UK-US Pre-Launch Tests on Anthropic's Claude Signal AI Safety Scrutiny for OpenAI
Pioneering joint eval: UK & US AI Safety Institutes stress-tested Claude 3.5 Sonnet pre-release for bio, cyber, and software risks—the first under...

Created by CuratorMaster
Track OpenAI closely: GPT releases, ChatGPT updates, Sam Altman moves, partnerships, and the company shaping the AI industry.
Explore the latest content tracked by OpenAI Watch
Pioneering joint eval: UK & US AI Safety Institutes stress-tested Claude 3.5 Sonnet pre-release for bio, cyber, and software risks—the first under...
Sam Altman's rare vulnerability ahead of GPT-5's August 2025 launch:
Strategic UK shift: OpenAI scraps massive Stargate data center over high energy costs and regulations, exposing UK infra gaps.
Escalating threats to OpenAI CEO amid AI tensions and legal fights:
Escalating regulatory heat on OpenAI: Florida AG James Uthmeier investigates alleged harm to minors, national security threats, and ChatGPT's link to...
Heated rhetoric escalates as trial nears on April 27:
ChatGPT's internet browsing feature is back after months offline, letting users access real-time info beyond the model's September 2021 training cutoff. Huge upgrade for staying current.
Key privacy wins for ChatGPT via Apple Intelligence:
Strategic pricing fill: New $100/mo tier slots between $20 Plus and $200 Pro, offering 5x Plus usage for intensive devs while $200 gives 20x.
OpenAI eyes ads as a revenue juggernaut:
High-value bet tested by scrutiny:
Sora discontinuation hits enterprises hard—web/app ends April 26, API September 24, 2026, forcing rushed migrations for integrated workflows.
-...
Competitive landscape heats up with self-reliant AI infrastructure reducing costs and boosting performance:
OpenAI's content moderation balancing act amid adult mode rollout:
ChatGPT's Advanced Voice Mode could soon add a 'Live Camera' feature, with vision mode imminent per code hints—unlocking both teased GPT-4o components from last spring. A big step in real-time multimodal AI.
ChatGPT update: OpenAI quietly axed Study Mode without announcement.
Escalating safety fears: Anthropic's Claude Mythos outperforms Opus 4.6 but is restricted after autonomously finding thousands of zero-days, including...