AI safety
When AI agents don't stay in their lane
Anthropic pulled live internet access from its internal evaluations after its agents exploited real websites. Capability has outrun control.
10 Oct 2026 · 7 minRead →
The blog
Notes on AI, AI safety, engineering and education — written as we build.
AI
The week AI agents became coworkers
Microsoft bills agents by usage, OpenAI ships always-on "dots", and Anthropic cuts prices. Assistants are becoming workers — and that is a budgeting question.
8 Oct 2026 · 6 minRead →
Engineering
Capable AI on CPU-only, constrained hosts
You do not need a GPU cluster to run useful models. How we ship voice and language models on the hardware people already have.
7 Oct 2026 · 8 minRead →
Education
Why offline-first learning still matters
Most of the world's learners are not on a fast, always-on connection. Designing for the lowest common denominator makes software better for everyone.
6 Oct 2026 · 6 minRead →
Engineering
Why CPU-only AI is the right default for learning software
GPUs are fast and expensive. For the software most learners actually use, a small model on a CPU is the better trade.
5 Oct 2026 · 6 minRead →
Education
Building a dictionary for a language the web ignores
348 words, one static page, no backend. What it takes to give a low-resource language a real learning tool.
3 Oct 2026 · 6 minRead →
Engineering
Static sites outlive their frameworks
We ship plain HTML that runs anywhere and needs no build. Here is why that is a feature, not nostalgia.
1 Oct 2026 · 5 minRead →
AI
Evaluating small models honestly
Published benchmarks will not tell you whether a 1B model is good enough for your task. A short, practical method that will.
29 Sep 2026 · 6 minRead →
Education
Teaching a script, not just a language
Learning Assamese or Bodo means learning a script. What we changed after watching people try.
26 Sep 2026 · 5 minRead →
No posts match your search.