they’d done to me,” says Parker.

For those who don't know the story.

Mark Zuckerberg's Brutal Prank On Sequoia (2010)


the full excerpt is on Fortune... great read


Does anyone recall which VC firm Jim Clark had beef with after being betrayed during his time at Silicon Graphics?

New Enterprise associates. It's in the New New Thing by Michael Lewis.

I'm not entirely sure now. But it's covered in the book mentioned here and by the other poster.

KPCB, I believe. ("The New New Thing" by Michael Lewis is a great retelling of Jim Clark's adventures)

Huh interesting (and my first hunch was KPCB as well), since he works with KPCB in his subsequent ventures, including Netscape (I seem to recall that his beef was with one particular partner rather than the firm as a whole, so maybe that is why).

The valley grudges by top executives has always been amusing and entertaining to me -- for example, Ben Horowitz's refusal to work with E&Y after they nearly blocked their acquisition by HP.

As a side note the cap table of Netscape

Why was Netscape a weird example (to me) of Equity Sharing between Founders


I am not certain I understand how the reputation damage is supposed to occur in this case.

According to the article Sequoia had a long standing willingness to play hardball. A reputation earned through consistent long term behavior.

It's hard to see how a one time event in extreme circumstances leads to a change in reputation. Presumably any rational founder would assume that any rational VC firm would act the same way in similarly extreme circumstances. One doesn't earn a reputation by being struck by lightning or spraining your ankle falling over a turkey in a city street or any unrepeated unlikely events.

And, if you believe you can clearly prove "fraud, break of contract, and breach of fiduciary duty" presumably you are only discouraging bad actors with whom you wouldn't want to get into bed with in any event.

Suing a company you funded is extreme enough that doing it even once attracts a reputation. Would future founders considering potential investors assume that Kalanick was so far out of line that there's no way anything similar could happen to them? Or would they think that there's probably blame on both sides, and there are better investors around?

Why does funding a company have to mean that you are forever beholden to the CEO's bad decisions? Extreme this is not, and founders should take this as a signal that Benchmark will bring reason to the board.

>Why does funding a company have to mean that you are forever beholden to the CEO's bad decisions?

Huh? It doesn't?

Remember CEO=founder in this case. It's not "Benchmark will keep your board in line", it's "Benchmark might sue you".


It does. Benchmark played along with Kalanick until they realised this guy isn't in for making money (right now) but highly egoistic like Bezos to build a global economy. But now that they're suing the guy whom they supported for fraud means they play at crucial right time and no Founders vision matters to them!

Agreed. Speaking strictly personally, as a wannabe entrepreneur dreaming of billions, this doesn't give me the feeling I should avoid Benchmark.

That is the first time I've heard the theory that the lawsuit might be a sort of stealth way of Benchmark getting on the record as not having anything to do with the current Waymo issue. Does anyone know if a board member can separate themselves that way from a suit in order to shield themselves from damages?

If I understand the risk correctly it would be that Waymo prevails, gets a huge damage settlement which the company has to pay, but doesn't have the cash to pay so the board of directors have to go to their resources to pay. Is this understanding correct? Does anyone have a precedent for this? I'd be interested in reading it if they did.

Breaching the corporate veil is definitely possible. An article from 1991 mentions ~2000 cases where it was attempted, 40% of which were successful[1]. However, this is much more common with smaller and more closely held corporate entities. None of the piercings attempted against public companies were successful. If I were in Benchmark's shoes I would not be particularly worried about that.

I think personal liability for their director is a somewhat more likely concern in the same vein. Smith v Van Gorkam[2] is a relevant case for that. Intentional misconduct is definitely something that can trigger this. I'm not sure that getting their obliviousness on the record is a particularly helpful defense against this though.

Overall, I doubt this theory is a major motivation on Benchmark's part. Benchmark obviously has billions at stake here and has also obviously concluded Kalanick is toxic to those billions. Seems reason enough to me for scorched earth legal tactics.

[1] http://scholarship.law.cornell.edu/cgi/viewcontent.cgi?artic...

