The Self-Cancelling Subscription | Hacker News Reader

The Self-Cancelling Subscription

predr.ag

188 pointssurprisetalk75 comments

>"Working" is not the natural state in a complex world! It's a testament to the combined energy and skill of many people that systems are built and kept working well enough for long enough so as to become invisible.

I just have to take issue with this as someone who grew up in a very rural, natural area and was enamored with biology, biological, and ecological systems as a kid (8-12).

The statement that "working" is not the natural state in a complex world? You're showing your ignorance of complex systems.

What of the Ogallala Aquifer, a massive underground cave and sediment system that stores and filters water over hundreds or thousands of years? It's massively complex and in its natural state it's working but we're draining it.

What about the weather systems in the atmosphere? Could you argue that one of the most complex systems (maybe only second to the ocean current system) on the planet is not "working" in it's natural state? Don't take an anthropomorphic perspective of it working for you. It is a complex system whose natural state is "working". If it breaks down for our purposes at this point, it is due to our combined energy pulling it from it's natural state.

Your limbic system is very very complex and is naturally in a state of working. No human intervention.

It's a testament to our combined lack of regard for the true complexity of systems that we consistently build systems that fail in opaque ways, and through our actions destroy long-running complex natural systems that we don't fully understand.

He speaks as if becoming invisible is a matter of transparency, but it functions more like a veil.

Hi! Author here. You're right — but two things to consider here.

First, that quote is referring to human-made systems, not natural ones (as is the rest of the essay!) and I think our views align on whether human systems regularly work.

Second, natural systems (and all complex-enough systems) are always running in some degraded fashion. So what "working" means is ambiguous: they are broken, yet accomplishing the goal. The quote from the essay refers to "working" in the "free of faults" sense, in which I again think our views align.

"they are broken, yet accomplishing the goal."

Are they really either of those two things? Natural systems have no "goal", they just are. If they change, they change. If they stay the same, they stay the same. Because there is no goal, there is no "broken". It is only we who assign some sort of meaning to them and characterize them as "working", either because they meet our needs, or just because we are inherently impressed by complex systems.

If you follow through on there being no goal, there is no working or not-working state, in which case it bears no relevancy on the question of whether "Working" is not the natural state in a complex world!

Or rather, you could say TFA is made more correct, by virtue of “working” not being a natural state in the first place.

But if we allow room to anthropomorphize, we can basically state that the natural goal of a natural system is to keep doing what it do, at least in regards to the larger outcomes. And for some strange reason, these systems are shockingly difficult to influence at meaningful scale in ways that are rarely true for the systems we design. In one sense, they continue to operate despite continuous minor and possibly major (but not catastrophically so, by definition) perturbations to their state

You need to burn ridiculous quantities of dino juice to influence the weather system. You need to look at windows a little funny to bring it to a complete halt. You need to bully only few substations to bring down the electrical grid.

I’ve read recently about natural systems in the book Antifragile. It’s interesting how those systems can become better.

holy unrelated discussion, batman! it sure yaps

> Your limbic system is very very complex and is naturally in a state of working. No human intervention.

It has evolved over millions of years. The evolution included billions of variants that didn't work and died before being able to reproduce.

And even then, are you saying everyone's brain is perfect and never needs any external intervention?

On top of that, we have to continually tend to our bodies, feed them fuel (sometimes at risk of life and limb), exercise them, clean them, tend to them, visit the doctor, take medicine/drugs, etc., just to keep them in good shape. And they eventually have a 100% failure rate.

"Working" here means "operating according to specification or intent", not just "operating". Natural systems have no specification or intent.

Nature as we experience it now took millions of years to get to this point. The "working" state we have now is a result of a great many things coming together into a - for us humans - local minima that just happens to support life as it exists. And yet for humans in particular, things still don't work that well per se. There are generic diseases, mutations, etc that we have found varying ways to offset artificially. We - and all other animals - also get old, which is just short for all systems breaking down over time. Maybe all that can be considered "working" at a macro level, but it sucks at the micro.

I guess this is the distinction between a complex system and a complicated system.

Simple made easy would agree with you

I didn't read the article, but I read your response and it feels a bit heavy handed.

> What about the weather systems in the atmosphere? Could you argue that one of the most complex systems (maybe only second to the ocean current system) on the planet is not "working" in it's natural state?

