How Netflix got its creatives to trust AI

For this week, what we're seeing is that AI changes our people's work significantly before it changes our companies' numbers. Here's a look at our favourite posts from the week:

Let your sceptics argue it out in front of the room

First Round Review's account of Netflix's AI overhaul follows Mckenzie Lock, who ran product, operations and creative there and had thousands of sceptical creatives to win over.

For each proposal she picked two people with opposing views and had them argue both cases in front of the room. Every strategic bet was written as a choice between two attractive options: lower-stakes titles before a release the size of Stranger Things, and proof of member impact before scale.

The piece gives no figure for what members gained, but the method still travels. Give your largest AI proposal's most credible sceptic the floor before you approve it, and write down the option you turned down, and why.

Watch: Molly Graham on grieving the work you are losing

The grief, loneliness, and burnout sweeping through the tech industry right now is the focus of Lenny Rachitsky's conversation with Molly Graham, who wrote the "give away your Legos" advice a generation of managers learned to delegate by.

Some of the work you were great at is going away. It's just the way the world is moving. Graham's advice is to grieve it properly, and stay open-minded and curious about what replaces it. If we take the right attitude, we can stay on top of the shifting landscape of how we do work.

Also worth your time

AI doesn't create leadership problems - it exposes them, Kate Leto. People copy what leaders visibly do with AI, so your own curiosity sets the standard for everyone below you. Leto also warns that automation often moves the work, instead of removing the work. As leaders, we'll need to ensure our team's psychological safety as they experiment with new technologies, making it okay to learn, and sometimes to fail.

Standards Create Trust. Trust Creates Scale, Cook's PlayBooks. Jim Cook, the former CFO of Mozilla and Netflix, argues that people resist process and accept standards they helped to write. He wants the CFO to agree decision, data and AI standards with each executive, before individual habits and group dynamics take over and set the standards. He also asks a great question, "Which parts of our business or your team still depend on heroics rather than standards?"

Who Gets Paid When AI Does the Shopping?, a16z. If assistants become where people decide what to buy, marketplaces lose the discovery and advertising they earn most from. Great reading when your CMO still needs a position on what you insist on keeping direct.

The state of the tech industry in 2026, The Pragmatic Engineer. At the companies the author visited, software engineers have mostly stopped writing code. Now instead they run five to ten AI agents at once. Nobody can review that much code, so it is often approved unread. Which means that now the software that runs your business is shipping without anyone having read it. It's a new way of delivering software that makes all of us leaders squirm in our seats.

Copilot Gets a Seat in the Org Chart, Every. Microsoft's always-on agents now appear in the org chart, reporting to a named owner. The useful part is the practice Microsoft's Charles Lamanna describes: measure how well people do the task before an agent tries it. Fund no agent until you know the human error rate it has to beat.

Is your leadership team ready to light the fire?

Upskill with AI Firestarter bootcamps and workshops, built for leaders who are too busy leading to be experimenting with AI.

As a leader, you set the tone for how AI is used in your organisation.

Until next time,
Ryan Snyder
ryan@aifirestarter.com