Hacker News new | past | comments | ask | show | jobs | submit login
When I Took My Zipcar Into the Wilderness (theatlantic.com)
227 points by curtis 14 days ago | hide | past | web | favorite | 178 comments

I got banned from using Zipcar last year.

I lived in a condo building with an under ground parking lot in SF. A few times I rented a Zipcar and left it parked in the parking lot at my condo building.

About 50% of the time, the car would not unlock. In all cases, I successfully contacted Zipcar who informed me the only thing they could do is tow the car (usually 24+ hours later).

The parking lot had no cell service.

Other times I extended the car’s reservation on the iPhone app, but because there was no cell service, the internal Zipcar rfid didn’t get the updated reservation time; and the car locked me out after the original expiration time.

Zipcar probably realized they were losing money due to these issues, and banned me from using Zipcar forever.

Also: Never leave your Zipcar parked very long without cell service... the cellular unit will drain the car’s battery very quickly when it’s constantly searching for a signal...

Honestly, the ban sounds justified to me. If you know that parking in the garage will lead to a 50% probability of them having to tow the car, repeating this is just unnecessary. It would obviously be better if they could handle such scenarios but ignoring a known limitation to incur cost to others (since you have no advantage as you also can't use the car in this scenario) sounds wrong to me.

Disagreed about the ban. Since when is shitty technology (which any engineer worth their salt should’ve predicted - this is not rocket science) the fault of the customer? Unless they say in bold in their marketing statements “Only drive within cell coverage” the customer isn’t at fault for trying to use it as a normal car (in fact it’s crazy to think that parking in a place with no reception is no longer normal, at least according to ZipCar).

I do agree it is not his fault.

But the ban also prevents that particular customer from receiving bad service. Some things are not meant to be.

ZipCar should have formally requested that the customer not park cars in that lot, or what have you.

Since the customer needs to use that lot, they would probably stop using ZipCar at that point—but that's better than being banned forever if their life situation changes.

Never play the "who's dumber" game with private companies you'll always lose. If it happens twice you should switch on your common sense and stop using the service, or at the very least change your behaviour/don't be surprised if you get banned.

> which any engineer worth their salt should’ve predicted - this is not rocket science

There are probably tons of other issues you don't know about. Making it work offline would probably increase the attack surface, &c.

Any engineer worth their salt knows that they have 100 things to work on and this guy's corner case didn't make the cut. Easier for them to ban him and work on a new feature that generates revenue for thousands of users.

My point is that this isn’t some minor bug that you can leave in the backlog because it’s just a minor inconvenience. This is a major dealbreaker that should’ve been thought of when designing the service in the first place.

Do you see declining to serve a customer whom you could not serve well as different to banning a customer?

Because companies routinely do the former.


No I’m trying to justify using the product as it’s advertised. Otherwise it’s too easy to just falsely advertise something and ban the people that manage to figure out the catch.

As advertised?

I wonder if there's a reason why traditional rentals don't lock you out when your reservation ends.

You said you did it a few times. How often do you have to do something with negative consequences before you learn you shouldn't keep doing it?

Other than being told its a signal issue, how do you determine that? IoT crap rarely, if ever, actually communicates with the user about error conditions. You usually find out shit 'no workey' after the fact.

There's a lot of articles and twitter feeds. Go look up the "Internet of Shit". There's a reason why it's called that.

And Zipcars are no exception, evidently.

This isn't internet of shit though...is an iPhone shitty if you can't use cell services when you don't have cell service?

It’s obvious that a phone won’t work without signal. It’s not (yet - and hopefully will never be) obvious that a car rental service requires cell signal to function.

If this limitation can’t be worked around and they still decide to go ahead with operating the service then they should advertise this limitation prominently. Until then, the customer should be compensated (your rental car letting you stranded is definitely an issue and can cause non-negligible monetary damage to the user), not banned.

I can see maybe it not being obvious to a random member of the public. Maybe they thought their card somehow had all the data it needed. But I guess I just expected more from someone who posts on hacker news.

Is an iPhone shitty if you can't use it at all if you don't have cell service?

Is a car shitty if it leaves you stranded due to a known issue?

It would only leave you stranded if you lock the keys inside the car, and expect the zip card to open it back up for you.

"Locking keys in car" is a well known failure mode of cars, and it is entirely OPs fault.

The comment you're replying to is not the person who was banned by ZipCar, but someone else showing why the logic behind the corporativist comment is wrong.

> If you know that parking in the garage will lead to a 50% probability of..

50% is GP's after-the-fact description of how often the car failed to start. It's not an actual failure rate that GP knew about before they parked.

Anyone can make an error once. Commenter talks about how they repeatedly parked the car in an underground garage that they knew had no cell signal, despite multiple towings being required.

What's the appropriate number of towings for one customer due to that customer's intentional actions before you kick them off the service?

I'm the original poster. I misspoke in the original comment (too late to edit it).

Now that I think about it, I believe unlocking does work regardless of cell service (as long as you don't extend the reservation, and try to unlock it after the original reservation time ends).

I had the car towed twice:

1) First time, the car wouldn't start because the battery was dead while parked underground with no cell service.

2) Second time, the reservation was about to expire while parked without cell service, and I extended it on the app, but the car couldn't know that since the cellular unit couldn't phone home.

