I've read less on X and LinkedIn lately.
This might sound self-satisfied, but imagine if you'd spent 20 years writing C# and suddenly your entire feed became posts about how a C# function should look or people sharing C# code examples. That's where I am now with AI development.
It's not that the posts are bad. It's that there's suddenly so much volume, and I already know the fundamentals. What I don't know is what I should build next. And for that, I need to build, not read.
In the last three months, I can see the acceleration clearly. Everyone is getting on the wagon. Many are sharing their discoveries. But the thing is — what took me three months to understand and integrate, people are getting now in a day with a little guidance.
The skills depreciate quickly. The tools improve constantly. Reading about what worked last month is less useful than experimenting with what works now.
What's interesting is watching the magic moments happen in real time. My colleagues are reaching those inflection points. They're starting to use the tools actively. The individual productivity gains are crystallizing across the team.
So I'm reading less. I'm building more. Not because I've figured it all out, but because the frontier moves so fast that the only way to stay relevant is to keep shipping.
Part of the #100DaysToOffload documenting agentic development in 2026