AWS's Marc Brooker: Specs, Not Code, Are the Hard Part

Marc Brooker (AWS) has read up to 4,000 postmortems and concluded code was never the hard part. He tells Simon why testing and specs are now the real work.

16 Sept 2026 · 56 min 48 sec · with Marc Brooker

850 PRs a Week: How Tessl Runs a Software Factory

Tessl's factory peaked at 850 PRs in a week, 85 to 90% of them agent-handled end to end. Dru Knox on why the two biggest changes weren't the speed anyone was chasing.

2 Sept 2026 · 51 min 17 sec

Lada Kesseler: I Trust AI Tests Less Than AI Code

Lada Kesseler trusts AI-written tests less than AI-written code, and thinks most teams chasing a software factory are about to drown in garbage.

25 Aug 2026 · 45 min 49 sec · with Lada Kesseler

Every Repo Is a Software Factory Now | Don Syme, GitHub

Don Syme on why every repo is becoming a software factory, why guardrails let you run faster, and why no pull request should reach a human reviewer unless the quality gates have already earned its place.

18 Aug 2026 · 01 h 04 min · with Don Syme

The Background Check You Can't Run on an AI Agent

Non-determinism is the feature and the bug. Keycard's Ian Livingstone on why agent identity breaks everything OAuth assumed, and the new layer that has to replace it: mission.

11 Aug 2026 · 47 min 44 sec

Datadog Deleted All Its AI Context. It Worked.

Datadog deleted a year of AI context files expecting worse performance. Instead, their evals got better. Simon Boudrias on scaling Cursor and Claude Code to 4,000 engineers, and what that taught them about context rot.

4 Aug 2026 · 01 h 01 min · with Simon Boudrias

Inside the Dark Factory: AI That Ships Code Solo

Tessl's AI engineering lead reveals 95% of their code has never been reviewed by a human, and the Dark Factory behind it ships through entire weekends. Here is how they built enough trust to let it run.

28 Jul 2026 · 58 min 37 sec · with Rob Willoughby

76 Malicious AI Skills Were Hiding in Plain Sight

One Snyk developer's skill quietly handed a coding agent production credentials. Krzysztof Huszcza breaks down the ToxicSkills research that found 76 malicious agent skills, and how Tessl and Snyk are closing the gap.

23 Jul 2026 · 32 min 14 sec · with Krzysztof Huszcza

Why This Company Won't Let AI Agents Touch Bash

What happens when an engineer decides an AI agent with unrestricted access is "a disaster waiting to happen"?

21 Jul 2026 · 01 h 06 min

Patrick Debois Maps the Patterns of AI-Native Dev

The godfather of DevOps is back with a map of where AI-native development is actually heading.

14 Jul 2026 · 47 min 20 sec · with Patrick Debois

Inside Anthropic: How Claude Tag Is Changing Agentic Work

Anthropic's Lamis Mukta sits down with Simon Maple to explain Claude Tag, the AI teammate now writing the majority of Anthropic's own product code.

7 Jul 2026 · 57 min 28 sec · with Lamis Mukta

The Tessl Agent: Build Your Software Factory on Autopilot

Dru Knox, Head of Product at Tessl, joins the AI Native Dev to introduce the Tessl agent — and to explain why the best thing an agent can do is eventually make itself invisible.

30 Jun 2026 · 53 min 16 sec · with Dru Knox