AI
Local AI tools are getting good enough for everyday notes
Smaller models on laptops are no longer toys — if you pick the right jobs.
The Story
Running models offline used to mean research hobbyism and loud fans. Newer quantized builds on modern laptops handle summarization, tagging, and light rewrite help without a cloud round trip.
They still choke on deep research synthesis and multi-step reasoning. Pick jobs that fit the model size and you get privacy, lower cost, and snappier loops.
Why It Matters
Not every note needs a frontier model. Meeting cleanup, bullet extraction, and "make this less corporate" are perfect local tasks. Sending every scratchpad to a hosted API is a habit, not a requirement.
Desk Take
Start local for note cleanup and meeting summaries. Keep research synthesis and coding agents online until your machine proves it can keep up. Measure wall-clock time once — feelings about "local AI" are usually wrong.