The dominant narrative surrounding artificial intelligence is one of extreme concentration. We are told that AI is a race that can only be run by a handful of multi-billion dollar companies in California or Seattle, using massive server farms that consume megawatts of power and millions of litres of water. In this frame, the rest of the world is reduced to mere consumers of foreign models, source material for scraping pipelines, or low-paid data labelling labour.

In part one I explained why I keep doing things the hard way, and in part two who is actually prying the doors open. This is the hard way itself: the choices I made turning a Pixel 10 Pro running GrapheneOS into a phone I can live with, what worked, what didn’t, and what I still haven’t solved. If you’re thinking about making the switch, I hope it saves you a few of the Saturday afternoons it cost me.

In part one I argued that convenience is the bait, and that doing things the hard way is worth the friction. I also conceded Cory Doctorow’s point: individual choices won’t fix enshittification. Policy will. This post covers why GrapheneOS needs that policy to survive, who is actually writing it, where they should point it next, and what you can do to help.

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.

I should declare a bias before I start. I grew up in regional Austria surrounded by agricultural co-operatives. The Lagerhaus sold the seed, the fertiliser and the diesel. The dairy co-op collected the milk. I’ve known them from the farm gate and, briefly, from inside the farmers’ league that runs alongside them. I’ve seen what they do well when their members actually run them, and I’ve seen what happens when governance gets comfortable and they start behaving like the corporations they were founded to keep at bay. They remain, perhaps sentimentally, a favourite of mine.

Part one ended in a Tuscan osteria where everything on the table came from the valley outside. This one starts in a pub in an Australian wine region, vines on every side, where I asked what local wines they had and got a blank look.

I’ve just spent some time in the hills between eastern Tuscany and north-western Umbria, and I kept running the same small experiment without meaning to. Walk into an ordinary osteria in a town nobody outside the province has heard of. Ask for the house wine. Ask where the oil on the table came from, and the cheese, and the meat in the ragù. The answer, almost every time, is a place you can see from the window.

In 2018, IBM and Maersk launched TradeLens: a blockchain platform meant to digitise global shipping, replacing the paper trail that follows a container from a Vietnamese factory to a Rotterdam warehouse with a single, shared, tamper-proof ledger. It had the backing of two of the largest logistics operators on earth. By 2023, it was dead.

I’ve spent a fair amount of ink on this site comparing LoRaWAN, LoRa mesh, and Wi-Fi HaLow: LPWAN and radio protocols shaped mostly by hobbyists, network operators, and the odd telco. There’s a fourth contender I’ve been putting off writing about, largely because it comes from a different world entirely. MIOTY didn’t grow out of a maker community or an open-source project. It grew out of Fraunhofer IIS research labs and the boardrooms of German industrial automation, and it arrived at the LPWAN table years after LoRaWAN had already set out the cutlery.

In the last post I left a sentence hanging on purpose: durability is a design question, and it matters who’s actually asking it. Here’s the honest version of why, and here’s where this whole run of posts actually lands.

In the last post I ended on a line I’d been putting off: the story has to go somewhere durable enough to survive the person who told it. I meant that more literally than it let on, and at two different scales: one person’s archive, and then something much bigger.

In the last post I named who FediFlick Archive is actually for: people using a photograph to hold onto memory, family, and place, not to make it look better. I ended by promising to explain why the phone, not the desktop, is where this actually starts, and why that’s not the whole story. Here’s that promise kept.