The 40 days of Lent are typically associated with the days Jesus spent fasting in the desert, resisting the temptation of Satan. It has become a religious tradition in Christian religions to also fast from Ash Wednesday to Easter.
I think it’s unfair to compare the Devil to AI, but I digress.
However, to keep in tune with the story of Lent, starting this Ash Wednesday (18th Feb 2026, coincidentally my 30th birthday), I will keep a 40 day AI Lent. No AI help, no Claude Code, Copilot, Google AI search.
Why? I miss coding, manually. It’s that simple.
I want to spend some time going back to one of the activities that brings me the most joy in my life, being frustrated at myself for writing bad code that doesn’t work, and eventually reaching a state of flow attempting to untangle the cause of my frustration.
I’m also a bit concerned that the engineering know-how I’ve built over the past 12+ years will atrophy and become outdated1. I think these skills are what allow me to move so swiftly and correctly with AI coding agents. I want to slow down, forego false productivity gains through managing fleets of AI agents2, and Sharpen the Saw.
Meanwhile, I will track the following:
- My satisfaction with software engineering (which has been slumping with AI)
- Mental fatigue/cognitive load (daily journal)
- Number of PRs merged
- Time to complete defined features
- Confidence in released features
I will try to update this blog to track these metrics over time, but the qualitative review of my experience will be available in this ai-lent series.
This post was originally a Mastodon toot
Anthropic’s own research shows coding and debugging skills degrade when engineers are introduced to AI-assisted code. - How AI Impacts Skill Formation - Anthropic ↩︎
Engineers routinely, and wildly overestimate how productive AI coding tools make them, sometimes taking 19% longer to solve tasks than when manually coding. Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity - METR ↩︎