EDIT: For what it's worth, I don't blame Zipcar for banning an unprofitable customer. I acknowledge that i was definitely pushing my luck by parking there a 2nd time after the first incident.

“I ordered the tuna salad every time I ate at the restaurant. About half of the time, the manager informed me that they were out of tuna, and had to send a truck to fetch more.

After a while of this, they banned me from the restaurant for ordering the tuna salad.”

Actually, that analogy would only make sense if "you" refused to order anything but tuna salad and expected them to send the truck out without fail when they were out. And they had to comply with your request for some reason.

In that case, yeah, they should ban "you".

The fellow is parking the car in his own garage, and ZipCar claims to be in the car business, and it is normal to park cars in garages.

I say they have tuna salad on the menu, but they apparently can't always serve it to everyone.

Your suggestion that he order something else to eat is like telling a ZipCar customer to move to a different condominium or whatever.

They don't "have" to do anything, provided they aren't discriminating against him on the basis of a prohibited factor like race or gender. Just as a restaurant doesn't have to serve you (subject to the same caveat).

But the important thing here to note is that the hypothetical restaurant has organized their business in such a way that some customers cannot get what they claim to serve, just as ZipCar has organized their business such that some customers cannot enjoy the service they claim to offer.

They don't have to do anything, but I object to blaming the customer for their choice to run their business in such a way that a very ordinary thing that some customers can do, others cannot.

A better analogy would be that ZipCar has a known failure mode that will leave you stranded. And they don't feel the urge to fix it.

> Honestly, the ban sounds justified to me

based on the story above I agree, knowingly (and after a few occurrences you should understand there's a problem) damaging the service should lead to a ban or at least a warning.

No, it should lead to someone fixing the problem. The customer centric thinking in this thread is amazing.

One facet of the problem (the cell coverage) is largely out of Zipcars control, no? I suppose they could add some functionality to their app so that you're alerted when you park at a location without cell coverage?

Lack of cell coverage is out of their control, but that doesn't excuse them from designing around the problem.

It appears they're aware of the problem, but certain known issues prevent access (cellular unit draining car battery seems to be the common one and sounds like something that should be fixable).

Add a keypad like Ford has had on their vehicles for like 20 years. Have the computer keep a short list of one time pass codes (and maybe some sort of super long master code) that can be updated whenever the car is in cell service. So in the event that you get into the car, Zipcar can provide one of the one time use codes to get you in the door.

My imagined internal security review to this issue Hard "no", closing WONTFIX. This change would significantly increase the attack surface, risking a lot for very little upside. See "Epic 769 (Offline Support - roadmap Q3 2020)" for blocking & related issues

The common man would think a company would try the fix with the utmost priority a mission critical issue...

Don’t pretty much all phones have Bluetooth now? Couldn’t it backup the authorisation and allow him to connect via Bluetooth and verify and unlock? How is this absolutely his fault?

There's at least two ways of handling serious product failings: "Oh, shit, we have to fix that!" or getting rid of the customer exposing the problem and trying to conceal the issue. One is the action of a healthy company, especially one that is trying to grow, while the other maximizes current profit levels.

Doesn't parking in same place twice, knowing full well the likely cost to the person they're doing business with, seem unethical? Would you want to do business with someone who carelessly took advantage of you?

I'm not sure I would do that, but I would expect them to fix their service it if it is costing them noticeably.

AFAIK they did not say you shouldn't park underground nor did they explain their various failure modes and how to avoid them.

Interesting. So it caches the authorized zipcards offline for this case, right? Why did it fail to unlock 50% of the time? If it failed, could you just retry, or did it fail 100% of the time in specific bad locations?

That's why I always took the key with me, despite the fact that they tell you not to.

Interesting. If that really works, perhaps they should add a special asterisk that you should always leave the key in the car unless you're parking in an area with poor or nonexistent cell reception.

Zipcar should get banned from such stupid practices and almost liquidating fine for customer deception.

The bit about the cell phone discharge seems pretty unlikely. A charged car battery has at least 2MJ of stored energy. GSM in its highest-power mobile station state draws 800mW. Add maybe another 150mW for the GPS, call it 1W total. The battery should power such a phone for weeks.

You should be able to smell-test this fact by considering that your mobile weighs nothing and runs all day whereas a car battery weighs 20kg.

Car batteries are generally a starter battery (minus EV ones). They don't like being closed circuit for extended periods of time.

A low current is usually allowed (200mW is the top end in my experience), but anything over will, over time, sulfinate the plates in the battery and cause it too loose charge and capacity rapidly.

Generally if you want to run a low current device on a lead acid battery, you should use a deep cycle battery with plates designed for it. Starter battery have large, flat plates to carry a lot of current for only a few seconds tops, then be charged up again.

It looks like Zipcar should have a separate smartphone-like battery to power GSM and GPS.

How long does it take to damage the car battery?

How long does it take to discharge the car battery?

Depends on the load.

Your smell test is a bit misleading in that most car batteries have a drastically lower energy density than Lithium Ion cells. There’s still a lot more energy in the car battery, of course, but not by nearly as much as a naive comparison of the weights would suggest. If it had the energy density of your cell phone, the lead acid battery in your car would only weigh 5 or so kgs.

Your cellphone will also last a lot less than “all day” if you take it somewhere where it has to run the radio at full power to try to maintain a connection. I’ve seen otherwise reliable and long lasting phones easily end up burning 1% a minute in areas of really bad signal.

