Writing

Learning in the open

A running LinkedIn series on what AI-assisted development does to leadership, organizational learning, and engineering culture. Written as I research, including the half-formed parts. Each post lands here as it goes live.

4 / 26 published so far

The series

  1. IntroductionWeek 1

    Intro: why I'm sharing this

    Watching agents do the work of months in mere days was an exhilarating and scary experience. The scary part was not the speed, but the consequence for the developer, the team and the organization. After 20-odd years as a software developer and leading

    Read on LinkedIn
  2. Org learningWeek 2

    Cheap code, expensive decisions.

    Will we soon see companies multiply their output, now that code has become cheap and developers can write it far faster than before? Development departments have traditionally felt like the bottleneck in modern technology companies. My department at

    Read on LinkedIn
  3. Engineering cultureWeek 3

    The 45% problem (insecure AI code).

    How secure is AI-generated code, really? In my most recent project with agentic development I had not specifically focused on that aspect, and a study shows why that is a mistake. Like every other foundational part of a technology product, security has

    Read on LinkedIn
  4. Engineering cultureWeek 4

    Bot-sitting: the new day job.

    You sit there waiting for the agents to come back with their results. You use the wait to kick off new things. Now you're not running one track, but four. Bot-sitting has become the new way to work, and done right you're very productive. Done wrong

    Read on LinkedIn
  5. Engineering cultureWeek 5

    Raise the floor — but look at what the floor is made of.

    Coming soon
  6. Engineering cultureWeek 6

    The escalator (AI and the lazy default).

    Coming soon
  7. Engineering cultureWeek 7

    It has my name on it, but I don't understand it.

    Coming soon
  8. Org learningWeek 8

    What's stable when everything moves?

    Coming soon
  9. Engineering cultureWeek 9

    The best at AI aren't the best prompters.

    Coming soon
  10. Engineering cultureWeek 10

    What AI quietly broke: mentorship.

    Coming soon
  11. Engineering cultureWeek 11

    When code is free, craft becomes the expensive part.

    Coming soon
  12. Org learningWeek 12

    Cheap code makes good process matter more, not less.

    Coming soon
  13. Engineering cultureWeek 13

    Half the work was cleaning up after myself

    Coming soon
  14. Engineering cultureWeek 14

    Fixing code in a loop makes it worse.

    Coming soon
  15. Org learningWeek 15

    Tell your AI what you've decided.

    Coming soon
  16. Org learningWeek 16

    Small PRs, big reviews.

    Coming soon
  17. Org learningWeek 17

    Build the toolchain team, not just the tool.

    Coming soon
  18. LeadershipWeek 18

    Don't ask what AI saves. Ask what it opens.

    Coming soon
  19. LeadershipWeek 19

    'Buy and hope' is not a strategy.

    Coming soon
  20. LeadershipWeek 20

    Who actually makes AI work?

    Coming soon
  21. LeadershipWeek 21

    Do agents replace middle management?

    Coming soon
  22. LeadershipWeek 22

    Going headless.

    Coming soon
  23. LeadershipWeek 23

    Governing the harness.

    Coming soon
  24. LeadershipWeek 24

    What the patent landscape tells us.

    Coming soon
  25. LeadershipWeek 25

    How do you measure success with AI development?

    Coming soon
  26. LeadershipWeek 26

    Three levels: strategy, tactics, operations.

    Coming soon

Follow along

New posts go up most weeks, published in the open as the research moves. Read the research they draw on, or get in touch to compare notes.

Read the research