Daniel moved house this week. New place, new mobile provider, the usual chaos. And he hits the first wall: to activate the SIM, he has to speak to a human. There's no self-service path. The activation line closes at four in the afternoon. That's wall one. Wall two, his credit card company's call centre also shuts around four. Wall three: he spends an entire day chasing authorisation for an early refill of Vyvanse — a controlled substance he takes regularly. All the paperwork is in order, but some system somewhere says an authorisation is missing that nobody flagged. Every office he needs to reach either closes at two or simply doesn't answer, including ones whose posted hours say they're open until eight. He solved it, but it cost him a day.
And he's been here ten years.
Ten years, and it still amazes him. He grew up in Ireland, and the contrast keeps catching him off guard. Here's what he actually wants us to dig into: how does a developed economy end up normalising a situation where essential services — pharmacy authorisations, telecoms activation, banking — are reachable only through a person, only for a few hours a day, and closed more or less at random the rest of the time? Mobile is supposed to be one of the better-served consumer markets in Israel. The sector was deliberately reformed, there are several competing operators, and it's the example people reach for when they argue competition fixed Israeli customer service. And yet the only way to turn on a SIM card is a human being who goes home at four. His question cuts deeper than the frustration: if a system can't absorb events that are routine but not standard — and he means both of those words deliberately — is it a working system at all, or just one that happens to function for people whose lives never deviate from the median?
Routine but not standard. That's the phrase that sits with me. Early refill on a regular prescription. Activating a phone line when you move house. These aren't edge cases — they're things that happen to everyone eventually. But the system treats each one like an anomaly.
So what does it mean when the system only works for the median life?
Let's start with what we're actually looking at. Essential services — telecom, banking, healthcare — designed for a narrow operational window. Not designed for resilience, not designed for the person whose schedule doesn't line up with office hours. Designed for the person who can call between eight and four on a weekday and wait on hold.
And if you can't, the service simply doesn't exist for you that day. This isn't about a power outage or a supply chain disruption — those are genuine crises. This is about a system that, on a completely normal Tuesday, cannot process a routine transaction because the clock hit four. The fragility is built in. It's not a bug that emerged — it's the architecture.
Daniel's framing is careful here. He's not describing a broken system in the sense of something that collapsed. He's describing a system that functions exactly as designed, and the design excludes a huge swath of normal life.
The Israeli context makes this particularly strange because the country markets itself as a tech powerhouse. Startup nation, innovation everywhere, and you can't activate a SIM card without a human being who keeps business hours.
The mobile sector is the interesting case because it's the one everyone points to as the success story. The 2012 reform, new operators entered the market, prices dropped dramatically. Competition worked — or so the narrative goes. It worked on price. And it really did — mobile bills in Israel fell off a cliff. But competition didn't touch service design. It didn't touch hours. It didn't touch self-service.
So you got cheaper access to the same bottleneck. Let me give you something concrete. Golan Telecom's customer service hours — phone support is eight in the morning to four in the afternoon, Sunday through Thursday. WhatsApp support runs until ten at night, which is better, but for the actual activation of a new line, you're back to the phone channel and the four o'clock cutoff.
So there's a channel open in the evening, but it can't do the one thing a new customer needs?
That's the pattern. Different channels have different capabilities, and the one that handles the essential transaction closes earliest. It's not that the company lacks the infrastructure to be available — they've already built the WhatsApp support system, staffed it until ten. But they haven't connected that channel to the activation workflow. So it's not a resource problem. They have people working until ten. They just can't use them for the thing Daniel needed.
And that tells you something about how these systems grow. They're not designed top-down with the customer journey in mind. They accrete. Phone support was built first, with certain capabilities. WhatsApp was added later, with different capabilities. Nobody went back and unified them. So you get a service that looks modern — WhatsApp support, very slick — but the actual transaction path is still the old one.
The Vyvanse case is even starker. Daniel had the paperwork. The prescription was valid. The only problem was a missing authorisation in some internal system — something nobody had told him he needed, and something that apparently nobody in the system could see until he called.
