The AI Tea
AI news, quickly brewed
Chronology
Important AI events from the briefing archive, stacked in chronological order for quick context.
Last updated April 26, 2026Google DeepMind's April 21, 2026 launch of Deep Research and Deep Research Max shows a clear shift toward long-horizon, tool-connected research agents built for real operations.
Anthropic's April 17, 2026 Claude Design release introduces a conversational visual workspace for prototypes, decks, and one-pagers with brand-aware iteration and export support.
OpenAI's January 29, 2026 retirement notice for GPT-4o and related ChatGPT models underlines the need for structured migration planning across prompts, QA, and user workflows.
OpenAI’s April 23, 2026 GPT‑5.5 launch (updated April 24 to add GPT‑5.5 and GPT‑5.5 Pro in the API) reinforces a shift toward agent-style models that can plan, use tools, and finish multi-step work.
OpenAI’s April 22, 2026 Privacy Filter release provides an Apache‑2.0 open‑weight model for detecting and masking PII in text, aimed at high-throughput pipelines that can run locally before data is stored, searched, or shared.
In an April 24, 2026 update, Anthropic detailed how Claude handles election-related queries with banners, policy enforcement, and evaluations—including tests showing Opus 4.7 and Sonnet 4.6 trigger web search for many midterms prompts.