GPT-5.3 Instant Focused on Clearer Answers and Fewer Dead Ends
OpenAI released GPT-5.3 Instant on 3 March 2026, updating ChatGPT's everyday model with changes to accuracy, tone, and refusal behaviour.
Claude Code Security Put AI to Work Finding Vulnerabilities
Anthropic announced a limited research preview on 20 February 2026 that reviewed code for vulnerabilities and suggested patches.
Anthropic Published a New Constitution for Claude
The 22 January 2026 document explained the values and priorities Anthropic wanted to shape Claude's behaviour during training.
Manus Joined Meta as General-Purpose Agents Drew Big Attention
Manus announced on 29 December 2025 that it was joining Meta, closing the year with a major deal around general-purpose AI agents.
GPT-5.1 Arrived with a More Natural Everyday Voice
OpenAI released GPT-5.1 Instant and Thinking on 12 November 2025, with changes to instruction following, reasoning time, and conversational style.
Are you using AI yet?
You do not need to use AI for everything. Start with one low-risk task, keep private information out of the prompt, and judge the result by whether it genuinely saves time or improves your work.