Category: Tech

Eighteen months ago the wow was a picture. Type a sentence, watch software paint you an image that would have taken a decent illustrator an afternoon, and the whole internet performed the same double-take at the same time. That was the trick. Impressive, a little uncanny, mostly a party favour.

That is not what is happening now. Now I hand a piece of work to something that plans it, executes it, checks its own output, and only comes back when there’s a decision that actually needs me. The distance between “AI made a nice picture” and “AI ran a client’s Tuesday” has been covered in about a year and a half, and most small business owners I talk to are still catching their breath from the first part.

Which raises the obvious question, because it should: are we actually getting somewhere, or are we just running faster on the same wheel?

I can only answer for my own patch of ground, and there the answer is genuinely yes. I run this as a one-man operation, and on certain categories of work I am moving through five to ten times what I could a couple of years ago. Not by working five to ten times harder — by handing off the parts of the job that were never the valuable part. Running a fix past three plausible causes before I commit to one. Chasing down an obscure error that would once have meant an evening lost to forum posts and stack traces. Querying a client’s database for the answer to a question they asked me on a call twenty minutes earlier, while we’re still on the call.

None of that is the work I’m actually paid for. It’s the toll you pay to get to the work you’re paid for. And that toll has gotten a lot cheaper.

Two examples

A client sells through a vendor catalogue — hundreds of products, updated constantly, and a brand voice that needs to hold across every one of them. The old way of turning that into a week’s worth of social content meant someone sitting down, product by product, writing copy and pulling images and matching a style guide by hand. I built a workflow that takes the catalogue feed straight through to a hundred branded, on-voice posts a week, ready to schedule. That’s not a person’s job anymore. It was never a job anyone enjoyed doing — it was a job someone tolerated because the alternative was an empty content calendar.

Second: email triage. Every small business owner I know loses part of every morning to a spam problem that keeps evolving faster than the tools built to stop it. I set up an agent that reads intent rather than just keywords, and it’s catching what static filters miss — the plausible-looking invoice, the fake vendor follow-up, the thing that would once have needed a human eye and a human minute to delete. That minute, multiplied by every message, every day, adds up to something worth having back.

Where the attention goes

What both of those have in common is that they’re not really about output. Nobody’s business gets better because it posts more, or answers email faster, in the abstract. What they’re actually about is where the owner’s attention goes. Every hour that used to disappear into the catalogue or the inbox is an hour that can now go toward the customer relationship, the pricing decision, the thing only the owner can actually judge. That’s the real trade being made, and it’s a good one when it happens on purpose.

When it happens on purpose

When it happens on purpose is the part worth sitting with, though. The same surveys showing most small businesses reporting real productivity gains this year also show a third of the tools they buy sitting unused within three months. Adopted out of nerves, not need. Kept because cancelling feels like admitting defeat. That’s the wheel — not AI itself, but the reflex to reach for a new tool every time a headline suggests you’re falling behind, without ever asking what specific problem you were trying to solve when you picked it up.

I don’t think the honest answer is that we’re all hamsters, or that we’ve all found the accelerant for real growth. I think both are true depending on which drawer you’re standing in front of. The catalogue workflow and the triage agent earned their place because I could name the exact hour they were going to save before I built them. Most of what causes the fatigue in this conversation is the stuff that skipped that step.

The test

So the test I keep coming back to, for my own work and for clients, is embarrassingly simple: can you say what this is replacing, specifically, before you turn it on? If yes, it’s probably not the wheel. If you’re not sure — that’s worth noticing too.

There is a particular kind of excitement that comes from installing something new — a fresh tool, a rewritten app, a framework that promises to change how you work. I have felt it more times than I can count. And I have paid for it more times than I care to admit.

The pattern is always the same. The new thing is shinier, faster on the demo, and full of features I will never use. Six months later it is abandoned, broken, or both, and I am back to the dull, dependable program that was sitting there the whole time.

What “boring” actually means

Boring software is not bad software. It is software that has stopped surprising you. The keyboard shortcuts are in your muscle memory. The file format has not changed since 2009. When you open it, it does what it did yesterday. That predictability is a feature, not a lack of one.

The boring tool is the one that is still maintained by someone who cares, but who is not trying to reinvent the category. It gets out of your way. You stop thinking about the tool and start thinking about the work.

The hidden tax of the shiny

Every switch has a cost that the launch video never shows. There is the migration, the re-learning, the plugin you have to rebuild, and the moment three weeks in when you discover the one thing the old tool did that this one simply will not. None of that shows up in the “Why we switched” blog post.

The most expensive software in my life has never been the expensive software. It has been the free, exciting tool I adopted, rebuilt my workflow around, and then watched rot.

When boring is the right call

If a tool touches work you cannot afford to lose — your writing, your client records, your financials — boring wins. Stability beats novelty the moment the data matters. A plain-text notes file from 2014 still opens. The proprietary app that stored everything in its own cloud may not.

Boring also scales across time. The format that survives a decade is the one you will never have to migrate again. That is a gift to your future self.

When the shiny is worth it

I am not anti-new. The exciting tool earns its place when it does something the boring one genuinely cannot — a real capability, not a prettier button. Try it on the side. Keep the old one running until the new one has earned the promotion. Promote rarely.

The quiet advantage

Most people chase the next release. If you instead pick tools that will still be here in five years, you inherit a quiet advantage: you spend your attention on the work, not on the tools. That is the whole case for boring software. It is not glamorous. That is the point.