Jason Says
Google adds remote MCP to Gemini while Anthropic brings Cowork to mobile — the agent war is no longer about model strength, but who can embed deeper into real workflows.
Anthropic expands Claude Cowork to mobile and web, letting users start tasks on desktop, track progress on phones, and pick up results later — even with laptops closed. This marks coding agents moving beyond IDEs into general office workflows.
Google adds background tasks, remote MCP, and more to Gemini API's Managed Agents. Remote MCP support means Gemini agents can now tap external tool chains, directly competing with Claude Code's MCP ecosystem and further cementing MCP as the de facto agent interop standard.
WSJ reports AI giants are giving away massive free computing credits to capture startup customers. This is a golden window for indie developers to validate products on free GPU credits, but lock-in risks are real as platforms compete for long-term retention.
Microsoft is cutting back AI spending by shifting from third-party to in-house models, joining a broader cost-cutting trend. The signal for developers is clear: multi-model flexibility and cost optimization are becoming core competencies in enterprise AI architecture.
TechCrunch analyzes why open source models and frontier labs aren't in a zero-sum game. Open source captures cost-reduction demand for mature use cases while frontier models capture premium for new capabilities — explaining how both Llama and Anthropic can grow simultaneously.
Meta launches Muse, a new AI image generator targeting advertising, decoration, and creator workflows. With Instagram and Facebook as built-in distribution, Muse embedded in Meta's social ecosystem poses serious competitive pressure on standalone AI image tools.
Qualcomm acquires edge AI startup Nexa AI and open-sources GenieX, a GenAI runtime optimized for Hexagon NPUs. This gives developers an official, optimized path to run LLMs on Qualcomm-powered Android devices — a significant infrastructure upgrade for on-device AI apps.
The EU AI Act becomes enforceable on August 2nd. An engineering-focused compliance checklist covers risk classification, transparency requirements, and technical documentation. For AI teams targeting European markets, this is the final compliance window before significant fines kick in.
Investors: Khosla Ventures
AI法律科技独角兽，用AI自动化法律工作流，$1.2B估值标志着AI Legal赛道进入成熟期。
Subscribe for daily AI updates + free playbook
📘 Subscribe Free