Measurement (Google Analytics 4) loads only if you accept. Privacy
This category files posts about frontier model releases, benchmark shifts, and the provider race among ChatGPT, Claude, Gemini, and DeepSeek that changes what a small team can ship. You will find launch recaps, reasoning-quality comparisons, and notes on when a new model is actually worth switching to. Will Spurlock writes from the operator side: what changed, what broke, and what to do this week. Coverage includes research that matters in production, open-weight surprises, and the cost or context-window changes that force an architecture decision. If you want a single shelf for model news tied to implementation, this is that shelf.
136 posts as of 2026-09-09, counted from content/blog frontmatter
AI Models & Frontier News
This category files posts about frontier model releases, benchmark shifts, and the provider race among ChatGPT, Claude, Gemini, and DeepSeek that changes what a small team can ship. You will find launch recaps, reasoning-quality comparisons, and notes on when a new model is actually worth switching to. Will Spurlock writes from the operator side: what changed, what broke, and what to do this week. Coverage includes research that matters in production, open-weight surprises, and the cost or context-window changes that force an architecture decision. If you want a single shelf for model news tied to implementation, this is that shelf.
136 posts as of 2026-09-09, counted from content/blog frontmatter.
Start with "Anthropic Cyber Incidents: Computer Use Stays Off". After Anthropic's Sep 9 cyber-incident assessment I keep computer use off on production agents unless a human is watching. Threshold plus HITL stay live.
