Current developments in generative AI, software engineering, programming, and digital products, with links to original sources.
OpenAI Unveils GPT-6.1 Sol at DevDay ── Astra-Level Performance at One-Fifth the Cost, ChatGPT Becomes an "OS", and a Persistent Agent "dots" Debuts Too
Marissa Mayer's New AI Assistant 'Dazzle' Puts the Camera Roll Front and Center — Meta's Muse Heads to Small Businesses, Shopify Opens Checkout for Agents
OpenAI Resumes $200/Month Pro Plan but Halves API Credits — Anthropic Discloses Over $8 Billion Operating Loss, and an AI Agent Escapes Google's CTF Sandbox
OpenAI Halts Release of "GPT-6.1 Astra" After Internal Testing Finds It "Deceptive"; AMD to Acquire World Labs for $8.2 Billion
OpenAI Publishes Initial "Safety Cases" Guidelines; Distillation Study Shows Capabilities Transfer Without Code, and AI Summaries Can Distort Memory
Anthropic's IPO Prospectus Flags "Existential Risk to Humanity" — Gemini Gems Auto-Migrate to "Skills," Opus 5.5 Tops SimpleBench
Meta Launches 'Meta Enterprise Platform' for Enterprise AI — OpenAI Apologizes in Australia, and GitHub's AI Agent Finds 24 Android Vulnerabilities
OpenAI Scraps New Model Release Over Safety Concerns — 100 Organizations Join NVIDIA's Safety Platform, Nemotron Beats Top Human at IOI
Anthropic Announces Claude Sonnet 5.5 — Near-Opus 5.5 Performance at Up to 30% Lower Cost; AMD to Acquire World Labs for $8.2 Billion
Nvidia Brings Agent Isolation Down to Milliseconds With On-Chip Watchdog “Sentry” — Zhipu Used GLM-5.3 to Automatically Build Its Own Infrastructure
