[recalibrator]

Sam. You own a Tesla Roadster, a car for people who don't have to niggle over money.

Most people a) don't live in SF b) have your money c) have access to Uber and d) have a Tesla Roadster.

You're essentially talking to yourself and a few choice friends who live in the same bubble as you.

[tdicola]

Not to mention since the car is paid off/owned outright the calculation doesn't take into account the price of the car. I would wager a majority of car owners are financing or leasing their cars and pay a monthly cost to do so. I guess depreciation kind of counts in his calculation, but it also applies to financing a car. Adding this cost would actually help his argument.

[mikeash]

Depreciation is the ongoing cost of the price of the car. Depreciation applies to financing a car too, but you don't count it in addition to your principal payments. Quick smell test: if you finance a $25,000 car, by the time you've reduced the value to zero it will have depreciated $25,000 and you will have made $25,000 in car payments on the principal, but you will not have paid $50,000 total, you will have paid $25,000 plus interest. Depreciation is ultimately just "take the price of the car, average it out over time" with potentially some games to take into account the nonlinearity of the curve.

I'm not sure if $5,000/year is right, though. The car cost at least $109,000 new and probably more. At that rate it would take 21 years to fully depreciate, but the average life of a car is more like a decade. It's nonlinear, but the car is at most six years old. I think this may be an underestimate.

While we're at it, I'm struck by the maintenance figure of $3,500/year. That's a big pile of cash! I had a (much more boring, plain old gasoline) car of similar age until last year and the yearly maintenance on it averaged well under $500 and I drove it twice as many miles per year. I don't know enough about the Roadster to say whether that's correct or why it would be that way, but it reinforces that this is a highly unusual calculation whose numbers will apply only to a tiny number of people.

Taking a quick whack at the calculation for myself, I currently do around 8,000 miles/year, insurance is about $1,000, maintenance is $28 (preventive maintenance is bundled in the price of the car for another year, and nothing should break, so I'm just on the hook for the state inspection), depreciation is, somewhat wild guess vaguely informed by my perusal of the used car market last year, maybe $2,000, parking is zero, and my cost per mile is about 8 cents. (Amusing side note: I did that last calculation with Google and it defaulted to giving me the result in dollars per meter.)

Putting it all together, and I have about $3,000 in fixed costs and $640 in per-mile costs, so $3,640/year total for ownership. At $2/mile, UberX would cost me $16,000/year for the same distance driven, about five times the cost.

[jonknee]

Which makes HN the perfect place to talk about it.

[cjbarber]

To be fair: It seems like Sam wrote this in response to queries after he posted this tweet

"just did the math. it's now cheaper for me to get rid of my car and uber everywhere (gas, parking, maintianence, insurance, depreciation) ."

And then he posted the calculation on Twitter.

[mikeash]

That would be fine if the article was just about his personal costs. But it's not. The last paragraph makes it clear that he thinks this calculation applies to enough people to make a difference of billions of dollars in Uber's valuation.

[rdl]

There's an argument that while Sam's living in the future, other people will be joining him.

[mikeash]

I really hope the future doesn't involve an order of magnitude increase in the cost of car ownership.

[plorkyeran]

Not to mention that if the cost of car ownership does go up an order of magntiude, the price of uberX would probably also go up quite a bit.

[mikeash]

That's the big problem I see with getting any significant number of people to use the service. Most people use cars to commute. An UberX driver is generally only going to be able to handle one person's commute, since most people commute at roughly the same time. Most of them won't find any work at other times, because demand is much lower. So for that to work out, most UberX drivers would effectively be one person's private chauffeur, and somehow this is supposed to be cheaper than owning a car yourself, even though you'd effectively be paying for your own car and a driver.

I'm sure it can scale up more than now, but how much farther?