[2] https://en.wikipedia.org/wiki/Smith_v._Van_Gorkom

It would be highly unusual and I dare say impossible for Uber shareholders to ever have to pay a settlement out of funds not already invested in the company. This is the essence of limited liability.

The exception here is if the firm is found guilty of criminal activity and evidence emerges to suggest that Benchmark knew about it. In that case I don't know that they would be financially liable, but directors and board members would stand the risk of jail time.

AIUI the "Limited" company in the UK limits a shareholder's liability to the value of the shares they hold - so if you hold £1M of shares that's your liability. I assume it works similarly in USA?

Yes correct. The worst financial outcome for holders of Uber stock is that their shares become worth zero. They can't go negative.

If the Saudis were misled in their $3.5B investment? IANAL, but individuals might be held accountable.

Maybe Benchmark is hoping for a situation like Zenefit's. Ousted the CEO and accused him of fraud. Then they effectively increased their investor and employee ownership stakes at the former CEO's expense.

The blog pushes two potential outcome, but fails to realize a third - Benchmark, as one of the most successful VC firms ever, is messaging to CEOs that certain behavior will not be tolerated, even if they are going to be the next Google and Facebook.

That is an awfully expensive way to send a message to someone that may not exist

Maybe as a by product :)

Realistically, how long do I have before Silicon Valley stops subsidizing my transportation?

Honestly, I think they are subsidizing the drivers. I think the play is to subsidize rides to prop up user counts until they can sell those users to a self-driving car fleet. (or own one themselves).

This would be my guess why a GM or ford (Can't remember which) would invest in someone like lyft.

There's nothing to crazy here outside of an app that pairs two GPS coordinates. Access to user pools for self driving car fleets is going to be big money though.

This is why uber could be in so much trouble over the otto/waymo thing. If they get barred from self driving cars, their competators will be able to offer the same or lower rates without having to pay drivers. Uber will have to continue to subsidize until then or loose mass user base.

The numbers just don't add up for this scenario:

Current valuation is 70 billion, Last known monthly active users is 40 million. Monthly spend per user is $50.

Meaning: instead of buying those users for 70 billion, a car company could offer each one of those users $1700 of free credits, or about 3 years of their typical use.

Considering that switching is as easy as installing some other app, I'm pretty sure they could buy marketshare for less than a 10th of that.

...and I haven't even factored in that the novelty of self-driving cars would probably make it incredibly easy to sign up users for whatever car company is first-to-market.

But the drivers really don't make anything.

Haha I have the same question. Until self driving cars are common presumably, but I think a lot of people are deluded about how soon that will be.

I think it will be soon because there is a huge number of markets to test in. You have every state in the US, every country in the world, and a dozen players all working on it. In an n*n space, you just need one technology+municipality pair that says "ok, we'll risk it" and you have a very public, newsworthy demonstration.

If there are real benefits to the tech, it will be impossible to keep them off Instagram. From there, it will spread to opportunist governments... Places who will try anything if it draws interest to their municipality.

At some point after that it becomes a liability to not allow it.

Now, how long before each of these stages? I don't know, but there's pressure there.

This explains how self driving will be legalized, not how it will be perfected, and it's not even close yet.

Each forces the other

I'd love to be proven wrong. But, progress made to date notwithstanding, getting to Johnny-cab as commercial service (i.e. door to door in city without a human even present) seems an incredibly high bar to be remotely relevant to a VC's planning horizon.

Not sure I would ever want to step inside a Johnny-cab https://www.youtube.com/watch?v=eWgrvNHjKkY

Tesla sells "Full Self-Driving Capability" option for Model X and S on their website right now.

The summary on TIT FOR TAT is not complete:

when you’re playing for the long run it is better to be nice — you’ll make up any short-term losses with long-term gains.

It also depend on the population distribution. If everybody else is mean, being nice in long run still lose. So the optimal strategies in different culture are obviously different.

I think TIT FOR TAT is good because: 1. it give a gesture for cooperation at first 2. but respond immediately if the response is hostile.

There were some variations of TIT FOR TAT, like being nice for several rounds, then revenge later. Those variations didn't work well.