I also seem to recall that standard car batteries respond poorly and inefficiently to trickle-discharge, being designed mainly for providing a lot of amps over a short period. Providing long term off-grid power is what deep cycle marine batteries are for, and they are a separate product.

As far as I know standard car batteries are fine with this kind of trickle discharge. Where they fall short compared to deep cycle batteries is providing a more substantial current (think tens of amps) over a prolonged period of time.

Trickle discharge up to a few dozen milliwatts is generally fine for car batteries. They don't like being closed circuit in general but if you draw more than about half a watt for a longer time period, the battery plates will start to desintegrate.

OK, here's another test. A common way to end up with a dead battery is to leave the lights on overnight. The lights on a car draw over 100W (standard halogen headlights and incandescent tail and marker lights). But you can leave your lights on for hours and still start the car. Therefore it stands to reason that the same battery could power a GSM radio for hundreds of hours.

That assumes the battery behaves the same for low and high current (they don't).

Even lithium cells already don't; high currents causes them to warm up, reducing capacity. So a device that draws 1 Watt will last longer than one that draws 10 Watts from a small lithium cell.

Lead Acid responds even worse, especially starter batteries, less so on deep cycle, and it's capacity rapidly changes, without temperature influence. A 1 Watt device might only last a day while it's fine starting up an engine with 10 times the total power consumption

This assumes the battery is fully charged, if it's cold or if the battery is slightly defective, it could be a lot less than full charge.

Searching for signal uses some kind of exponential backoff algorithm (somekind because of patent wars between Qualcomm Broadcom and others), the phone doesn't simply sends the maximum power every fraction of a second. 1% per minute means something is wrong with your phone, or a weird edge case.

GSM phones don’t send anything at all in the first phase of registration.

Anecdotal, but if I leave a USB charger plugged into the cigarette socket, without even anything plugged into the usb, it drains my battery overnight.

Get a refund.

I leave a phone plugged into my car's USB port 24x7 and the car has its own GSM modem too and obviously it starts right up every time.

The difference is that you're using a USB port directly built into the car (which is "smart" about not draining the battery) while the other user is using a cigarette lighter to USB adapter (the cigarette lighter is a "dumb" circuit).

getting your battery tested might be useful too.

If it's not a car from the 80s there may be a fuse you can add/remove which converts your cigarette lighter from always on to only when engine is on. Check the back of your car manual.

Friend in Seattle had this experience when trying to use a zipcar truck out of a garage while moving (i.e. Limited time window to get stuff out of one area to another). Ended up calling support, dealing with multiple hours of delays. Despite the issues which has led him to avoid Zipcar these days, he uses ReachNow (now defunct) and car2go (now called ShareNow) regularly and AFAIK ReachNow didn't have these limitations when on long out-of-range trips - they solved the problem by doing sortof a two-factor of having an RFID card (to unlock, on window) and a pin code (entered via the BMW iDrive system to start engine).

I could argue Zipcar has a deep integration problem needed to solve the out-of-comms scenario. They try to support several car vendors in order to have different car types (toyotas, mercedes, ford vans) which limits them to having this cellular-dependent setup. Whereas Reachnow (BMW), car2go / Sharenow (mercedes / smart), and limepod (fiat, for now) have limit the car type for better integration.

In a related front, a friend moved into an apartment building with some bluetooth-based laundry vending machines. The machine requires blootooth, and the payment is done via phone which requires a cellular connection. The laundry room (like any basement area) has a spotty cell connecition, so the friend does some weird positioning exercise to find the area where he can have a cellular connection while still being connected over bluetooth to the laundry machine. Strange future.

Zipcar has RFID-based card access now. Sounds like they just need to cache an auth token locally when there is cell service saying how long the car may be used for. Although there's still an edge case that parking out of cell service range when the token expires will disable the car even if you extend the contract since there's no way to refresh the token.

Perhaps they could create an RFID token in their app and let the RFID-capable phones bear that to the car since it's much easier to move a phone than a car, not to mention potentially get wifi service where there is no cell service.

I haven't used any of the services you mention, but an on-line rental service I have used has a simple system: they unlock the car remotely when you start your rental, and the key is in the glove compartment. You may run into issues starting your rental, but lack of connectivity won't stand you in the middle of nowhere.

Is there a particular reason these services can't just leave the keys in the car? Theft?

ZipCar has keys in the car and it's part of the service rules that the key CANNOT LEAVE THE CAR. I have no idea if they use rfid or something to check, but I felt it was really clear in their "how to use" videos that the keys never leave the car, you can can get banned from ZipCar.

So while there are keys in the car that you could have held onto to use during a cell service outage, you're told to never remove them from the vehicle.

That's pretty dumb. Car2go (Vancouver) has keys in the glovebox in a special holder-device: you're free to take the keys with you if you want to keep paying for the car as it's parked, but it won't let you "end session" or whatever unless the keys are replaced into the holder thingy.

This also avoids some of the problems mentioned in the above article. I assume they have a remote-kill but unless there's a serious problem, you always have the keys so you can always start the car.

If the car has no signal, how is it unlocked?

That's a problem when starting the rental and returning the car. Pickup points should be placed in areas with good reception. But in any case, not being able to start a rental is a nuisance. Not being able to use the car once you've driven it somewhere is potentially dangerous. I don't see why the car should need a connection to unlock once the rental has been started.

