At some point last year, agentic coding became the default in my workflow. This causes a lot of FUD for some developers. Will I forget how to code? When I learn new things, will I actually understand how they work?

Honestly, I’m not sure of the answers to these questions. But the productivity gains are too incredible to ignore. I can kick off brand new projects and make sizeable progress in minutes; I can finish projects in hours. I can throw OpenCode (yep, I migrated from Claude Code in December) at a problem and Opus 4.5 will keep it cruising for 10-20 minutes without me. I can run multiple OpenCode instances for multiple projects at once, and jump around doing some light context switching to ship at a pace that is probably 100x what I could have done before.

I don’t really know what that means for the future of software development. Things are clearly changing, and if you aren’t interested or trying these tools, I think you’re going to be left behind. Will my skills - the ones I spent over a decade building from scratch - become rusty? Probably. But does that matter when AI is consuming software, creating software, and becoming the way our tools interact with each other? Maybe not.