Convenience Is the Bait: Why the Hard Way Is Worth It

by Leo Gaggl • WOL Notes

If I had a dollar for every eye-roll my refusals have earned me, my new phone would have cost me nothing. My partner’s arrives whenever I say “no, I’m not on WhatsApp”. My tech colleagues have their own version, usually when I explain, again, that I’m not on Twitter, or X, or whatever its owner has decided to call it this week. Friends ask why I can’t just use the family group chat like a normal person. Collaborators send me Facebook event links and then a second message, slightly weary, with the same details in plain text.

The latest round started with a phone. My trusty old Pixel’s battery began ageing faster than I did, sealed into its case like every other modern phone, so I bought a Pixel 10 Pro for its camera. The first thing I did after unpacking it, before I’d signed into anything, was wipe the factory Android firmware and install GrapheneOS. It was a deeply satisfying job. It was also the start of several weeks of research, trial and error, and explaining to the people around me why I keep making my life harder than they think it needs to be.

A feathered fishing lure with a bare hook dangling from a line against a cloudy sky, by Yortw, on Flickr
"Fishing Lure" by Yortw (CC BY-NC 2.0)

The short version: convenience is the bait. It’s the first stage of a pattern we now have a name for, and the inconvenience is the price of not taking it.

A Word for What We Already Felt

Cory Doctorow named the pattern “enshittification” in 2022, and in 2024 the Macquarie Dictionary made it its Word of the Year, defining it as “the gradual deterioration of a service or product brought about by a reduction in the quality of service provided, especially of an online platform, and as a consequence of profit-seeking.” Doctorow’s own description is more vivid: “first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves.”

The same year, the Australian National Dictionary Centre picked “Colesworth” as its word of the year, for the supermarket duopoly that sets what most of us pay for food. Two duopolies, one year. Coincidence, but a telling one. Coles and Woolworths sell us convenience by the kilo. Google and Apple sell it by the app. In both cases the convenience is real, and in both cases it’s how they hold on to us once the deal stops being a good one.

Stage one of enshittification is the good part. The free email account with more storage than you’ll ever use. The phone that sets itself up in three minutes because it already knows who you are. The group chat everyone is already in. People get hooked by the convenience, and by the time the extraction starts, leaving costs more than staying.

Convenience Sends Its Own Bill

Enshittification is only half the problem. Even when a platform behaves itself, convenience sends a bill of its own, and we already know how that works from the physical world.

Colesworth’s shelves are a good place to start. A 2019 study of the national nutrition survey found that ultra-processed foods (packaged, ready-to-eat products assembled from industrial ingredients you’d never find in a home kitchen) supplied 42% of the energy Australians ate. In 2024 an umbrella review in The BMJ, drawing on meta-analyses covering almost 10 million people, linked them to 32 different health problems, with the strongest evidence for early death, heart disease and, tellingly, anxiety and other common mental disorders. Those are associations, not proof of cause, but none of the evidence points the other way. Add the car to the corner shop, the lift instead of the stairs and the drive-through instead of the kitchen, and you get a life engineered to take the effort out, with a body that pays for it later.

None of that convenience is forced on us. It’s offered, it’s pleasant, and it’s cheaper in the moment than the alternative. That’s exactly why it works.

Online life runs on the same recipe. The infinite feed is a packet of chips that never empties. Autoplay is the next episode nobody chose. Notifications are a vending machine at the end of every corridor. And the newest product on the shelf offers to do our thinking for us. The evidence here is younger, so here’s how solid it is. A 2020 McGill University study found that drivers who leaned heavily on GPS had worse spatial memory when they had to find their own way, and in a small group retested three years later, more GPS use went with a steeper decline. A 2025 survey of 319 knowledge workers by Microsoft Research and Carnegie Mellon University found that the more people trusted generative AI, the less critical thinking they reported doing. No single one of these studies settles the question. Together they point the same way as everything we know about muscles: what you stop using, you lose.

Canberra has already conceded half the argument. Since 10 December 2025, social media platforms have had to keep Australians under 16 off their services, on the grounds that the feed harms developing minds. I have my doubts about the ban, but not about the diagnosis. And the feed doesn’t turn healthy on anyone’s 16th birthday.

The nuance matters, because not all convenience is junk. My washing machine doesn’t make me any less capable, and nobody needs to churn their own butter to stay sharp. The test is simple: does this convenience free up effort for something you care about, or does it remove effort you need to stay capable? A washing machine gives back an afternoon. A feed built to keep you scrolling takes the afternoon and your attention span with it. Finding my own way across town, doing the sums in my head and writing my own first draft are the mental equivalent of taking the stairs. My phone does sit in a dashboard holder with a map on it, but its voice stays silent for everything except alerts. I’d rather read a road sign than be talked at.

