- Anthropic expands its Claude ecosystem, launching the agentic tool Claude Cowork on mobile and web, introducing a dedicated version for educators, and transitioning Claude Fable 5 to a pay-as-you-go model.
- OpenAI introduces GPT-Red, an AI model engineered for red-teaming other artificial intelligences, proving significantly more effective than human teams at identifying system vulnerabilities.
- Mira Murati's Thinking Machines Lab launches Inkling, a 975 billion-parameter open-weight AI model, establishing it as a prominent US contender despite remaining behind China's largest open-source offerings.
- Google DeepMind encounters further setbacks with Gemini 3.5 Pro, missing its third release deadline, even as it enhances its Gemini Enterprise Agent Platform with advanced Parallel Web Search grounding.
- Meta's Infrastructure VP reveals a projected 20-month rebuild requirement for AI agents, whilst the company investigates AI token spending limits and implements new parental alerts for adolescent distress in Meta AI interactions.
Sources
- Anthropic (https://www.anthropic.com/)
- OpenAI (https://openai.com/)
- Google DeepMind (https://deepmind.google/)
- Meta AI (https://ai.meta.com/)
- Mistral AI (https://mistral.ai/)
- Hugging Face (https://huggingface.co/)
- The Decoder (https://the-decoder.com/)
- TechCrunch AI (https://techcrunch.com/ai/)
Book the two-hour diagnostic