As a retired pediatrician, this one hits close to home. Vyvanse is a controlled substance — lisdexamfetamine — used for ADHD. The controls around it exist for good reason. You don't want people stockpiling amphetamines. But the control mechanism has to be proportionate. When a patient with a valid prescription, a documented history, and all the right paperwork still has to spend a day on the phone because some internal flag got tripped and nobody can explain why — that's not drug control anymore. That's administrative failure wearing a security hat.
And the offices he needed to reach. Some closed at two. Some didn't answer despite posted hours saying eight PM. That's the "closed more or less at random" part. The randomness is almost worse than the limited hours. If you know a place closes at four, you can plan around it. If the posted hours say eight but nobody picks up, you're just shouting into a void. You don't know whether to keep trying or give up.
Daniel spent a day on this. He's self-employed, he can shift his schedule. Someone working a shift job, someone who can't take personal calls during the day — they just don't get their medication. This is the point where "routine but not standard" becomes a real filter. The system works if your life matches the assumptions: you're available during business hours, your prescriptions renew on schedule, you never need an early refill. Deviate from any of those, and the system doesn't absorb you — it rejects you. Not maliciously. Just structurally.
Let's pull on the mobile thread a bit more. The 2012 reform is genuinely interesting because it's held up as proof that competition solves consumer problems in Israel. What actually happened?
Before 2012, the mobile market was essentially a triopoly — Cellcom, Partner, and Pelephone controlled something like ninety-five percent of the market. Prices were high, barriers to entry were enormous. The reform, driven by the Ministry of Communications under Moshe Kahlon, reduced those barriers. New mobile virtual network operators could enter. Golan Telecom and HOT Mobile launched with aggressive pricing — unlimited plans for something like ninety-nine shekels a month when the incumbents were charging two or three times that.
And the incumbents had to respond. They slashed prices. Within a year or two, the average mobile bill dropped by something like fifty to seventy percent. It was a genuine consumer victory. But here's what didn't happen: nobody competed on service hours. Nobody said, "activate your SIM at midnight through our app." Nobody made self-service a differentiator.
Partly because price was so dominant that everything else faded into the background. When your bill drops from three hundred shekels to ninety-nine, you don't ask about the activation process. You celebrate the savings. The operators knew this. Competition was a price war, and price wars don't produce service innovation — they produce cost cutting.
And a human-only activation line that closes at four is a cost structure. Staff it from eight to four, five days a week, and you've covered most of your customers. Extend it to eight PM or add weekend hours, and your costs go up. Add self-service infrastructure, and you've got upfront development costs plus ongoing maintenance. The economics are real. But the minority who need those extra hours aren't a random fringe. They're people who work, people who travel, people with medical needs that don't follow a nine-to-five schedule.
And the cost calculus doesn't capture their experience. If you're the mobile operator, you don't see the day Daniel lost. You don't see the medication that didn't get picked up. Those costs are externalised — borne by the customer, not the company. So the market, even a competitive market, doesn't price that in. Not unless customers start choosing providers based on service availability rather than price. And the evidence suggests they don't — price dominates, coverage dominates. Service hours are way down the list until you need them, and by then you're already a customer.
There's another layer here. The human-only path. It's not just that the hours are limited — it's that there's no way around the human. No automated activation, no self-service portal, no eSIM download that just works. In a lot of countries, you can switch providers by scanning a QR code. The phone downloads the profile, and you're active in minutes. The technology exists and is widely deployed.
So why not here? Some of it is regulatory. Israel's telecom regulator has been cautious about eSIM adoption — there are security concerns, identity verification requirements. The mobile operators are required to verify the identity of the person activating a line. A human on the phone can check an ID number against a database. An automated system can too, but building that system costs money, and the regulatory framework has to approve it. So the human bottleneck is partly a regulatory bottleneck wearing a customer service uniform.
The medication authorisation is even more clearly regulatory. Vyvanse is a controlled substance. The Ministry of Health sets rules about prescribing, dispensing, and refilling. Pharmacies and HMOs build processes around those rules. Each layer adds a human checkpoint. The result is a chain of offices, each with its own hours, its own phone queue, its own internal systems that don't talk to each other. Daniel's missing authorisation — some flag in some system that nobody could see until he called — that's the classic symptom of systems that don't interoperate. Each office has its own view of the data. None of them can see the whole picture.