The same test applies to AI, and I won’t pretend I don’t use it. This year of working out loud would have been far less productive without the LLM box I built at home. It helps me check facts and rough out a first structure, which I then argue with. Self-hosted and offline-first is where I start, and I’m careful never to get locked into one tool or one company. The current AI bubble has plenty of problems of its own, which I’ve been writing about since Don’t Let the Asphalt Bury the Garden. But the convenience trap is a specific one: handing your thinking to someone else’s cloud until it’s the only way you know how to work.

A pantry full of packaged food and a phone that does everything for you run on the same business model: make the easy option so easy you forget there was ever another one. Fixing my own phone is the digital version of cooking from scratch. It’s slower, occasionally maddening, and it keeps me capable.

Twenty Years of Trying

I didn’t arrive at GrapheneOS fresh. I’ve been trying to get out of the Google and Apple duopoly for most of the smartphone era, and this blog’s archive is some of the evidence.

In 2010 I wrote about the iPad as a Trojan horse, worried that schools would hand every student a device that only worked through one company’s store and needed an Apple ID to function. I gave iOS a fair go in those early days. I lasted about six months before the walled garden drove me mad. Android was the open alternative, and I spent the next few years discovering how much Google’s version of open came with strings attached, and how far its appetite for data kept growing.

So I tried the alternatives. FirefoxOS on a Nexus S. Ubuntu Touch on a Nexus 4, and later an actual Ubuntu phone, the Meizu MX4. CyanogenMod on a Nokia X and a Sony Xperia Z2. A Fairphone. In 2015 I wrote Good bye Android? Hello Ubuntu! Not yet unfortunately…, which included this line: “I am quite prepared to forego some conveniences I got used to on CyanogenMod (Android) over the years.” I meant it then, and I mean it now. But the Ubuntu phone locked up on its home screen, the web apps refused to load, and I went back to Android for work.

Each of those projects was built by people who cared, and most of them are gone. FirefoxOS was discontinued. Canonical dropped Ubuntu Touch (the UBports community still keeps it alive, to its great credit). CyanogenMod’s company collapsed, and the community carried on as LineageOS. Each failure made the next attempt harder. Every abandoned phone left behind lost data, workarounds built around it, and a slightly lower tolerance for hope. The convenient path doesn’t just win by being easy. It wins because every failed alternative raises the cost of trying the next one.

Why GrapheneOS Finally Stuck

GrapheneOS is the first alternative that has stuck, for two reasons. It doesn’t ask you to trade security for freedom, and it doesn’t ask you to start from scratch.

The old custom ROM routine meant unlocking the bootloader (the low-level gatekeeper that decides what software a phone will start) and leaving it unlocked, which made the phone less secure than the stock one I’d just replaced. GrapheneOS installs from a web page, then you lock the bootloader again with GrapheneOS’s own keys, so verified boot still checks every start-up. Compared with the bootloader wrestling and sideloading of 2014, the install was a breeze.

The second part is the pragmatic genius of the thing. FirefoxOS and Ubuntu Touch asked people to move to a whole new platform with a handful of apps and a promise that more would come. They never did. GrapheneOS takes the world’s most widely used phone operating system and strips the bad stuff out. Underneath, it’s still Android, built from Google’s open source code, so almost every app still runs. You get a phone that works for you, without giving up the app catalogue that killed every earlier alternative.

The design choice that matters most is how it treats Google. GrapheneOS ships with no Google services at all. If you need them, sandboxed Google Play installs as an ordinary app, with no more access than any other app you install, and you can run it without ever signing into a Google account. On a stock phone, Google Play Services runs as a privileged system app, with special access to location, identifiers and the network. On GrapheneOS it has to ask, like everything else.

Then there’s the reason that drove the whole decision more than any other: I don’t want a so-called ‘assistant’ listening to me. Not in my home, not in my car, not in my pocket. On a stock Pixel, Google’s assistant is woven into the system. You can switch off “Hey Google”, but you can’t uninstall the Google app that listens for it, and getting rid of the rest of the bloat around it is close to impossible. Since September 2026, Google has been replacing Assistant with Gemini on Android phones, and once a phone has switched, there’s no going back. The worry isn’t hypothetical. Google agreed to pay US$68 million to settle a class action alleging that Assistant recorded conversations when nobody had said the wake word, and that people then saw ads based on what they’d said. Apple paid US$95 million to settle near-identical claims about Siri. Both denied wrongdoing. GrapheneOS doesn’t ask me to trust that the microphone is behaving itself. There’s no assistant there to listen.

