The benchmark showed tiny local models *can* do real maintainer work. We measured *why* — pre-registered macro ablation, same corpus, temp 0:• MACRO-OFF: model composes the workflow from primitives live• MACRO-ON: workflow pre-encoded; model just dispatchesI(X;Y) raised 1.24–1.62× for every model that routes; 2.0–5.1× info-per-second for the 1.5B/3B/8B. Negative reported: 7B inverted at 0.72× (latency noise). Demonstration on one task, not a law.🔗 https://macrokit.dev #LLM #opensource
Related
Soofi S shows that Europe can build fast, efficient and fully open AI on sovereign infrastructure. Now scale it beyond o...
Soofi S shows that Europe can build fast, efficient and fully open AI on sovereign infrastructure. Now scale it beyond one country: shared compute, multilingual data and public fun...
Inside Anthropic's state-by-state plan to ratchet up AI rulesBy pushing for ever-tougher AI safety laws, Anthropic is dr...
Inside Anthropic's state-by-state plan to ratchet up AI rulesBy pushing for ever-tougher AI safety laws, Anthropic is drawing a distinction between its state lobbying strategy and ...
Google Photos could soon let you edit photo details with AI by just asking for itHey Google, update the metadata for my ...
Google Photos could soon let you edit photo details with AI by just asking for itHey Google, update the metadata for my photos.https://www.androidauthority.com/google-photos-ask-ph...