In the Netherlands, we have Greenwheels, which in addition to being accessible using an app, also allows you to open the car with an NFC-enabled card (I'm assuming for the case where my phone doesn't have reception?). If the car doesn't have reception either then I'm not sure you would be able to open it.

EDIT: Actually, after you pick up the car and start driving it, you (un)lock the car using the keys that are in the glovebox. In principle, you should be able to drive into dead zones without an issue

They could give your phone a signed token indicating your rental times, and then let your phone directly connect to the car to hand it off.

Presumably it is easier for a human with a phone to move to a location with service than moving a locked car, usually just out of the cave you're in.

I just don't see why that complexity is necessary, when you could just give the user the key to the car.

The whole point of zip car is that you aren't reserving it a week in advance and want to wait for a key to be mailed to you.

Why can't they just leave the key in the glove compartment? That way, Zipcar only needs to unlock the car remotely when the user starts the rental, and lock the car remotely when the rental ends. In between, the user can use the key. like with every other car.

Probably to prevent accidentally taking the key. If you've ever used a zipcar, the key is affixed to the steering column. I assume they have an ignition interrupter that requires a proper unlock signal because otherwise stealing one would be matter of jimmying a lock or breaking a window.

Good question, but if there is no signal how could the previous driver have locked the car? Similar services I'm using don't let me end the rental if there is no signal, simply because either the car or the shop on my phone can't talk with the server.

This sounds like a pretty common scenario for zipcar and certainly cell phone coverage coming in and out is something that they should be planning and designing for! Maybe a series of one-use pass codes or something like that could help. I always make sure there is a failsafe when I have two microcontrollers talking on the same circuit board! Never fully trust anything external.

There are many things they could do with software only without having to change their current scheme.

- Use exponential backoff when looking for network (although I find this part of the story hard to believe, my phone survives much longer when there is no network)

- Have a larger grace period. Let the key card just work for an extra day or indefinitely if the car's last known location is not the garage. Charge the customer for the usage.

- Fail open especially if the battery is low. Once the car starts, you have all the electricity you need to do updates.

- Make the updates small and efficient, so long as a few signed UDP packets get through, you can update the schedule. You can even use SMS. Sync non critical things (e.g. logs) opportunistically and asynchronously.

>my phone survives much longer when there is no network

Interestingly I've had the opposite experience (with a Pixel 2 XL, and before that a Nexus 6P). With both phones, when going on out-of-service hikes the battery drained significantly faster than usual -- despite not using the phone at all, I'd see a 40-50% power drain over ~8 hrs. Putting the phone in airplane mode when I knew I'd be out of service solved the problem, as expected.

My Pixel 3, on Fi, also seems to heat up quite a bit whenever I am at a bad network location!

In the good old days, you could touch your TV's rabbit ear antennas and do some weird positioning exercises to improve your service.

i had a similar experience with Turo. The car would lock and unlock with a bluetooth connection and would not start unless the bluetooth connection was made. I went on a fifteen minute hike about an hour off the main road to see a redwood grove and when I got back the app had logged itself out with no action taken on my part. We had to hitchhike for 4 hours back to a hotel so that I could log in using my Google account, then hitch hike back to the car. The car itself was unlocked the entire time because we could not lock it. I wonder if anyone else has had an experience with Turo being logged out. Could a certificate or token have expired?

I hope you’ve reported that to the app developer. That’s a potentially fatal problem.

Bluetooth? Holy shit, it can barely keep an audio connection, why would anyone use it for something important?

> then hitch hike back to the car.

Sounds like it would have been easier to just rent another, closer car at that point, assuming you took your stuff with you since you couldn't lock it.

I've rented Zipcars here in the UK many times and never encountered this.

The cars have a key in the glove compartment. After you initially unlock it at the beginning of your reservation you take that key with you. You put it back in at the end of the reservation. In between you use it to unlock like a regular car key/fob.

Am I missing something here?



> The key is in the glovebox in ALL vehicles. Throughout your reservation you should use the key to lock and unlock the doors. For a vehicle that has a start button instead of a key, the Zipcard or app can be used throughout the reservation to lock and unlock the doors.

OK so this situation is plausible when the car doesn't have an actual key.

Technically, according to their rules, you are supposed to leave the key in the car. I didn't think anyone actually paid attention to that rule.

You're supposed to put it back _at the end of the reservation_. Not every time you lock the car.

Given human nature, I'd assume you would be missing the key frequently.

Many of the reported issues seem to involve renewing an expired reservation. Also, how's your cell coverage there?

You will be charged quite a bit if the key is missing at the end of your reservation. They do have your ID, driving license, and credit card on file after all.

If this article is true, Zipcar in the US must work in a fundamentally different way from how it does here in Europe. Every Zipcar I've ever rented, once you've gained access with your card or the app the first time, you have the actual keys in your hand until the end of the rental. You don't need a mobile signal to lock and unlock the car wherever you go, the key fob does that.

Of course, in one of the scenarios the article addresses - car battery failure in the middle of a wilderness - it doesn't matter whether it's a Zipcar or not, that's a serious problem for any car.

In the US the keys that are in the car are physically tied to the car with a string. You are expected to use the card to lock and unlock during your reservation

Wow, if I came across that without some prominent warnings, I'd probably assume some kind of key setup mistake and cut the string.

