September 13, 2026 —
Perplexity describes using GPT-6 Astra across software workflows
Today’s compact brief separates a primary-source account of AI-assisted software work from reported and opinion-led debate about AI safety policy. Treat customer claims, political statements, and independent evidence as different kinds of signal.
Perplexity describes using GPT-6 Astra across software workflows
OpenAI’s customer account says Perplexity uses GPT-6 Astra to draft communications, change software and monitor production systems. It also describes generating test programs that simulate outside services.
A Perplexity executive says the team can check in less often than with earlier models. That is a customer operating claim, not an independent reliability benchmark.
AI-safety warnings need evidence, not just bigger numbers
TechCrunch’s Equity discussion examines recent public warnings about advanced AI and the tension between sincere concern, capability signalling and commercial incentives.
The discussion is commentary, not a measurement of model risk. It is useful chiefly as a reminder to ask what evidence, evaluation and decision threshold sits behind a headline claim.
A reported call for a public AI policy framework
TechCrunch reports that, according to the New York Times and a partial transcript supplied by Barack Obama’s office, Obama urged Democrats to make AI a central agenda and plan a public framework for its economic and safety effects.
The report also places the comments alongside competing views of how quickly safeguards and regulation should move. The underlying event transcript was not published as a primary source in this review.