The random strategy(being nice or mean randomly) also doesn't work because it's difficult for others to predict your behavior.

The experiment was designed for long term survival, so it doesn't apply to these cases:

Being nice for all the time, then use the reputation to scoop the biggest gain in the last step.

So the experiment is very interesting and revealed a lot of points, but the simple takeaway of "being nice is good for long term" is just too limited.

Here is a fun 'game' to illustrate some scenarios around this. http://ncase.me/trust/

Here it's framed as the game theory for the Evolution of Trust.

How interesting. If I read between the lines, it would make fiscal sense to be ruthless and Cutthroat early on. Then after guaranteeing some certainty of income, sell the business to a nice player.

That way, you could get the best of both worlds being a bad actor in the prisoner's dilemma and being a tit-for-tat member. They only hope would be is that nobody finds out this grift.

>when you’re playing for the long run it is better to be nice — you’ll make up any short-term losses with long-term gains. It also depend on the population distribution. If everybody else is mean, being nice in long run still lose. So the optimal strategies in different culture are obviously different.

Hence the age-old debate over whether God is vengeful or compassionate

> Does it give the firm a bad reputation, potentially keeping it out of the next Facebook? Unquestionably.

I question this assertion.

As others pointed out in the HN comments for the lawsuit announcement, the only people who have to worry are those with questionable ethics. We need fewer of those people in startups so this is a positive signal to me.

This is compounded by the fact that brand names don't get tracked but this post, ostensibly by someone in the VC business, doesn't even mention: who is Benchmark? Starting a new firm with a new name might have zero reputation, but it's better than a negative reputation, and who's going to do extensive research on the founders of a new VC fund, as long as they're interested in your new app, and you weren't personally aggrieved by them in a previous round.

> and who's going to do extensive research on the founders of a new VC fund

...Anyone who actually cares about their firm and is smart enough to actually attract VC interest?

False although I grant some will.

Uber already is one of the grand slams, so it would make sense to take action to protect that success (I'm not saying Benchmark is definitely doing that, but I'm sure they think so). Even if you don't agree with what Benchmark is doing, I think it's hard to see this as anything other than a fight between billionaires. If you're just starting out, deciding to hold a grudge against them over this is getting way ahead of yourself.

This kind of stuff plays out at much smaller scale too. You just don't hear about it in the national news. So some due diligence on your investors is worthwhile if you are going to pour the next 5 years of your life into something.

So cash out or grow, but they're already shrinking and cashing out is looking increasingly hard and with diminishing returns. This may not end well, but presents a lot of opprotunities for those with the right eyes.

So Lyft played nice and is having more long term success as a result?

Not really.

Yes Lyft has gained some marketshare in the past 6 months. But Uber still dominates the US rideshare market (Uber by public accounts still has greater than 70% Market Share). Uber dominates in several other regions (Western Europe, Latin America) and is competing in several other huge markets (India and SouthEast Asia). Uber also has significant stakes in the dominant ride hailing companies in China and Russia.

This is not to diminish what Lyft has done. They're successful company. But they are in no way seeing more long term success compared to Uber as of today

Any pointer to "Uber dominating in Western Europe?" Only in a two horses race with Lyft, but as you know there are many other racers, including traditional taxies which are well organized in Western Europe, not only a vested interest.


To be fair, this doesn't (I believe) take into account if traditional taxis are still dominant in some countries

Did anyone do the math on using Uber vs buying a car? I probably spend like 100 per month on Uber

This is incredibly specific to an individual, so specific that the math would have to involve an algorithm with probably 15+ inputs.

Consumer Reports and the like have calculated "true cost of ownership" for different model cars you can use as a rough guideline. They estimate depreciation, taxes, financing, gas, maintenance, and insurance over 5 years. However, that's just a rough estimate and only a first step, it doesn't take into consideration individual variables. If you are a young male, have to drive in stop-and-go traffic every day, and pay to park everywhere your cost of ownership will probably be close to double.

Auto insurance alone is probably close to $100 a month for the average American - depending on your car, location, and driving record, of course, I'm just making an educated guess. IIRC, last I ran the numbers with my car (average budget model) cost of ownership was around ~$20,000 over 5 years. That's amortized to $333 a month.