I use a zipcar occasionally and this is exactly what happens when I'm driving with a friend. Once they actually did manage to yank on and snap the string. They said something like "Wow, the previous driver tied the keys to the center console, what an asshole" pull, snap

It has a warning. Something like “leave these keys here! It seems weird but don’t worry about it.”

I can imagine that a law could be made to mandate car manufacturers or dealers or rental companies to list all the possible failure scenarios that cars might not work properly, and present the drivers some warning labels or sort. It shouldn't take that much effort to list all the possible cases by hierarchically inspect smaller modules (just like checked exceptions in Java), and it doesn't have to be very precise. Like most drugs warn about possible death due to its use, there's no reason that we couldn't expect cars and other safety related industry to follow a similar standard. Just like drugs, scaring potential customers away is a part of its cost (and benefit).

The car manufacturers probably already do a FMECA (https://en.wikipedia.org/wiki/Failure_mode,_effects,_and_cri...) analysis.

In my younger years I worked at a startup that had an advisor that was one of the original technical architects at Air Touch Cellular. For those that do not know they were one of the original cell phone companies in California and the west (now AT&T). He did not own a cell phone, a fact which I found strange. I asked him why and his answer was “they screwed it all up..the deployments and coverage, it’s a solved issued but they screwed it up anyway. I will not own a cell until it works everywhere.”

I'd love to hear more about this.

Did he mean that there is a solved and known way to get uniform national cell phone coverage without blackspots? If so, do you know more about what his solution was.

No magic here. The gist of what he said was that they did a design that had 100% coverage but the company pulled back from doing that.

Sounds like he was talking about something specifically to their network.

There's not some magic thing that everyone should be doing that will give 100% cell coverage.

It's a two way radio. The transmitter needs to talk to the phone. But the phone also needs to talk back.

Also carriers have different frequency spectrums allocated to them.

A carrier using 900MHz has to have fewer base stations than one operating using 1800MHz because lower frequencies transmit further.

They also are able to go underground and through walls better.

It's a solved issue, in that the cell company "simply" has to spend enough on infrastructure to cover the desired area.

Justifying the business case to cover the low value (i.e. few customers) areas is another story. However, even here, it might be possible to justify it if the rest of the customer base sufficiently values wide coverage to pay a premium. Particularly if the competitors don't have the same level of coverage.

> Still, cars without reception become vulnerable in a few scenarios: ... when the vehicle battery dies. That last scenario was the one my family and I found ourselves in

This seems like a major detail that should have been mentioned near the headline. Even if they got into the car, they would still have been stuck because the car battery was dead.

Every new technology comes with its own unique foibles and failure scenarios. One can easily imagine the very first automobile drivers getting frustrated by how their car becomes a useless lump of metal within minutes as soon as they run out of gas. "If I just had my trusty horse, this would never happen!" Fear mongering around these new failure modes makes for a fun pastime, but fundamentally pointless, unless we also consider the benefits they bring to the table.

Jumper cables and spare gas cans have been a thing for quite some time; and Li-Ion battery boosters which can start a car multiple times are small enough to carry in a pocket. Everyone driving in the wilderness should have these things in the trunk in addition to a spare tire and some water. The majority of problems which strand people are flat tires / dead batteries / overheating / no gas; which are all trivial to fix on one's own with preparedness appropriate to the environment.

Furthermore, mechanics are readily available in rural areas; and when horse transportation was common, farriers and vets were also common. These people drive out to your location and fix stuff, although you might need to walk to them without a phone. And any rural mechanic can hack a fix to the mechanical components of an internal combustion engine well enough to get it home.

So, if you accept Zipcar's terms of service ("don't remove our interlock / hack the ECU"), this is fundamentally different because it's a failure mode which only Zipcar can fix.

It's inevitable that someone will die because of this design at some point, and the various articles and online discussions about these problems and the company's arrogance and incompetence at addressing them will mean they will be found criminally negligent.

> making sure that the cars work when out of reception is a “mission-critical success factor” for the company

I guess mission failed then...

It is so easy to call something mission-critical. I am always tempted to ignore it when I hear it, because most of the time it is a red herring. Instead I use it as a launching off point for a conversation with clients about priorities for a given solution.

If the car has a big yellow sticker on the dash saying, "if you drive this vehicle to a location without cellular coverage, you may become stranded", would you rent it?

Sure would. Not for every trip, though.

A lot of comments are focussing on a Zipcar not starting due to insufficient battery power. However, that’s only one aspect of the article (a personal anecdote of the writer). The other scenarios highlighted the key take-away from the piece:

> But the problem with using services dependent on a network is that you are then dependent on the network.

This – for me – is the main problem with IoT devices, e.g., many privacy issues would go away if devices were designed to operate without needing to regularly “phone home” (excepting explicit, intentional data-mining by the manufacturer).

> many privacy issues would go away if devices were designed to operate without needing to regularly “phone home”

This is the sole reason why I do not buy IoT devices. I would love to have a resource which lists devices that do not try to contact the manufacturer over the internet

Both car2go and Evo here in Vancouver appear to have the right idea - when you rent a car, cell access is only needed for starting and ending the rental (which must be done within city limits). For a stopover (stopping the car without ending the rental), you just take the physical key with you.