This is the part that drives me up the wall as a former clinician. You have a patient with a legitimate need, a valid prescription, a documented history — and the barrier isn't medical. It's not a clinical judgment call. It's an administrative ghost in a database that nobody can locate or explain. The system is so fragmented that even the people inside it can't navigate it. And Daniel's experience — calling offices that are supposed to be open until eight and getting no answer — that suggests the fragmentation goes all the way down. The posted hours don't match reality. Maybe the person who usually answers left early. Maybe the office is technically open but the phone isn't being monitored. Whatever the reason, the result is the same: the system as experienced by the user is less functional than the system as described on paper.
Let's step back to the bigger question. How does a developed economy normalise this?
I think there are a few things happening at once. One is simply inertia. These systems were built at a time when business hours were the default assumption for everything — banks, government offices, doctors. The infrastructure, the labour contracts, the regulatory frameworks all assume a world where services are available during working hours because that's when the workers are there. Updating that assumption touches everything — employment law, union agreements, IT systems, training. None of it is impossible, but each piece is friction.
The second factor is what I'd call the resilience of low expectations. When everyone around you also expects services to close at four, the inconvenience stops feeling like a failure and starts feeling like the weather — something you complain about but don't expect to change. Daniel's perspective as an immigrant is useful here because he still has the contrast. He remembers what it was like in Ireland, where the baseline assumption is different. Most people who grew up in this system don't have that contrast — or they've been in it long enough that the contrast has faded.
The third factor is that the people who design these systems are, almost by definition, people whose lives fit the median. They work office jobs. They're available during business hours. They don't experience the system's failures personally, so the failures don't feel urgent. The manager who sets the call centre hours is someone who can make personal calls during the day. The regulator who approves the activation process has never tried to activate a SIM at six PM. The distance between the decision-maker and the edge case is structural.
And the edge case isn't even an edge case. Daniel's examples — moving house, travelling, needing an early refill — these are normal life events. They happen to a huge proportion of the population every year. Calling them edge cases is part of the normalisation. "Edge case" implies rare, unusual, not worth optimising for. "Routine but not standard" is more honest. It acknowledges that these events are common while recognising that the system isn't built for them.
So let's tackle Daniel's question directly. Is this a working system? If you define "working" as "processes the median transaction successfully most of the time," then yes. Most people get their SIMs activated eventually. Most prescriptions get filled. The system doesn't collapse. But that's a pretty low bar. A bridge that's passable most days but floods whenever it rains isn't a working bridge. Right now, it's fair-weather infrastructure. It handles the standard case and dumps the cost of every deviation onto the user.
The cost isn't just time. For Daniel, it was a day of work. For someone on Vyvanse who can't navigate the authorisation maze, the cost might be going without medication. ADHD medication isn't optional for a lot of people. Going off Vyvanse abruptly can cause withdrawal symptoms, rebound ADHD symptoms, mood disruption. A system that makes refills contingent on a day of phone tag is a system that's creating medical risk.
The system doesn't see that risk because it's not measured. No HMO tracks how many patients missed doses because they couldn't get through to the authorisation office. The metric doesn't exist. These systems are fragmented across multiple organisations — the HMO, the pharmacy, the Ministry of Health, the prescribing doctor. Each one is responsible for its piece. Nobody is responsible for the patient's experience across the whole chain. So the failure has no owner.
The mobile activation is less consequential but structurally identical. The company is responsible for the activation process. The customer is responsible for being available during business hours. If the customer can't be available, the failure is framed as the customer's problem, not the company's. And the company probably has metrics showing that ninety-something percent of activations complete successfully within a day or two. Those metrics don't capture the people who gave up, or who delayed their activation by a week because they couldn't call during hours. The data looks fine because the data is shaped by the system's own assumptions.
What would a resilient system look like? Self-service is the obvious starting point. eSIM activation that works through an app or a website, with identity verification handled automatically. The technology exists. It's deployed in dozens of countries. The barrier isn't technical — it's regulatory and organisational.
