The full version of the old line is "jack of all trades, master of none, but oftentimes better than a master of one" — and the truncated version everyone actually quotes tends to get used as an insult. Twenty-plus years into independent IT work, it reads more like a fair description of how the useful problems actually get solved.
Independent practice doesn't hand you a narrow lane the way a large organisation does. One week it's a firewall config that's quietly letting the wrong traffic through, the next it's a compliance team trying to make sense of a data set nobody structured properly, the next it's a piece of tooling that needs to exist and doesn't yet. Specialising early would have meant saying no to most of that — and no to the pattern that turned out to matter most: almost none of it is really a different problem.
The one deep thread
Underneath firewall logs, compliance records, and communication histories sits the same actual work: finding the signal in something dense, messy, and not built to be read at a glance. That's the specialism — not a technology stack or a single industry, but a way of looking at data that generalises across all of them. IT security just happens to be the domain where that instinct got sharpest first, because security work punishes sloppy pattern-reading faster than almost anything else does.
Breadth, in that light, isn't the opposite of depth — it's what let the depth show up in the first place. Working across enough different messy datasets is what makes it obvious which parts of "finding the signal" are domain-specific and which parts are the same problem wearing a different hat. You don't see that from inside one lane.
Why it turned out to matter
It's the reason the current product line looks the way it does — a security tool, a compliance platform, a case management system, and a communication-analysis tool, built by the same practice instead of four separate specialists. Not because any one of them is easy, but because they're the same underlying skill applied to four different piles of dense data. Master of none, maybe. But the breadth is exactly what makes the pattern recognisable in a new pile before anyone else has named it.