Tools don’t have taste

A new tool shipped this week. There was one last week too, and another will land before this post cools off. Our stack has changed more in the past two years than in the previous ten, and anyone in this business who claims otherwise is either lying or napping.

We try most of them. That tends to surprise people who expect a craft argument to show up wearing suspenders and mourning the good old days. No mourning here. The new stuff is frequently astonishing, and pretending otherwise is its own kind of posing. Layouts in seconds. Code scaffolded before the coffee finishes. Video that used to require a crew, a permit, and a prayer.

Here’s what none of it ships with: the part where somebody looks at forty plausible options and knows which one is right.

What compresses and what doesn’t

Tools compress execution. They don’t compress judgment. The gap between those two is where this whole profession actually lives, and the gap is getting wider, not narrower, because when producing options becomes nearly free, choosing among them becomes nearly everything. A generator that hands you fifty directions hasn’t done the hard part. It’s moved the hard part, from making to deciding, and deciding was always the expensive bit dressed up as the easy one.

What taste actually is

Taste is the word we use for that deciding, and it’s worth being precise about what it is. Not preference, not aesthetics, not knowing which fonts are fashionable this year. Taste is knowing what to want before you’ve seen it and what to kill after you have. It’s the ability to look at something technically flawless and feel that it’s wrong for this client, this audience, this moment, and then to say why in words that survive a meeting.

The eye that decides what’s good doesn’t come bundled with anything.

Nobody downloads it. It gets built the slow way: from shipped projects and failed ones, from watching real users ignore the thing you were sure they’d love, from a thousand small calls that turned out right or wrong and, crucially, from finding out which. Every veteran you admire is carrying around a private database of consequences, and no vendor sells access to it.

The flood, again

This has all happened before, for whatever comfort that’s worth. Desktop publishing handed everyone two hundred fonts, and for a while documents looked like ransom notes, and good typography got rarer and therefore more valuable. Cheap cameras made everyone a photographer except in the ways that mattered. Each time a tool democratizes execution, the market floods with competent-looking work, the average gets louder, and the difference between competent and good becomes the entire product. We’re in that flood again, just faster and wider.

There’s an uncomfortable corollary, too. Bad taste scales now. A mediocre eye used to be throttled by slow hands; the work took long enough that only so much of it could exist. Give that same eye a modern toolchain and it can produce mediocrity at industrial volume, all of it polished, none of it right. Speed is an amplifier, and amplifiers famously don’t care what they’re amplifying.

So no, we’re not worried about the software, and we’re not precious about it either. When a tool makes execution faster, we take the speed gladly and spend the savings where they’ve always belonged: on the thinking before and the judgment after. What we’d never do is confuse a shorter distance between idea and artifact with a better idea.

Craft has outlived every instrument it’s ever used. The people who made the work you still remember worked in media that barely exist anymore, and the work survived the tools the way a song survives the piano it was written on. Whatever we’re all using two years from now, someone will still have to look at the output and make the only call that ever mattered.

Is it good. The box has never once known.