The AI Tool I Quietly Dropped This Year (and the One I Can't)
I signed up for everything and kept two. The one I dropped wasn't bad, it was just solving a problem I'd already solved, and that's the most expensive kind of subscription.
I went on an AI-tool binge like everyone else and then did the adult thing: looked at what I actually opened. The result was two keepers and a graveyard of monthly charges I'd stopped noticing.
The one I dropped was a fancy writing assistant that rewrote my sentences into something smoother and less mine. It was good. It was also redundant, because the model I already used could do the same when I asked. Paying twice for the same capability is the silent tax of this boom.
The one I can't drop is my transcription-and-summary combo for meetings. It doesn't wow anyone in a demo, but it quietly returns the hours I used to lose rewriting notes. Boring tools that compound beat exciting tools that don't.
The lesson I'm keeping: a tool earns its subscription by removing a task I'd otherwise do, not by impressing me on the first day. If I haven't opened it in a month, the renewal email is a goodbye.
Frequently asked questions:
Q: How many AI tools should one person keep?
A: As many as remove a real task. For most of us that's three or four, not thirty.
Q: Is it wasteful to try many?
A: Trying is fine; forgetting to cancel is the waste.
Q: What makes a tool a keeper?
A: It does a job you'd do badly or not at all, without needing a babysitter.