External signal·Exponential View·Jan 10, 2026·Azeem Azhar·8 min read
The work after work
“Both use AI, but only one has a human signature.”
Summary
Azhar revisits his nine-year-old "artisanal cheese" thesis — that as machines master the specifiable, value migrates to what resists specification — and argues AI has now collapsed the cost of execution in knowledge work. He describes building three custom apps for his DJ workflow in about an hour for a couple of dollars, clearing a backlog he would never have paid a dev shop to build, and points to the Claude Code lead developer whose December code contributions were 100% AI-written, BCG's 36,000 custom GPTs, and Stack Overflow's decline as people stop asking questions the model answers in context. His central distinction is a spectrum of use: at one end "prompt and publish," where the model's output is treated as finished; at the other, treating output as a draft reshaped by human judgment and experience. With pseudo-authenticity now cheap to manufacture — AI can fake texture and idiosyncrasy — the essay asks what still makes a real human contribution legible.
Predictions for the future of work
Azhar's framing for the future of work is that execution becomes near-free and value concentrates in judgment, taste, orchestration and authenticity — an "artisan premium." He reports personal velocity equivalent to "50, maybe 100 people" on his team, implying small high-leverage operators displacing larger ones, and argues the engineer's "translator" role of converting real-world needs into code is becoming obsolete as people communicate through working prototypes. Near-term, advantage shifts to those who can direct and edit AI rather than those who merely generate raw output.
Originally published by Exponential View · Jan 10, 2026
Read the original at Exponential View