Keep Up, Don't Rush

There's a particular anxiety that comes with working in AI right now, and it has a rhythm. Every morning there's a new model, a new tool, a new benchmark, a new thread explaining why the thing you learned last month is already obsolete and you, specifically, are falling behind. The firehose never slows, and it never will. The anxiety it produces is real — it's fueling real burnout across the industry — and most of the advice about it is useless.

The confusion at the root of it is that we treat two very different things as the same: keeping up and rushing. Keeping up is staying oriented — knowing roughly where things are and where they're going. Rushing is reacting to everything the instant it appears, adopting each new tool the day it ships, trying to drink the whole firehose. Rushing feels like keeping up. It's actually just thrash, and it's exhausting in a way that orientation never is.

Jeff Bezos has a line I keep coming back to: stress doesn't come from having a lot to do, it comes from not having taken action on something you can control. That reframes the whole problem. You cannot control the pace of the industry. You can't read every paper or try every model, and no amount of anxiety will change that. What you can control is whether you did the one thing that actually mattered to you today. The dread lifts not when the firehose stops — it won't — but the moment you act on your own small slice of it.

Bezos has another idea that fits here: most decisions are two-way doors. You can walk through, look around, and walk back out if you don't like it. Trying a new tool is almost always a two-way door. Treat it like one and you can move fast without stress, because nothing is at stake that you can't reverse. Rushing is what happens when you treat every door as if it were one-way — when adopting a library feels as weighty as betting the company. Save the real deliberation for the few decisions that genuinely are irreversible, and be quick and cheap about everything else.

So keeping up looks calmer than people expect. Pick a couple of sources you trust and ignore the rest without guilt. Go deep on the few things that are yours to own, and let the rest wash past. Act on what you control, treat most choices as reversible, and measure a day by whether you moved your own work — not by whether you kept pace with a timeline nobody can keep pace with.

You will never consume the whole firehose, and you were never meant to. Keep up by acting on what's yours. The rest is noise wearing an alarm. Don't rush — just don't stop.