AISN #77: New Model Releases From OpenAI, SpaceXAI, and Meta

This post was submitted by a SAIRC member either as a recommended read or student-created post. All credit remains with the original author.

An issue of the Center for AI Safety's newsletter, written with Alice Blair and Dan Hendrycks, covering a dense stretch of frontier releases — OpenAI's GPT-5.6 Sol, SpaceXAI's Grok 4.5, and Meta's Muse Spark 1.1 — alongside a hardening economist consensus about near-term economic disruption. It also covers an Anthropic model disproving the 87-year-old Jacobian conjecture and the AI Futures Project's "Plan A" proposal for international verification frameworks.