The more I talk and write about this, the more I am convinced that we must use AI as an opportunity to do things differently, not simply to perpetuate more of the same but faster and harder. As always, see below for the summary if you prefer to read than wathch.
Every conversation about AI centres on the same promise: do more, move faster, scale up. But what if that’s the wrong question entirely? In this solo reflection, the case is made for a counterintuitive idea — that slowing down, stepping back, and doing less might be how we actually get more out of AI, and out of ourselves.
AI Summary
The pull of AI is almost always toward more — more output, more speed, more scale. But there’s a growing tension between that pressure and what actually sustains high performance over time. This talk explores why “doing less with AI” is not a cop-out, but a strategic and deeply human response to the moment we’re in. The argument moves through several layers: the collapse of the traditional junior engineer learning path, the intensification of work that AI research is already documenting, the Red Queen trap of competing on volume, and ultimately why human vision, values, and intuition are the real differentiators between two equally capable AI systems. The closing idea is both practical and philosophical — a dinghy pointed in the right direction will always outrun a speedboat going the wrong way.
Chapters
(00:00:00) - Introduction: The Pull Toward More
(00:01:08) - Rethinking the “Army of Interns” Metaphor
(00:02:34) - What We Lose When AI Skips the Drudgery
(00:03:46) - Junior Engineers and the Forced Jump Up the Stack
(00:06:20) - AI Intensifies Work — and That’s Not Entirely Bad
(00:09:14) - The Paradox: AI Lets You Do More, So Why Not Do Less?
(00:10:26) - The Red Queen Trap and the Race to the Bottom
(00:14:40) - Vision, Vibes, and What Only Humans Can Contribute
(00:16:44) - Where Vision Comes From: Embodiment, Rest, and Inner Life
(00:19:27) - The Ergodicity Argument: Why the Best Skiers Aren’t the Fastest
(00:24:11) - Audience Q&A: AI as Mirror and the Question of Consciousness
(00:29:09) - Value-Based Work and Doing Less to Achieve More
Key Moments
The intern metaphor we’ve never questioned (00:01:08) — “This idea that, you know, they’re essentially just slaves for us, right — we can get them to get our coffees or do all our boring tasks.”
Junior engineers skipping the queue (00:04:21) — “One of the junior engineers I’m working with at the moment is having to make trade-off decisions and think about things — the conversations we have are at a level I would have previously been having with at least a mid-level, if not a senior engineer.”
The Red Queen trap (00:12:09) — “You have to work harder and harder — it’s called the Red Queen paradox, right? You have to run harder and harder just to stay still.”
The dinghy vs. the speedboat (00:23:25) — “A dinghy going in the right direction will always beat a speedboat going in the wrong direction. You might have a speedboat with AI, but if you’re not pointing in the right direction, what’s the point — you’re just going the wrong place way faster.”
AI as the ultimate mirror (00:25:00) — “AI is like the ultimate mirror — this is ultimate in reflection. And I think these are ultimate mirrors, so it’s incredibly difficult not to anthropomorphise them and to read things into them that we want to see.”
The key message (00:30:36) — “If we can step back and do less, I think we’ll actually achieve far more.”
Sharing is caring!
If this conversation resonated, please consider sharing it with two friends who might find it useful.
