https://www.jasonbelnick.com/writing/ Practical notes on AI adoption and local models. en-us Wed, 01 Jul 2026 00:00:00 +0000 https://www.jasonbelnick.com/writing/ai-adoption-missing-90-percent/ https://www.jasonbelnick.com/writing/ai-adoption-missing-90-percent/ Wed, 01 Jul 2026 00:00:00 +0000 73% of surveyed enterprises use AI but only 10% run on it. The gap is not model quality. It is ownership, sign-off paths, and measured workflows https://www.jasonbelnick.com/writing/the-open-weight-frontier-mid-2026/ https://www.jasonbelnick.com/writing/the-open-weight-frontier-mid-2026/ Wed, 01 Jul 2026 00:00:00 +0000 In one June window, open weights reached the coding frontier. What I'd build on, what I'd admire from a distance, and which numbers to trust https://www.jasonbelnick.com/writing/local-llms-on-a-64gb-mac-studio/ https://www.jasonbelnick.com/writing/local-llms-on-a-64gb-mac-studio/ Wed, 01 Jul 2026 00:00:00 +0000 The awkward middle tier of local AI in mid-2026, the capacity and bandwidth math that governs it, and the honest 75 percent verdict https://www.jasonbelnick.com/writing/the-week-the-cloud-blinked/ https://www.jasonbelnick.com/writing/the-week-the-cloud-blinked/ Wed, 01 Jul 2026 00:00:00 +0000 Fable 5 vanished for 18 days under an export-control order. Frontier model access is now a dependency risk you have to engineer against https://www.jasonbelnick.com/writing/routing-work-between-local-and-cloud/ https://www.jasonbelnick.com/writing/routing-work-between-local-and-cloud/ Wed, 01 Jul 2026 00:00:00 +0000 Model rankings churn monthly; a routing policy survives. Classify work by risk, privacy, latency, and volume, and keep the decision log. https://www.jasonbelnick.com/writing/self-refinement-that-knows-when-to-stop/ https://www.jasonbelnick.com/writing/self-refinement-that-knows-when-to-stop/ Wed, 01 Jul 2026 00:00:00 +0000 Nous Research's Autoreason finds naive self-refinement makes output worse, and fixes it with a blind tournament where doing nothing can win