They’re also quite convenient since you can pick up and drop off the cars at almost any street or parking garage. Because Vancouver doesn’t (yet?) have Uber or Lyft, they serve a really important role here.

well um, the car's battery was dead. so how exactly would it open via RFID? of course if you put that piece of information first, the headline isn't so catchy. i digress.

> A tow truck took us to a lot with reception, where the rental failed to start... Still, cars without reception become vulnerable in a few scenarios... when the vehicle battery dies.

> I shudder to think about limping back to a trailhead with no more water in my backpack, only to find a car that would not start. Or getting locked out and marooned in Death Valley, perhaps with medicine trapped in the car.

Yep, "car died at the trailhead" is hardly a novel or Zipcar-specific problem. I've known plenty of people who camped for few days, packed up at dusk, and found themselves with a dead battery and no one around to jump it. Taking a Zipcar with essential medicine into Death Valley is a scary thought, but it's recklessness in a dangerous place rather than some esoteric flaw of Zipcar.

It's a shame, because there are real issues here. Lacking a physical carkey is a real drawback that means you can't wait in your dead car or open the hood to jump it. Having your car die in the woods because you exceeded your reservation time is bad for the user, the next user, and the company. But the article spins those narrow issues as life-or-death flaws, conflates "dead car" with "Zipcar broke", and puts "Zipcars normally work without service" down in paragraph seven after implying in paragraph two that this is a consistent problem.

The headline was "When I Took My Zipcar Into the Wilderness." It's a pretty underwhelming headline and I have no idea why anybody would possibly object to it.

How is that clickbait? Even the subheadline is a pretty fair description of events: "In an area without cellphone reception, I was unable to open the car."

yes if you want to get pedantic the subheadline could more aptly read, "i was unable to open or start a car with a dead battery". or for the main headline, "when i took my non-zipcar into the wilderness (and its battery died and i was stranded)", would anybody read it? no.

ZipCar seems to be unique (at least among cars I've used) in that you can't open the doors with a dead battery and it seems there's no mechanical backup. As a result the car needed to be towed. Additionally, you seem to be ignoring that the battery may have died because it couldn't find a cellular signal.

And yes, I would have read it if the subheadline was instead something about modern cars that weren't zipcars that causes this to happen, because it's never been possible with a car I've owned and is pretty interesting.

> the battery may have died because it couldn't find a cellular signal.

I had totally overlooked this myself, thanks. If Zipcars without service are constantly phoning home and running down the battery, that could make the issue way more common, and explain why regularly-serviced cars still die on people pretty often. It would also escalate ZipCar's role from locking the car when the battery dies to actually killing the battery.

Beyond that, I do wish the social tie-in hadn't just been a line about networks failing us, but a more substantial look at the limitations of relying on power and data access. It's not just a sharing-economy issue; wholly-owned IoT devices, keyless-start cars, iPhones earbuds, etc. increasingly lack analog fallbacks. They're not just neglected when designing cheap IoT goods, they're being stripped out of existing products like phones and keyfobs for sleeker design.

It's a topic that's been on my mind a lot; whenever I travel abroad or to remote bits of the US, I notice that an increasing fraction of both services and products aren't viable. Beyond my inconvenience, it makes me wonder what will happen in 10 or 20 years. It's not uncommon to go somewhere remote and see everyone using the previous generation of products, either bought used or sold off cheap as they aged. But that's going to be increasingly impractical as those things demand nonstop data access.

A key element here is that the battery likely died _because_ there was no service. When the car's cellular unit is scanning for service, it uses significantly more power than usual—quickly draining the battery.

when you leave your vehicle at a trail head for a few days you can disconnect the battery leads, and greatly increase your chance of coming back to a live battery.

I too rolled my eyes when I read that sentence..

I recently went into a parking garage in Amsterdam, where they have a partnership with a parking app; you press a button in the app to open the garage doors.

The reception underground is nil, so it took me quite a while fiddling with phone position, with my arm through the entrance gate, to be able to get out...

Zipcar also has a dumb system in-place where you need to use the card first to start the reservation following which you're allowed to use the app to lock/unlock the car. The whole point of using the app is so that I don't have to carry a physical card with me.

Based on my experience, this is not true, I could always open the car via the application.

The first time you unlock the car, it always has to be using a physical card. I learnt this the hard way.


This has actually been relaxed of late, though I agree it used to be a requirement earlier in Zipcar history.

Ok, so you say I just imagined it.

All car sharing platforms* I have used in Amsterdam/The Netherlands/Europe have a key in the glove compartment that you can take out when you park the car. Why doesn't Zipcar do the same?

* i.e. Car2Go, ConnectCar, MyWheels, Fetch, SnappCar

Doesn't the key go missing often?

I'm reminded of European hotels that want you to leave the room key at the front desk.

They plug in to a socket in the glove compartment, and I don't think the rental is counted as completed until they are returned there. I'd naively assumed this was standard across the globe, but clearly there must be other factors to warrant not allowing you to remove the key.

On the other hand, the fuel payment card does missing - but at least Zipcar in the UK are efficient at refunding your own payment when provided with a receipt.

In basis of this case, vehicle battery was actual problem, as well as the Tesla outage there was a backup.

That people choose not to have it available isn’t completely the manufacturer or providers fault is it?

I've used LimePod in Seattle, you can't lock your car without cell service, which guarantees you'll only be able to unlock your car if you had service prior. The key is permanently attached to the ignition, so you could leave your car unlocked and start it later if you were in the wilderness. I wouldn't recommend taking any of these free-floating cars outside of the city because of problems like this, though.