Over geologic timespans, this is all just temporary thermodynamics. We will not always receive the same amount of energy over the sun. It will vary dramatically as the sun moves through its solar lifecycle.

Everything evolved around these dynamics in their current metastable states. Lots of optimization flux ebbing and flowing around the larger gradients and salients. All subject to perturbation such as asteroids and atmospheric carbon and solar death.

Life on earth has existed a third the age of the universe. That's a long time. Complex life has less than 600 million years left, and that's an upper bound estimate.

> Your limbic system is very very complex and is naturally in a state of working. No human intervention.

If you move around enough. If you vegetate on the couch all day, it's not.

We evolved biology that works under the conditions we evolved under. Through us into new conditions and you stretch the behaviors of these systems.

We evolved under the gas energy exchange of our gravity well. Put us into a different environment and everything breaks. Give us little pocket dopamine rectangles and suddenly we stop reproducing.

I don't have this much patrience, I had a really similar issue, I had a streaming service as an extra perk for somethimg I have- When It stopped working - Something hit me, this isn't a troubleshooting session but twas the call of the seas.....

I would have given up after the first failure, and used a different streaming service. I have zero patience for consumer technology that doesn't work, after spending every work day dealing with enterprise technology that doesn't work.

Is your stack really that rough?

9900 users have working laptops and 100 have hardware problems.

The laptop repair team's work queue, though? 100% broken laptops.

Yep, large diverse org, something's always broken somewhere.

Well said xD

Completely off topic but the title made me wonder if there’s any subscription service that cancels you if you don’t use it? Not quite usage based billing - plans that cancel (or pause) without use? I can’t think of any - terrible business model of course

> plans that cancel (or pause) without use?

Kagi is one of them.

[1]: https://help.kagi.com/kagi/faq/faq.html#fair-pricing

I recall a db service does that too long ago. Although I'm not sure if they changed policy as it's been a while.

In 2020 Netflix claimed they would start to automatically cancel inactive accounts [1], but the post has since disappeared. I also remember Microsoft saying the same thing about Xbox Game Pass but have not searched for their statement.

[1] https://web.archive.org/web/20200522032356/https://media.net...

Kagi arguably “pauses” your subscription if you don’t use it in a month. They give you a credit at the end of the month that then applies to the next month, so that people aren’t charged if they aren’t using it.

Slack. Deactivates a seat after 28 days of inactivity. A really good practice.

edit: and obviously reactivates after activity

Uber dropped me and closed my account. They gave me warning, but I was fine with it and took no action to prevent them from doing it. The email said something about being based on no activity.

I know it's not exactly a subscription service though where they were making money from me whether I used the service or not. It was just surprising to me that someone felt the need to develop a cancel user purging system at all.

Uruky [1] (a Kagi alternative, but simpler and EU-based) does this in a more intentional way: You can only pay once for a period of time (one, three, six, or twelve months), there is no subscription! If, after that period, you want to keep using it, you have to pay again.

This was mainly done for privacy reasons, so less data needs to be stored in the account (it also only uses a randomly-generated account number, like Mullvad), but honestly, we don't want people paying that don't use the service.

And you're right, it makes "making money" much harder, but also, feels more honest.

[1]: https://uruky.com

Kagi does something close to that. https://help.kagi.com/kagi/plans/plan-types.html#monthly-pla...

tailscale used to do this for teams ("active user billing"), but recently changed pricing models to be purely seat-based.

they had a whole webinar about it with all sorts of justification, although most of it sounded like mba-isms to me.

We take a slightly different approach for https://pico.sh -- no automatic subscription, but we charge for an entire year. It's great for us because each sub is a year and if someone truly isn't using our services then it'll quietly drift into the background for the end-user.

Tarsnap has you put money into a stored balance, and when that balance goes to zero (and after a grace period), they delete your backups.

Not exactly a subscription since it's a stored-balance system, but still.

Morning brew (the email newsletter) iirc would unsubscribe you if you didn't open their email for a while. Not a paid sub though

Xbox/Microsoft Game Pass actually automatically canceled for me when I hadn't used it.

netflix did this. I didn't use streaming for a (long) time, they turned it off. Kudos.