I wish I'd done the math on Uber when I bought my car some years ago. I ended up paying ~$80 a month on insurance alone. Add in the initial cost of the car, maintenance and fuel, it was not at all worth it in comparison to just shelling out for a Lyft whenever I needed one. So by my math if you're spending $100 a month on rides, you're well in the black in comparison to owning a car.

The main (and perhaps only) advantage to having a car is the convenience and the ability to go where ever whenever.

You can certainly save a lot of money (depending on your location / usage pattern) by going carshare-only; but one thing I've found is that it can result in a sort of cognitive fallacy (not really a fallacy per se; rather a thought pattern that can be tricky to avoid) where every trip you make becomes subject to economic considerations, far more than it would have if you'd owned a car.

For example if you have a friend who lives 15 miles away - it becomes "do I want to spend $50 to visit my friend?". If you drive your own car, the economic calculation is invisible - it's about $2 of gas, a bit of depreciation that you don't even think about, and [huge already-paid upfront cost for vehicle/financing/insurance that you don't even think about] - "of course I'll pay $2 to visit my friend!"

One thing that helped was making a spreadsheet that showed how my total cost of transportation went from $8k/year to about $1200/year when I switched to Lyft/ZipCar and ditched my own vehicle. So even if I were to add in a weekly $50 friend-visit, I'm still paying less than half for my transportation than I previously was.

Based on the standard IRS deduction, a 30 mile round trip would be about $15. (And the marginal cost is less than that--your insurance at least doesn't generally go up by mileage except to the degree it slightly increases the probability of an accident.) But I agree with your basic point. If people had a clicking meter whenever they got into their car, they'd think a lot more before they turned the key.

Some might say this was a good thing. Personally, I think mobility is generally positive.

My insurance has always wanted to know annual mileage to determine the insurance tier to put me in. The tiers are pretty wide though, so I never put any thought into visiting a friend. Just sometimes which of the family vehicles to drive across the state.

> where every trip you make becomes subject to economic considerations, far more than it would have if you'd owned a car.

As it should, as well as an environmental consideration.

A lot of people (including myself) live in locations and have activities (including commuting) that pretty much make not owning a car a non-starter.

But it's certainly true that for people who don't need to drive on a regular basis, a lot of reason to own a car may come down to things like not wanting to deal with Zipcar or a regular rental car every time you want to do a big shopping trip or get out of town on a Saturday or for the weekend.

It really depends on a lot of things.

-- Amount that you drive

-- Availability of Uber when/where you need it

-- Whether you have to pay for parking

-- Whether you also take a lot of weekend trips

-- etc.

$100 per month is actually not very much. That suggests you're only traveling 100 miles per month or so on Ubers which is nothing. By contrast, that would basically take me to the nearest large city and back once or get me back and forth to my office a couple of times.

If you live in a city, you'd probably pay a pretty decent chunk of that amount just in auto insurance.

Do you live in a dense city? $100 would not get me anywhere close to 100 miles in the city I live in.

I don't live in a city and have only used Uber a handful of times. But UberX has been in the ballpark of about $1/mile when I've taken it to and from airports and that seems consistent with numbers I read online. Perhaps the cost is higher in a dense city with lots of lights and traffic.

$100/month is a pretty low transportation expense. If you buy a car, you're going to spend more than that for the next few years, but if it was a used car, you might pay ~$250/month for 3 years and come out ahead in 8-10 years.

If you go for a new car, you could easily pay $600/month - at that point it would take > 18 years to break even.

Of course, there are other things to take into account: convenience on the one hand, maintenance, parking, & insurance on the other.

But, my gut instinct is that you'd be better off to keep taking uber and invest the difference in an index fund. Keep that up for long enough and you can buy a new car in cash (or retire early).

[edit]: Don't forget that you can bicycle also. A $400 bicycle has an incredibly high ROI if you replace a few regular car trips with bicycling. It's a net gain for your health too! (You may already be doing this, but it's worth pointing out anyways ;)

> If you buy a car, you're going to spend more than that for the next few years, but if it was a fairly inexpensive used car, you might pay ~$250/month for 3 years and come out ahead in 8-10 years.