Whenever I used Zipcar, I only used the car at pickup, and then always take the key out for any stops during my rental. Would that have prevented this?

Allowing people into the wilderness with a car that by design has these issues is very close to negligent homicide.

What exactly do you propose? A zipcar that doesn't allow allow you to drive outside of the city?

Sure. Shut down the engine when it loses cell signal, with some warning so you could turn around. It'd be inconvenient, but not more than the alternative.

Shut down the engine when someone is still driving? That's not just inconvenient, that's suicidal. What if you're not able to get back to a place with cell coverage in time? Is the car just going to stop in the middle of the highway or the middle of a tunnel?

The title is misleading

> Still, cars without reception become vulnerable in a few scenarios: when members lose or do not have their physical Zipcard with them, when they exceed their reservation time or want to extend their Zipcar reservation, or when the vehicle battery dies. That last scenario was the one my family and I found ourselves in, though we did not know it at the time.

Can someone change the title to reflect that cell phone reception had nothing to do with the inability to open the car in this case?

How then do you explain this part: "A similar thing happened to Tom Coates, another Californian and, as luck would have it, an expert in the internet of things. He used a Zipcar to get to the Getty Villa, part of the J. Paul Getty Museum in Los Angeles. His mistake was to park in a garage, where his bars went to zero and he found himself stranded as the sun went down and the museum closed and the on-staff security guards started to circle in annoyance." ??

> How then do you explain this part

All we know from that anecdote is that the car left him stranded in the garage and that he had no cell phone reception. Nothing of what is presented suggest that there is a directly causal relationship between those two circumstances. In fact, to the extent that it's described it's consistent with the situation referred to in the title of this post.

To make an uninformed guess, maybe the car battery drains quicker when cell reception is poor, but that's conjecture based on little information.

Someone mentioned above that the battery most likely died quickly due to constant hunting for cell signal, so it’s 100% related.

Even my cell phone can handle no reception for a few hours—no problem. It seems strange to me then to assume that it is 100% related based on what "someone" said was most likely.

I believe the issue was that, after losing both reception and battery, the car needed reception before being unlocked.

Looks like it. They got a tow to somewhere with reception but the car wouldn't start. That implies they did get into it, but were still stuck due to the dead battery. So presumably the door unlock system has it's own backup battery?

No, the power required to unlock the door (a small motor) is usually a lot less than the power required to ignite the fuel. You'll probably find that the radio and internal lights still worked too (though likely not for long).

To turn the engine. Lighting the fuel is easy enough. But overcoming the pressure inside the engine and getting ~50lbs of metal reciprocating and spinning from 0 to 1500rpm in a few seconds takes a bit of a kick.

TL;DR: the car battery was dead

That's actually a problem with all modern cars: If the battery dies, the key fob won't open the door. In fact, if the battery dies in the car or the fob, you can't open the door.

But, there is a mechanical backup with an ordinary car. In my 2012 Volvo, I can open the driver's door with the mechanical key. If the fob is ded, I can physically insert it and start the car. If the car is also dead, I have to sort that out to drive it, obviously.

I like to MTB and I used to climb, so I carry a backup battery that can start the engine if need be. But I appreciate that with a ZipCar, the whole point is that you don't need to go somewhere and pick up a mechanical key. So it's really the same risk profile as if I were to drive around with the key fob, but leave the mechanical key at home.

The other worrisome thing is if a car is designed that it needs the Internet to start. If you drive it somewhere remote, you have a problem. Or if there is a service outage, you have a problem.

If I buy another car, it will probably brag thatI can use my phone as my key. I will still want to carry the mechanical key, but my phone will then become a single point of failure. With my fob, and my 2012 car, I can start the car even if the fob battery dies. But if my phone dies, I will be SOL.

Luckily, said backup battery that I carry in my car has a USB port to recharge my phone. But even then, what if the network is down? Am I to trust Bluetooth to do its thing?

These failure modes require some thinking through before going anywhere remote. I'd want to work all this out before taking another driving holiday with my family.

I'm amazed at how easy it is to kill the battery on modern cars. I would think that they would monitor the batter and shut-off all non-ignition systems when it gets low. Or at least turn off the headlights when the car is off.

FWIW audi does do this rather effectively in my 2011 audi A4 and it also comes with a huge battery compared to other cars I've owned. You can tell if your battery is near EOL by the fact that the entertainment system starts taking 30s to come on after starting the car, but it also shuts off most other electronics as the battery dips lower as well.

When a battery was on the edge of failure, I was also amazed at how many days it continued to be able to start for, driving a mere 3 miles per say and starting it twice. The voltage gauge on an accessory I had would show the battery under 10 volts, I'd push the ignition button and silence would follow, then seconds later the faintest sound of an engine would start. Every day, for nearly an entire month of winter.

If the battery is flat on the car then you are still pretty much out of luck even if you can get in. The only advantage to being able to open it is to use it as shelter in a emergency situation

That depends on the vehicle. If it's an ICE with a manual transmission, and you can get it moving fast enough, you can pop the clutch to start it. Back in the day, with a ~dead battery, I'd park on a hill. Even without a hill, 2-3 people can generally just push.

> That depends on the vehicle. If it's an ICE with a manual transmission, and you can get it moving fast enough, you can pop the clutch to start it.

