A realistic timeline for what AI can (and can't) do this year

Every January someone predicts AI will replace some entire profession by December. Every December that profession is still there, slightly changed. I've stopped taking the big predictions seriously and started tracking the small, boring shifts instead.

This year the boring shift was AI quietly showing up inside tools people already use — the spreadsheet, the email client, the search bar — instead of new standalone apps nobody opens twice.

My bet for what actually happens next: less "AI replaces X," more "AI becomes an invisible feature of X that you stop noticing within a month." Unglamorous, but that's usually what real adoption looks like.

Part of the deeper dive: The Practical Guide to Using AI Tools Without Getting Burned.