Not any more. They'll keep billing you, and they'll even follow you getting a new card even if you don't explicitly update your account with the new card info. That does not sound like the same mental space as not charging people for being inactive.

I’m stealing this idea!

Kagi does this

I can easily see myself failing to catch this type of bug, especially if when you run it locally, the latency on jobs from enqueue to finish is aboue 5ms, whereas it probably fluctuates in production from a few ms to 5 minutes. It probably passed QA when latency was low.

If the desire is to mostly keep this architecture, the flag in the DB for "has a streaming account linked" needs to not be a boolean, and then you could have a third state besides "Ready to link" and "Link": 'Pending unlink' which would cause the UI to ask the user to stand by until the streaming site confirms the unlinking. Mildly inconvenient for the 0.1% of people who need to unlink just to immediately re-link, but better than buggy.

In the spirit of yes, and: how about a subscription similar to a pay as you go phone plan? Pay for the month, and when you don't pay, then you don't get to keep going. After a couple of months, they unsubscribe you, get rid of your account, etc. More often than not, the first thing I do when I sign up for a service is cancel it (after confirming I can use it for the billing period).

This is one of my favorite things about Mullvad VPN.

> More often than not, the first thing I do when I sign up for a service is cancel it (after confirming I can use it for the billing period).

That's what I loved about Apple controlling their App Store subscriptions. All subscriptions keep working through the billing period and there isn't really anything an app can do about it. It also gave me an easy centralized place to view and cancel subscriptions.

Apple got smart on at least part of it and if you cancel a free trial/bonus for buying a thing, it cancels immediately - not at the end of the trial.

Uruky [1] (a simpler and EU-based Kagi alternative) does this! You only pay once for a period of time (one, three, six, or twelve months), and there is no subscription! If, after that period, you want to keep using it, you have to pay again.

[1]: https://uruky.com

I used to use one-off "virtual" credit card numbers for this kind of thing. My bank (MBNA, at the time) let me set a balance limit and / or expiration date. It was exceedingly handy, so the feature was axed when Chase acquired them.

Capital One has something like this but when I read the fine print I got the impression that it didn't work for this for subscription services. Something like how they automatically update businesses when you get a new card number due to fraud happens, defeating the purpose as long as it's marked as a subscription service somehow.

For quite a few services you can get the pay-as-you-go experience by getting yourself a gift subscription.

I can't imagine the frame of mind the author has to be in to think that there's moral value in not "naming names" of corporations that do things badly, as if they are people who can be offended. Although they also write cringe things like "to the builders <heart emoji>" so perhaps I will just never understand them.

I think it would distract from the points he's making. The article could be misread as a rant about a bad time he had, when it's actually meant to make a specific point about considering async vs sync transactions and what happens when they're combined in the same system.

And I don't believe that only one streaming service and one bank makes such mistakes.

Author here. This! The point isn't to name and shame, but to illustrate a complex failure mode. I too could see myself writing this bug!

There’s a lot of bleeding heart people like this. They add variety to the world. The downsides being things you mention, but it’s usually more palatable than someone on the other end of the spectrum.

My idea of the term "bleeding heart" is more like "painfully aware of the plight of people (and often wants to make sure you're also painfully aware of it)", whereas the author's tone struck me as simply charity, and I enjoyed it as such.

Sometimes they're worried about getting sued.

I don't think its for moral value but rather they want to make a general point. For example Netflix couldn't care less if they were named or not named in this blog so what purpose is there to "name and shame" them? Most normal people dont even know what a request is so its not like there is any reputation damage risk here for Netflix and the author can write without any bias and talk about general tech and its shortcomings/quirks.

That is the frame of mind and seems pretty reasonable.

This was actually detectable in the calls to the providers if they went as described. The credit card company tells them the perk subscription is active and the streamer says it has been cancelled. ("There was a valid activation of the streaming perk, and a confirmation from the provider" vs "The subscription had been activated, then cancelled in an orderly fashion about 5 minutes later.")

This is perfectly in line with the actual async problem, but differs from what they put in the summary ("Support on both sides saw an orderly activation followed by an orderly cancellation, with no errors").

Pirating content is so much easier than this

Seriously, download a few containers from docker, make some quick config changes, and now you have a comparable experience to these subscription services

> download a few containers from docker, make some quick config changes

You don't even have to do that if you don't want to. There are so many good piracy streaming sites which are on par, if not better, with commercial ones.

https://fmhy.net/video

I parked the boat, myself, for about a decade.

