AI
🤖Anthropic and Alibaba push AI agent capabilities forward
The Rundown: Anthropic launched computer control features for Claude Code while Alibaba released Qwen3.5-Omni, both advancing AI's ability to autonomously handle complex tasks.
The details:
- ●Anthropic added computer control to Claude Code v2.85, enabling full UI testing and debugging from single CLI prompts without human intervention
- ●Alibaba's Qwen3.5-Omni beats Gemini Pro on audio benchmarks and supports 10+ hours of audio processing with 256K token context
- ●OpenAI launched Codex plugin integration with Claude Code for enhanced workflow management and adversarial code reviews
- ●Microsoft introduced Critique and Council modes in Copilot that outperform single-model solutions by 13.88% using dual-model research systems
Why it matters: These advances signal AI agents are becoming truly autonomous development partners rather than just coding assistants. For founders, this means dramatically faster development cycles and the potential to build complex software with smaller teams, but also the risk of being outpaced by competitors who adopt these tools first.
📰 Source: Multiple sources