Our industry is in love with numbers. Higher sample rates, lower latency, the newest protocol — the bigger spec sheet wins the argument. But thirty years taught me something simple: the show doesn't care about your spec sheet. In all that time, an audience has never once clapped for 96kHz. They care about one thing — that the sound never stops. So I design for that, and it means making some unfashionable choices on purpose. I'll take a 48kHz system that runs twenty-four hours without a single glitch over a 96kHz one that doubles my network load and stumbles when the building gets noisy. The number that matters isn't the sample rate. It's the uptime. I slave every converter to one master clock. Two timing protocols fighting over who keeps the beat is how you get the little ticks and dropouts that take hours to chase. One clock, one boss — that's not bureaucracy, it's silence where there should be silence. So I leave myself margin. I don't run a network at the ragged edge of its capacity just to win a spec argument — I leave headroom, so when the room gets noisy or a cable goes flaky, the system has somewhere to give instead of falling over. Robustness isn't something you bolt on at the end. It's a choice you make at the start. None of this is me being against technology. I love the tools. It's me being against fragility. The best system isn't the one that benchmarks highest in a quiet shop. It's the one still standing at the end of a bad night.
Don't buy the spec sheet. Buy the night it doesn't fail. What do you reach for first — the numbers, or the uptime?