Then enshittification hit the forever-forking streaming services.

So it's often been back to the high seas, matey... hell, I even buy CDs, now (it's easier for rarer stuff, and then remains forever mine).

I thought this was going to be about a technique I've been using lately for free trials.

And it goes like this, I sign up for the free trial and immediately cancel it, so I don't forget. It will still let me have the free trial until the end of the time specified.

And then if I do decide to keep it, the worst thing that happens is it just stops working at the end of the free trial and then I go ahead and re-enable it.

Says me a ton of money and anger from forgetting to cancel a free trial.

Unfortunately that immediately terminates access for some subscriptions, such as Apple Music. :-(

I love stories like this. I loved watching House MD and I always fantasized about writing a script for a drama series about a software company (with a zany, but attractive, cast of characters) that debugged a different problem every week.

I even started collecting true stories of software bugs, like the time a county court system decided to send out 10x the number of jury summons and caused a major traffic jam.

That sounds amazing. Off-by-one disasters, time conversion problems, imperial instead of metric messups. Halt and Catch Fire & Mr Robot were great. We need more like that, and less like CSI:Cyber.

Shameless plug slightly related to this pain of subscriptions. I've been cooking https://github.com/gchamon/buzz for a few weeks. It's a replacement for zurg or debridmediamanager. It also serves as an alternate frontend for real debrid so you can load the legal copies of the movies you own using honest trackers.

Full disclosure, I haven't written a single line of code there, but it's been refactored and improved a lot, so it isn't your average vibecoded project, it's been brought up with agentic engineering and countless hours of manual testing.

Just want to say that I loved both this and the wifi raining story

> Here, purely-async makes more sense than purely-sync:

> From a user experience perspective, the user has no need to wait around until the link is severed. They expressed the intent to sever the link, and were told this would be accomplished. Generally, that's sufficient.

That's incorrect I'm afraid. The reason the flow is synchronous for linking is so that the user can consume the service as soon as they link it. Async means they would have to wait, no user wants to wait.

Similarly, cancellation is asynchronous so that the service doesn't stop immediately. This benegits both the service and the bsnk or credit card company since users often do change their minds and resume the service during the "cool-off" period.

tl;dr, the current logic is correct, it just does not work for your use-case, which is understandably frustrating.

From the linked article:

> Linking the accounts between the bank and the streaming provider is a synchronous process, for both technical and user experience reasons. For example, it makes sense to get the user access as quickly as possible! "Click here and you're done" feels good, "click here and we'll send you an email in a few minutes" does not.

Self cancellation sounds like a feature to me.

It's not self-cancellation in the way you're interpreting the title :)

It's a completely different failure mode, but this reminded me of an infuriating situation I had a while ago with single-use URLs and safe-links functionality in MS Outlook. I can't remember the exact details but the gist was that Outlook would replace the actual URL with an indirection which would scan the page to check it was safe when you clicked. The result was that every time I would request an activation link it would be expired by the time I clicked it. I ended up having to reverse engineer the safe-links encoding algorithm to get the original URL. I wasted hours.

The entire article reads as an excellent example why piracy continues to exist.

I.e.: https://xkcd.com/488/

Once you have the .mkv on your local computer system, then only actual hardware failures will prevent you from watching it whenever, wherever, and for as many times as you want to do so.

Yep, at some point I lost the few things I had on iTunes.

...yeah.

What's everyone's favorite torrent site these days? Mine is Bitsearch, it has absolutely everything

The author's LinkedIn style usage of emojis repulsed me and had me closing the article.

There is a coffee shop here that has a membership plan (you can roast at the shop it’s cool. Membership = no charge to roast and discounts on beans). It’s monthly and you have to re-up to keep it. It’s great and I’m happy to support them.

The fact that a subscription designed to cancel itself is considered innovative tells you everything you need to know about how low the bar is. We've normalized making it hard to leave to the point where 'letting you go' is a feature.

> The fact that a subscription designed to cancel itself

But it wasn't? TFA is describing a technical issue that kept cancelling a subscription. This is not a "we've noticed that you haven't been using the service and paused billing" situation.

Reply on news.ycombinator.com