For the medication case, it's about interoperability. If the HMO's system, the pharmacy's system, and the Ministry of Health's controlled substance registry could talk to each other, Daniel's missing authorisation would have been visible immediately. He wouldn't have had to call anyone.
Interoperability in healthcare is a massive challenge everywhere, not just in Israel. But the Israeli system has some particular features that make it harder. The HMOs — Clalit, Maccabi, Meuhedet, Leumit — each have their own IT infrastructure. The Ministry of Health has its own systems. Pharmacies have theirs. Connecting all of them is a project of enormous scope, and the incentives aren't aligned. Each organisation benefits from controlling its own data.
The patient falls through the gaps between them. Every time.
There's another dimension here. The "closed at random" phenomenon. Daniel mentions offices that simply didn't answer despite posted hours saying they were open until eight. That's not a policy choice — that's an execution failure. But it's an execution failure that's been normalised to the point where it's barely remarked on. Posted hours that don't match reality are a signal. They tell you that the organisation doesn't consider the posted hours to be a commitment — they're more like a suggestion. And when the organisation doesn't take its own hours seriously, customers learn not to either. You call, nobody answers, you shrug and try again tomorrow. The failure becomes part of the expected experience.
Which loops back to the low expectations point. If everyone expects the posted hours to be aspirational, nobody complains when they're not met. The system never faces pressure to improve because the failure isn't registered as a failure. Daniel's surprise is the tell. He's been here ten years, and it still amazes him. That means the contrast is still alive for him. He still has the expectation — formed somewhere else — that posted hours mean something, that essential services should be reachable, that a system should absorb routine deviations. Most people born into this system lose that expectation. Or never develop it in the first place.
The normalisation is partly a story about reference points. If you've never lived somewhere where you can activate a SIM at midnight through an app, you don't know that's an option. The inconvenience feels like the natural order of things. And that's how a developed economy ends up with this situation. Not through a deliberate choice to limit service, but through a combination of inertia, fragmented regulation, misaligned incentives, and the slow erosion of expectations. Nobody sat down and said, "let's make sure nobody can activate their phone after four." It's just that nobody ever sat down and said, "let's make sure they can."
The market reform story is instructive here. Competition was supposed to fix Israeli customer service. And in the mobile sector, it did — on price. Prices dropped dramatically. That's a real win. But the reform didn't touch the structural issues: the regulatory requirements, the human-dependent processes, the limited hours. Competition optimised the thing that was easiest to compete on. Price is easy to compete on because it's visible. You can put it in an ad. Service hours are harder to advertise and harder to verify — you only find out if it's true after you're already a customer.
The market didn't fail exactly. It just solved a different problem than the one Daniel is experiencing. Competition fixes the things that customers can easily compare and switch over. Everything else — the service design, the hours, the self-service infrastructure — those are second-order considerations that most customers don't evaluate until it's too late.
Let's bring this back to Daniel's core question. Is a system that can't absorb routine but not standard events a working system? I'd say it's a system that works only if you define "working" narrowly enough. It processes the median case. It doesn't collapse. But it externalises huge costs onto anyone who deviates from the median. Those costs are real — lost workdays, missed medication, delayed travel, stress. They're just not counted in anyone's metrics. So it's a lottery. If your life happens to match the system's assumptions, you're fine. If it doesn't, you pay a tax — in time, in money, in health. And the tax is invisible to the people who designed the system.
The lottery framing is right. And like a lottery, the odds are fine for most people most of the time. But "most people most of the time" isn't the same as "everyone always." A working system, in any meaningful sense, should handle the routine deviations. Not the once-in-a-century flood — the normal rain.
Hilbert, you've been quiet. You don't buy this?
Hilbert: I buy most of it. The description is accurate. The frustration is real. Where I part company is the conclusion that these systems aren't working. They're working exactly as designed — and the design is rational, not broken. A mobile operator that staffs its activation line from eight to four, five days a week, covers something like ninety percent of its new customers within twenty-four hours. The remaining ten percent wait an extra day or two. The cost of extending hours to capture that last ten percent is disproportionate. You're doubling your staffing window for a marginal improvement in service. That's not a failure — that's a business making a sensible tradeoff.
