There's a quiet shift happening that big vendors won't put in a press release: the free, open models you can run on your own machine are now good enough for a lot of real small-org work. Not everything. But more than you'd guess.
What's happening
A year ago, "local AI" meant slow, clunky, and noticeably dumber. That gap closed fast. Independent developers and open-source communities, not Big Tech marketing departments, pushed models that run offline on a decent laptop (via free tools like Ollama) and handle summarising, drafting, and data-cleanup well. For a nonprofit or a 10-person shop, that changes the math.
Why it matters for you
- Privacy without a lawyer. Client intake notes, donor lists, health-adjacent info: running the model locally means the data never leaves your building. That's a real answer to "can we even use AI with this?"
- No per-seat bill that grows with you. Grassroots orgs get punished by usage-based pricing. Local models flip that: one-time setup, no meter running.
- It works when the wifi doesn't. Community centres, field sites, church basements: offline is a feature, not a compromise.
Try this this week
Ask someone technical (or us) to install one local model and point it at one boring task, like turning meeting notes into action items. Don't boil the ocean. Prove it on the dullest job you have, then decide.
The bigger picture
The power dynamic is shifting toward small, nimble teams, the ones willing to try the un-hyped tool. You don't need a vendor's permission to start. You need one afternoon and the willingness to run something yourself.