My AI stack hasn't changed in months, and that's exactly why it works.

almost daily, someone asks me "what AI tools should I be using to keep up?"

Wrong question.

The right one is: which tools am I willing to actually get good at.

here are my 2 cents and 5 tools: (the combination below saves me 70-80% of the time)

I switched from typing to speaking almost everything I write, and it's not because Wispr Flow is the smartest tool out there. It's because I stuck with it long enough to stop thinking about it. I talk, a transcript appears, I move on. that one habit change gave me back time I used to lose to typing.

Same story with NotebookLM. Doesn't matter if you're a teacher building a lesson, a founder buried in documents, or a student trying to actually understand something instead of memorizing it - dump your files in and let it organize, summarize, turn them into slides or visuals you can look at instead of re-read. not flashy. Just useful every single day.

For writing and brainstorming, I use Claude. Could've been Gemini, could've been ChatGPT - genuinely doesn't matter which one you pick. What matters is you pick one and stop jumping.

For images, Google Gemini does the job fast enough that I don't think twice.

For code, Claude Code and Cursor, depending what I'm building.

Here's the part nobody wants to hear: the tools were never the bottleneck 🤷‍♂️

The jiggling is.

I watch people re-learn a new tool every couple weeks, chasing whatever had the loudest launch that month. they're not more productive, they're just busier learning. Meanwhile whoever picked five decent tools eight months ago and got fluent in them is shipping circles around them.

Of course this isn't a call to never try anything new.. but a call to stop mistaking exploration for progress.

Pick your five. Get boring with them. Watch what happens to your output.