The team also has its commercial head screwed on. In March 2026 they announced a partnership with Motorola, and the first Motorola phones with official GrapheneOS support are due in 2027. That’s what finally makes this more than a Pixel hobby. I’d have waited for one if my battery had let me.

GrapheneOS still stands on Google’s code and Google’s hardware, though, and that’s a risk I’ll come back to in part two.

What the Hard Way Actually Costs

To be fair to the eye-rollers: the inconvenience is real.

Installing GrapheneOS took an evening. Turning it into a phone that does what my old Pixel did, without handing my data back to Google, has meant a string of small, niggling fixes and one big fight with Android Auto. Most of my apps arrive through Obtainium, which pulls releases straight from each developer’s own site or GitHub page, or Aurora Store, which fetches them from Google’s catalogue anonymously. That works for most things. Then you hit something like Android Auto, which GrapheneOS will only accept from its own App Store or from a cable connected to a computer. That restriction is deliberate. Android Auto gets deep access to the phone, so GrapheneOS controls who can install it and what it can touch. I misread an error message, skipped the manual and spent hours installing it from my laptop over a USB cable, before finding it had been sitting in GrapheneOS’s own store all along. A proper RTFM moment, and an expensive one. Installing it was only the start. Then came the hunt for hidden permission switches that all default to off, and getting the phone and the car to talk to each other has been the hardest part of the whole exercise.

Something odd happened along the way. I started out fairly pragmatic, expecting to make a few compromises, and I’ve made some. But with every rock Google threw in my path, I got more determined, not less. Once the car finally connected, Android Auto hid every app I hadn’t installed from the Play Store. A hidden developer switch brought most of them back, but not my navigation app. That one, as GrapheneOS’s own documentation confirms, has to come from Google’s own store, with my Google login attached. Not even Magic Earth’s own fix for copies installed elsewhere gets past it. The app is the same, the code is the same, and it’s already approved for Android Auto (a fig leaf of “app store trust” I take apart in part two). The lengths Google goes to annoyed me so much that I’d rather put the phone in a holder on the dashboard for navigation than hand over my login to install it from the Play Store. So that’s what I do.

Notifications are another trade. Without Google’s push service, apps like the official Mastodon client poll for messages themselves, which costs battery, so I’ve spent time tuning which apps are allowed to run in the background and set up my own notification server. Some banking apps check whether the phone is “certified” by Google and refuse to run when it isn’t, and Google Wallet’s tap-to-pay won’t work at all. GrapheneOS has published a guide asking app developers to check the phone’s security directly instead. Some banks listen. Most haven’t yet.

So, yes: research, trial and error, and the occasional hour lost to a log file. I’ll go through the choices in detail, what worked, what didn’t, and what I still haven’t solved, in part three: . This one is about why it’s worth bothering.

The Case for Convenience

Credit where it’s due: there are good arguments for not bothering, and the people rolling their eyes at me aren’t wrong to make them.

Time is finite. Not everyone has the patience to chase down one niggle after another on a phone, and not everyone has the skills to try. I wrote about the one in five Australians who are digitally excluded a few days ago, and an approach that requires a laptop, a USB cable and a working knowledge of adb is not an option for most of them. Doing it the hard way is easier when you’ve had 30 years of practice.

Network effects are real too. The family group chat lives on WhatsApp because the family lives on WhatsApp, and my absence doesn’t move them anywhere. It just means someone has to forward me the photos. Opting out has a social cost, and some of it lands on other people.

Then there’s the systemic argument, and Doctorow makes it himself. Individual consumer choices won’t fix enshittification. What fixes it is policy: competition law, interoperability rules that let you leave a platform without losing your friends, and an end to the legal protections that stop people modifying the things they own. My phone choices don’t change Meta’s business model. Regulation might, and that’s part two.

Apple’s privacy pitch has some substance: an iPhone hands less of your data around than a stock Android phone. But the price is the Apple vortex. The hardware is overpriced, and it only works well with more overpriced hardware from the same company. The watch needs the iPhone, the iPhone wants the Mac, and every friend still on Android shows up as a green bubble. Apple doesn’t need your data as badly as Google does. It has your wallet instead, and every new device makes leaving dearer. That’s the same trap with nicer packaging.

Finally, nobody likes a zealot. If I made my partner feel judged for using Meta’s products, I’d deserve every eye-roll from her.

Why It’s Worth It Anyway

All of that is true, and the hard way is still worth it. Two reasons.

