This is a vision of a future that could be - one among many.

August 1, 2040

I wake up from a good night's sleep to a gentle alarm from my pocket computer sitting across the room. Yawning, I get out of bed and turn off the alarm.

I'm a programmer, fresh out of college. There was a time when some said we wouldn't need programmers anymore and that AI would do all the work. That turned out not to work quite so well - those companies that went all-in and fired their development teams found they had an unmaintainable mess on their hands a few years later. AI's still around in places, and some people do use it and like it, but it's far from universal.

The AI datacenters aren't around anymore, though. Casualties of the AI crash.

But the RAM crisis of the mid-2020s never quite ended. After years of super high demand from the AI companies for memory and initial hesitation, the few companies manufacturing DRAM finally gave in and ramped up production... just in time for the bubble to burst. Nearly all of them went bust, and RAM these days is especially hard to come by.

I get dressed, make sure I have my pocket computer, wallet, and keys, and head to the local coworking space.

When I get there, I sit down at one of the displays and plug my computer in. Years ago, most developers would have looked at this setup and found it unbearable. Surely that tiny computer can't be enough for programming, they'd say? But the RAM crisis would force something of a mindset change: we'd need to do more with less.

See, in the mid-2020s, the amount of RAM and storage in smartphones was already starting to approach that in laptops. And with the price of RAM on the rise, more and more people had to make the tradeoff: between only having a smartphone or only having a laptop, which would you choose? Most people, especially at the time, were going to pick the smartphone.

So a lot of people started finding ways to get their work done with only a smartphone, and it turned out to work better than a lot of people expected. There had been a small cadre of (perhaps somewhat masochistic) folks trying to make this work for years, and the result was that by and large, the software was mostly ready by the time it started to matter. Between spending $1500 on a laptop that would have been $500 just a few years before, or plugging your phone into a monitor with a $15 dongle, for most people it was an easy choice.

By 2027, Google had already put a desktop mode on most Android phones that could run Linux apps through a software update, and mobile Linux distros (which already had useful desktops) were starting to be capable phones on more and more hardware. Apple was the holdout, but it only took them a few more years to make iPhones act as Macs when plugged in after they realized people weren't likely to spend $1500 on an entry-level MacBook with an iPhone chip after having just spent $1500 on their iPhone.

It didn't take long for people to start thinking of their smartphones as just "computers," and calling them that.

It was true, though, that pocket computers didn't quite have the power for every task. Programmers using interpreted languages were often able to work locally without problems, but anything with long compile times got increasingly painful. So companies started setting up build servers for their employees, and hackerspaces got in the game shortly after for their members. Sharing memory and compute between multiple people turned out to be more economical than giving everyone a ton of RAM locally.

This did take a lot of adjustment on the part of build tools. Offloading builds to more powerful machines was something that had always been possible (distcc had been around for decades after all), but it wasn't common.

When I finish up work for the day, I ride my bike across town to the local library. I've got some writing I'd like to do, and I've been feeling a bit cooped up at home lately.

The library has a bunch of displays set up on desks as well, but I've just been working at a desk all day and I'd rather chill in a window nook, so I check out a lapdock from the front desk. The library bought these things about a decade ago (pretty cheaply since they don't have RAM or storage), and other than needing a few battery replacements over the years they still work just fine.

Once I'm done writing, I hit a switch in the corner to connect up to the social net. My computer starts syncing with my relay at home, and I start catching up on what my friends have posted lately.

Years ago, we were mostly using social networks hosted in datacenters somewhere. But when fascism was on the rise in the mid-2020s, governments started putting pressure on the big social networks to control the narrative. X was captured from the start of course, but Meta fell in line pretty quickly. Bluesky held out for a while, but their centralized AppView was a pretty big target, and eventually they capitulated. The few other providers that could afford to run alternative AppViews fell under similar pressure.

The Fediverse made for a harder network to control, without having just a few large companies to put pressure on. A lot of people moved there, and many of them stuck with it. But in part because of the RAM shortage and increasing hosting costs, a lot of small server admins found it hard to keep up with.

Thankfully, though, development on peer-to-peer social networking tools like Spritely's has progressed to the point where running a social network without servers in a datacenter is more and more practical. Instead of living on one server, your profile lives on your local device. You still need a relay somewhere with a more reliable network connection and power source, but it can live just about anywhere, can't see your messages, and you can switch relays whenever you want.

Some people use relays run by one of their neighbors or their local library, and some people run their own relays at home. I have my own, running on my old pocket computer. The last time I upgraded, I took the battery out of my old one and set it up as a relay. It's usually pretty easy to set one up these days if you have a device to put it on, it's pretty much like installing any other app. Then you plug it in, scan a QR code, and you're good to go.

Some devices are more locked down, but on most there are ways to unlock them, and there's a pretty good network of folks helping others out with that.

That weekend, I head over to my local hackerspace to help some folks repurpose their old pocket computers. Manufacturers got better about repairability after a while (nobody wants to be told they need to buy a new expensive device when it just needs a new battery) but a lot of earlier models with non-removable batteries are still floating around. You don't necessarily want to keep those plugged in 24/7 because of the fire risk, but that's what you need to do to use it as a relay. So those of us with the repair skills to do so have been volunteering to take those older devices apart and get them running without their batteries. Sometimes they can already run without a battery, and sometimes they need a dummy battery of sorts installed, but we've been doing this long enough that we've pretty much gotten it down to a science.

When I get home, I pop a controller on the back of my pocket computer and play some video games. I've been mostly playing retro games lately, but most contemporary games run fine on this thing too. Game graphics kinda leveled out around the PS5 era, since a lot of people couldn't afford new dedicated game consoles after that. But it didn't take too long for phones to catch up, and as with desktops, nowadays most people have one device to do it all.

I certainly don't expect everything in this little vignette to happen. Maybe RAM prices do actually recover in a few years (actually probably pretty likely). Maybe AI turns out not to be a bubble and it's just ever-present from now on. Maybe we do all lose our jobs in a few years. Who knows.

Maybe phones get so locked down that you can't run unapproved software on anything but a Linux phone, and manufacturers manage to close all workarounds. Likely Apple will never actually open up a desktop mode on their phones.

Maybe, hopefully, we successfully push back fascism, and hold on to democracy.

My goal in this piece was to explore a particular "what if" - what if some of the problems we're facing right now get worse from here? What if computers stay expensive? What if we need alternative social networks, because the ones we have today get captured?

And how can we collectively try to make the most of that possible future?