The cost of living in the future

I just finished reading Mitchell Hashimoto’s piece about his AI adoption journey. It got me thinking about my own experience over the last couple years: climbing the steep adoption curves of several AI products.

As Mitchell observed, adopting a new technology usually starts with a period of inefficiency. That’s when you force yourself to abandon a workflow you're familiar with (eg. asking code questions in chatGPT) for one you're not (eg. telling OpenCode what to do). First, your productivity dips; then eventually rebounds to a higher baseline.

I lived a version of this recently with OpenClaw. I spent two nights last week struggling to run it locally. I had several issues with plugins, the memory system, the TUI. (Their rebranding caused problems with the configuration folders too, which seem to have been patched.) I gave up on OpenClaw a couple times, but eventually came back and got it running.

This is part of the bargain that early adopters of technology have always made. We put up with rough edges and high cost in exchange for a glimpse of the future.

There's a certain clip of Steve Jobs that I keep coming back to. During WWDC 1997, shortly after the NeXT acquisition, Jobs did a long Q&A where he described his computer setup at home:

In the last seven years, do you know how many times I've lost personal data? Zero. Do you know how many times I've backed up my computer? Zero. I have computers at Apple, at NeXT, at Pixar, and at home. I walk up to any of them, and log in as myself; it goes over the network, finds my home directory on the server, and I got my stuff wherever I am.

And one of the things I’m excited about is to look at that personal computer and take out every moving part except the keyboard and the mouse. I don’t need a hard disk in my computer, if I can get to the server faster.

Managing a network like this is a pain in the butt. Setting it up, getting it all to work, is really complicated.

One of my hopes is […] that Apple can make that as plug and play for mere mortals as it made the user experience over a decade ago. It’s one of the things where I think there’s a giant hole. And I can’t communicate to you how awesome this is unless you use it. And what you would decide within a day or two is that carrying around these non-connected computers, or computers with a ton of state in them is byzantine.

An audience in 2026 would struggle to understand the battle he was fighting. File syncing is boring now; it's largely a solved problem. I don’t remember the last time I worried about backups, or having to access a work file from another device, or losing data because because of a bug in Google Drive.

Today we live in a cloud-first world, where data gets synced and backed up by default. This world was invented by people like Jobs — and surely many, many others — who paid the price of early adoption back in the 80s and 90s.

Jobs’ experience with the cloud at NeXT went into MobileMe and later iCloud, where I keep my files today.

The future is designed by those who get to live it first. The cloud of today was designed by people who got to use it ahead of everyone else. The same will be true for agents, whatever they look like in ten or twenty years.

And yes, keeping up with the bleeding edge of AI takes a lot of time; but that's the price you pay for a chance to shape it.