The first reason is who’s on the other side of the convenience. Look at what just happened in October. On 17 October, as the No Kings coalition ran a nationwide day of action to get out the early vote, Meta suspended the movement’s main Instagram account for about 90 minutes, and organisers say they were locked out of it for about four hours. Meta says an automated system did it in error and has apologised. The error may well be genuine. The problem is the power to make it, on the one day it mattered most. We’ve seen that power used here before. In February 2021, during its fight with the Australian Government over the News Media Bargaining Code, Facebook blocked news in Australia and, along with it, about 17,000 other pages, including hospitals, fire services, charities and suicide hotlines. Whistleblowers later alleged that the overreach was deliberate, a bargaining tactic. Meta says it was a technical error.

Then there’s Starlink. Overnight on 13 October, Ukraine’s military terminals on Starshield, SpaceX’s government version of Starlink, were throttled without warning at speeds above about 160 km/h, and drone missions failed. Elon Musk said limits are “decided by the US military and ultimately POTUS”. Trump denies ordering it. In 2025, Reuters reported that Musk himself had ordered Starlink coverage switched off over part of Kherson during Ukraine’s 2022 counteroffensive, which SpaceX called inaccurate. Whoever made the latest call, an army’s communications depended on the moods of one billionaire and one politician. Infrastructure that can be switched off on a whim isn’t infrastructure. It’s a favour.

The second reason is closer to home. Alternatives only survive if someone uses them. Every gardener who saves seed knows this: a variety that nobody grows is a variety that disappears, however good it is. FirefoxOS and Ubuntu Touch didn’t die because they were bad ideas. They died because not enough people used them through the awkward years. GrapheneOS got a deal with Motorola because enough people installed it on Pixels to prove there was demand. Every bug report, every email to a bank asking it to support hardware attestation and every write-up of a workaround makes the road a little smoother for whoever comes next.

Maurizio Cattelan's sculpture L.O.V.E., a giant white marble hand raising its middle finger against a blue sky outside the Milan stock exchange, by Ralf Steinberger, on Flickr
Maurizio Cattelan's "L.O.V.E." in Milan, photographed by Ralf Steinberger (CC BY 2.0)

Google won’t notice one middle finger. It notices when there are enough of them. Short of a regulator, that’s the only pressure a company this size responds to, and it’s why I’d rather be one of many than a lone crank with a strange phone. If this post helps one more person raise theirs, it’s done its job.

And none of this is purity. I run sandboxed Google Play, unsigned, because a few things still need it. I use a Pixel. I’m on Discord for projects that live there. Doing it the hard way isn’t about refusing everything. It’s harm reduction: finding where my convenience feeds the worst behaviour, and paying a bit of friction to stop feeding it.

The Slow Way Builds Something

My backyard in Adelaide started as bare dirt. Over the years, with constant ground cover, rainwater and patience, it turned into soil that holds water and teems with life. I now grow bananas there, in a microclimate that didn’t exist until we built it. There was no shortcut. You can’t buy that soil at Bunnings. It’s the result of choosing the slow way over and over again until the slow way becomes the landscape.

My phone is a much smaller version of the same thing. Every hour I’ve spent on it has built something I understand, that answers to me, and that I can explain to the next person who asks. I don’t expect my partner to install GrapheneOS, and I’ll keep forwarding the family photos the long way round. But if you’ve been feeling the enshittification and wondering whether there’s a way out, there is. It’s not as convenient as staying. That’s rather the point.

And for all the grumbling in this post, the result is the best phone I’ve owned. I’ve been through quite a few Pixels, and this one is better tuned than any of them: fast, smooth and remarkably easy on its battery, because almost nothing runs on it that I didn’t choose to put there. Android Auto, the banks and the rest still niggle, but I’m very happy with how it turned out.

That’s thanks to the GrapheneOS team, who build a hardened phone operating system and make it free for anyone to install, and to the free and open source developers behind nearly every app on my phone, many of them volunteers, writing software that answers to the people who use it rather than to advertisers. Thank you.

Start with one thing: one app, one account, one default you change.

Part one of three. This was meant to be a simple write-up of my GrapheneOS experience. Then every fix raised a why, and every why led somewhere else. Next: , on the regulators actually prying the doors open. Then , the technical detail: what worked, what didn’t, and the pitfalls I’ve already fallen into, so you don’t have to.

Sources

←The Network Is Not the Owner: Co-ops, Peer-to-Peer and Who We're Ready to Be

Comments

Be the first to comment! Reply to this post from your Mastodon/Fediverse or Bluesky account, or mention this post's URL in your reply. Your comment will appear here automatically via webmention.

Follow this blog on Mastodon at @gaggl.com@web.brid.gy or on Bluesky at @gaggl.com