This depends on just how dead it is on newer cars.

If it's able to run the electronics but just unable to crank enough to start, this can work. If it's dead enough that you can't unlock the doors it's probably also dead enough that the fuel pump, ECU, etc. aren't going to be able to do their thing.

Carbureted vehicles can be bump started from flat dead for the most part, but EFI and especially direct injected vehicles need some electricity to operate.

Good point. I should have noted that I've only done that with a 60s VW Bug. Just to see, I have turned off and restarted my ~10 year old Civic, going down a hill. But not with the battery disconnected. Maybe I'll try it. Not now, though.

Edit: And upon reflection, maybe not. Because I'd have to pay someone to reset the computer.

Edit: But wait a minute. If the vehicle is moving downhill, and you engage the clutch in first gear, the engine will be turning. And even if it's not firing, the alternator will also be turning. So you should have enough power for the computer, ignition and fuel pump. Or am I missing something?

It wouldn’t surprise me if newer cars require a higher battery charge to start, which means you would need to roll the car for longer, or (worse) don’t effectively charge the battery at all at lower speed because the partially booted system takes too much power, so that you need a higher speed to start charging the battery enough to power the ignition system.

Also, are there cars where you need the electronics to disengage the brakes?

Don’t do it with battery disconnected, it can cause the voltage to go too high and you may fry things. Back in the day only bulbs, nowadays who knows.

Thanks. I'm not going to try it.

The alternator needs to get some decent RPM to generate enough power. You'd have to go downhill quite fast.

In first gear, so not that fast.

Careful! Shutting off the ignition will lock the steering wheel on older cars.

That’s called powered steering.

It depends on the vehicle. Power steering can either be electric (EPS) or mechanical, with a compressor that's belt-driven off the engine. My Civic has rack-and-pinion steering with EPS. When the engine is off, it's still steerable, but there's no power assist. But you don't want to turn the ignition lock all the way off, because that does lock the steering wheel.

I don't remember ever having any problem starting my 92 Civic that way. It definitely had EFI and an ECU. I know I started it when the alternator was bad, as well as when the battery was dead. (not at the same time)

My Civic is mid 00s, but I'm pretty optimistic that I could manage it. Hopefully won't need to, though.

Interestingly enough, early automatic transmissions also allowed this, since they had a rear pump that could build up enough hydraulic pressure to power the rest of the transmission without the engine running first:


This feature disappeared in the early 60s for cost reasons.

And what percentage are manual transmission vehicles these days? Probably less than 1

90% of southern Europe?

Why just southern Europe? Here in West Europe there is still a vast majority of manual transmission.

Some models it’s much higher, but overall definitely single digits in the U.S.

Europe has a much higher take rate on manuals.

>>90% in Germany

True. But some of us still like them.

I'm not hard core enough to have a pre-computer vehicle, though.

> If the battery is flat on the car then you are still pretty much out of luck even if you can get in.

With an EV yes, but not with an ICE. Just keep a jump box in your trunk.

These "jump boxes" are amazing! I used to keep big jumper cables and have AAA in case of a dead battery. Last time AAA came (after I waited hours...), they just pulled out a little "jump box" and instantly started my car. I found and bought a random one on Amazon (there are dozens -- https://www.amazon.com/s?k=car+jumpstart). It's very small, can also be used to charge a cell phone, as a flashlight, etc., and really works. Every ICE should have one of these in the trunk.

Note that the battery in a "jump box" is probably capable of discharging a dangerous amounts of energy at very high current. BigClive measured[1] over >500A when he tried to discharge one of these batteries, which didn't actually work. The internal Li plates probably separated, breaking the circuit before fully discharging. After starting to disassemble the battery and realizing the that the plates had separated he attached a voltmeter and measured a full 3.6V across the supposedly-discharged, partially disassembled battery (squeezing the battery with his hands, "This still is fully charged, but it's just puffed up... that's a bit scary." Later, after the Li plates start changing color (!) - probably metallic Li starting to rapidly oxidize - they spontaneously ignited (pyrophoric).

The benefits of having a spare "jump box" in the trunk are obvious, but it's important to remember that this means carrying a dangerous amount of energy in form that is a very serious fire risk if anything damaged the battery.

[1] https://www.youtube.com/watch?v=0tGK1nqXr28

And having shelter from the rain while you wait for someone to arrive can be a great advantage.

I'm sure there are many other fantastic reasons to want access to your vehicle when the battery is flat.

Yep, that's potentially a huge difference.

If you can't open the car, you probably can't pop the hood latch to get a jump-start, which makes the difference between a tow and a simple drive to a store. And "stuck alone at the trailhead with a dead battery" is nothing new to hikers, but being able to spend the night in the car if you have to makes it a lot more pleasant.

Last year, we push-started my dads car where battery was so bad it could not be charged or jump started from another car. Push start worked the first time (car from 2009).

With my (relatively) old Ford, the backup key is a small one hidden inside the fob that you can pry out if needed. There are also backup methods to lock the non-driver doors and for the push-button start to work in the case of a dead fob. I'd be very wary of buying a car without those backups.

Can you tell me exactly where you saw that? I don't see it.

“or when the vehicle battery dies. That last scenario was the one my family and I found ourselves in, though we did not know it at the time.”

Applications are open for YC Winter 2020

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