Perplexity and Intelligence
On 2026-11-22, OpenAI published an article discussing how perplexity measures progress in AI models and how it could help developers establish a shared approach to evaluating emerging capabilities.
The latest from ChatGPT and Claude, in one timeline.
On 2026-11-22, OpenAI published an article discussing how perplexity measures progress in AI models and how it could help developers establish a shared approach to evaluating emerging capabilities.
Anthropic / ClaudeNo new briefings for this day.
On 2026-11-20, OpenAI reported MedChemBench, a benchmark evaluating how AI models interpret molecular structures, reason from experimental evidence, and select compounds in early drug discovery.
Anthropic / ClaudeNo new briefings for this day.
Reported on 2026-11-19: Vercel introduced a new message board that lets Claude Code and Codex agents work together.
On 2026-11-19, Anthropic CEO Dario Amodei reportedly reiterated at the Nasdaq opening bell that Anthropic is committed to pacing AI development regardless of business or political pressure.
Reported on 2026-11-19: OpenAI’s ChatGPT is rolling out Live Dictation, which transcribes audio prompts as users speak rather than after recording ends.
On 2026-11-19, OpenAI reported that it built the Performance Factory, a system of Codex-powered workflows designed to find and fix performance issues to help keep ChatGPT fast.
On 2026-11-18, TechCrunch AI reported that Snapchat lets developers build AR Lenses with AI coding assistants such as Claude Code and Cursor. Snapchat says creators can describe a Lens in plain language and have their AI assistant handle the process from start to finish.
On 2026-11-18, OpenAI reported that Sumitomo Life uses ChatGPT Enterprise to research insurers, prepare proposals, and draft code for investment analysis. OpenAI also reported that early executive use helped drive adoption.
On 2026-11-18, OpenAI reported measuring how coding agents are accelerating AI research, while noting that their performance remains uneven and examining implications for recursive self-improvement.
On 2026-11-17, OpenAI introduced GPT-6.2 Astra for demanding professional work and GPT-6.2 Sol for everyday work. OpenAI states that both models are available through the OpenAI API and to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex.
Anthropic / ClaudeNo new briefings for this day.
On 2026-11-16, OpenAI announced planned new tools, practical guidance, and partnerships to help prevent, detect, and respond to online child sexual exploitation and abuse.
Anthropic / ClaudeNo new briefings for this day.
On 2026-11-13, OpenAI confirmed it began embedded assessments with Apollo Research, METR, Transluce, and Goodfire to independently evaluate model alignment, safeguards, and monitoring systems.
Anthropic / ClaudeNo briefings in this part of the feed.