January 2026

Simon Willison's Weblog

Claude's new constitution

Anthropic has officially released Claude's full "constitution" document - a 35,000+ token training document that defines the AI's core values and behavior. This follows a leak last year where a...

Read summary
Anthropic Engineering

Designing AI resistant technical evaluations \ Anthropic

Anthropic's performance engineering team has redesigned their technical hiring test three times as each new Claude model has outperformed human candidates on the original assessment. AI capabilities...

Read summary
Simon Willison's Weblog

Electricity use of AI coding agents

AI coding agents like Claude Code consume dramatically more energy than typical chatbot queries, with heavy users burning through thousands of tokens per task. One user's daily coding agent usage...

Read summary
Universe of AI

AI News: Gemini 3.5 LEAKED, GPT-5.3 CONFIRMED and DeepSeek R2?

The AI development race is accelerating with three major model developments happening simultaneously. OpenAI confirmed GPT 5.3 is in development, DeepSeek dropped hints about a new model through code...

Read summary
Simon Willison's Weblog

Giving University Exams in the Age of Chatbots

A professor at École Polytechnique de Louvain conducted an experimental open-book exam that allowed students to use chatbots, requiring them to declare usage in advance and share their prompts. Only...

Read summary