But the cost of that tradeoff lands entirely on the ten percent. The company doesn't feel it.
Hilbert: The company feels it if those customers leave. But they mostly don't, because every other operator has the same hours. The market has settled into an equilibrium where limited hours are standard, so no individual company gains by breaking ranks. That's not fragility — that's a stable equilibrium.
And the medication case?
Hilbert: Controlled substances require controls. The authorisation chain exists because Vyvanse is an amphetamine. The fragmentation between HMOs, pharmacies, and the Ministry of Health is inefficient, sure. But the alternative — a single unified system that processes everything automatically — that's a centralisation of sensitive health data that comes with its own risks. Privacy risks, security risks, single-point-of-failure risks. The current mess is partly the price of keeping data compartmentalised.
Interoperability doesn't require centralisation. It requires standards — APIs, data formats, consent frameworks. Those exist. They're deployed elsewhere.
Hilbert: Fair. But deploying them costs money, and in a system where healthcare is already stretched, the investment in interoperability competes with investment in actual medical care. More nurses versus better IT. I know which one I'd pick.
That's a false choice, though. The IT failure creates more work for the nurses. Daniel's day on the phone — he was consuming staff time across multiple offices. If the systems talked to each other, that time would be freed up for actual care.
Hilbert: In theory. In practice, IT projects in healthcare have a track record of running over budget, underdelivering, and creating new problems. I'm not saying the current system is good. I'm saying the alternative isn't as clean as you're making it sound.
I'll concede that. Healthcare IT is a graveyard of ambitious projects. But the current equilibrium has a hidden cost that your framing doesn't capture. The ten percent who can't call during business hours aren't a random sample. They're disproportionately shift workers, people with disabilities, caregivers, people with chronic conditions that require frequent authorisations. The system's design imposes the highest costs on the people least able to absorb them.
Hilbert: That's a distributional argument, not an efficiency argument. The system is efficient for the median. The distributional problem is real, but it's a policy question — do we want to spend more to serve the tails? — not evidence that the system is broken.
I think the distributional question is the whole point. A system that's efficient for the median and brutal for the margins isn't a well-designed system. It's a system that's chosen a particular definition of efficiency that excludes a lot of people.
Hilbert: Nobody chose this. It accreted, as you said. And changing it means making explicit choices about where to spend scarce resources. I'm not convinced that extending call centre hours or building self-service infrastructure is the highest-value use of those resources, compared to, say, reducing drug prices or shortening wait times for specialists.
I don't think we need an obviously better design to call the current one inadequate. We just need to recognise that the costs it imposes are real and concentrated, and that the people bearing them don't have much voice in the design process.
Hilbert: On that, we agree.
Where does that leave us? The system isn't broken in the sense of non-functional. It processes most transactions. But it's fragile in ways that matter, and the fragility falls hardest on people whose lives don't fit the template. The question is whether that fragility is inevitable or chosen — and I think the answer is that it's chosen by default, through inertia and misaligned incentives, rather than through any deliberate weighing of costs and benefits.
The mobile case is the clearest example because the self-service technology exists and is deployed globally. The barrier isn't technical or even really economic — it's regulatory and organisational inertia. That's a choice, even if nobody is consciously making it.
The medication case is harder because it involves multiple organisations with legitimate but conflicting priorities. Drug control, patient privacy, clinical autonomy, cost containment. Solving that isn't simple. But not solving it has costs too, and those costs are currently invisible to the system.
Daniel's phrase — routine but not standard — is going to stick with me. It names something that's easy to overlook. These aren't rare emergencies. They're normal life. And a system that can't handle normal life is a system that's failing in slow motion.
This has been My Weird Prompts. Thanks to Hilbert Flumingtop for producing and for pushing back where it needed pushing.
If you've got a story about a system that worked perfectly on paper and fell apart the moment your life deviated from the script, we'd love to hear it. Email the show at show at my weird prompts dot com.
We'll be back soon.