Do those numbers include maintenance and insurance? IME just routine maintenance on an oldish (60-100k) Toyota runs into the hundreds of dollars. Insurance again something like ~$800-1000.

I got rid of my car after I did the math and realized I was spending $1500-2000 a year on just insurance and maintenance.

Not explicitly. Although in my case (2001 civic until about a month ago) car insurance was less than $300/yr, and repairs & maintenance was probably a bit less also, probably under $1000 most years.

What insurance were you using? I've looked into some of these cheaper insurance companies, but I worry about whether they would really stand up for me if something did go wrong. One scary story I heard from a friend of friend was related to this cut rate insurance company in NJ. They assigned 20% of the blame to the friend even though it was clear cut case where the other guy didn't yield on a left turn.

State Farm. Been with them for a long time and never had any complaints.

My wife and I have never had any significant trouble (tickets/accidents/etc.) though, and it was collision insurance only. We also don't drive much. (I bike into work or else work from home.)

(Collision = cover other people's bills when you cause an accident. Comprehensive also fixes/replaces your car if it's wrecked, but that doesn't really make sense on a 10+ year old car.)

We also have home and life insurance through State Farm, so there may be some bundle discount going on.

I did. It was cheaper to own a car (but not by much).

My specifics: I live in Cupertino and work at home. I have a wife and kids. About four times a month we have a situation where we need two cars, because one of us has to be somewhere other than where they kids need to be.

Unfortunately, most of the time that's San Francisco, which is about $80 round trip with Uber/Lyft. I could sometimes use public transit, but usually the times don't work out and it takes a really long time, and I'd still need a ride to and from the train on both ends.

I have a five year old car so the payments and maintenance still come out to a little be less than Uber/Lyft, and since I've already got another car, the cost of insurance doesn't add much.

It's pretty close in cost, and the convince pushes it over the breakeven point.

I bet you're underestimating what you spend on Uber. That's only about a dozen small trips.

12 trips a month makes the calculus easy: don't buy a car.

Depends if you need to travel long distances (over 50 miles). If the answer is yes, Uber is simply not a good option. Otherwise no.

Not quite, it depends how often you need to travel long distances and what the other options are for that travel. If you only need to drive 100 miles four times a year and otherwise have zero everyday commute costs because you ride a bicycle to work then renting a car four times a year is almost certainly a better option than buying a car just for those trips.

Additionally, if there is frequent and reliable train or bus service to your destination that may be a better option than driving. I know several people who don't own cars but travel 100+ miles via bus on a regular basis. On the other hand, I travel 100+ miles regularly but buses and trains don't go to where I am going.

(There's people who justify buying a truck because they need to haul stuff twice a year at the most. Renting a truck twice a year is almost guaranteed to be a better option financially. That being said, if you want a truck because you like trucks, that's also a perfectly valid reason to buy one.)

Agreed, I greatly oversimplified things. There's the whole matter of convenience too which I didn't even touch upon. As it happens I have a car (weekend trips) and use Uber too (during the week), and use public transportation. Convenience is the driving force behind my usage.

Thanks for an extra 1000 words explaining the prisoner's dilemma.

Edit: that was sarcasm

The article was mostly about VC-Uber tug of war, though. That's the essential point of the whole debacle.

I feel like these simulations refute "if you are not a liberal at 25, you have no heart. if you are not a conservative at 35 you have no brain":


The older I get, the more it seems to me that context is the basis of strongly held political beliefs. Growing up in Idaho, I'm often dismissed out of hand when I offer my (by regional standards) liberal opinions. I've met many people from liberal states (even California) who offer conservative opinions, which always struck me as ironic. How would our positions hold up when confronted with a change of context, if we saw how their application backfired?

I'm wondering if there is a duality here, where different approaches can both be right but one could and should be chosen based on desired outcome. People unaccustomed to pragmatism may be performing a logical fallacy by assuming that their way works in all cases, when it practice they would lose to experience (and in the case of the prisoner's dilemma, cooperation).

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