A project to build with System One models

Inbox triage for a small business, on a laptop CPU

Sort incoming email into invoices, customer questions, sales enquiries and noise with a model small enough to run without a GPU.

Tools and appsBeginner4 stepsSuggested by @biplov

The decision

A choice over folders (invoice or receipt, customer question, sales enquiry, newsletter, spam), plus a noul question: "needs a reply today".

Read new mail over IMAP with a short Python script.
Pull a small model. Julia 1 is 144M parameters and runs on a CPU: systemone pull supersonic-labs/julia-1.
Classify each message, move it to its folder, and flag the ones needing a reply today.
Run it every few minutes with cron.

Keep a "moved by hand" log: every message you drag to another folder is a label. Fine-tune on them once you have a couple of hundred.

Tools and appsAdvanced

Decide, for each field on a form, whether to type a value, tick a box, click or skip, using a model small enough to run in the loop.

Tools and appsIntermediate

Take a labelled dataset from your own work, fine-tune Laya on it locally, prove it beats the base, and publish it on the registry.

Tools and appsIntermediate

Score every pull request for risk from its diff and description, and ask for a senior reviewer when it touches something dangerous.