# News: February 12, 2026

> 8 curated links on AI agents, Python, security, Google Cloud, and engineering, selected by Médéric Hurier.

- Author: [Médéric Hurier (Fmind)](https://www.fmind.dev/)
- Published: 2026-02-12
- URL: https://www.fmind.dev/news/2026-02-12/

- [Accelerating Mathematical and Scientific Discovery with Gemini Deep Think](<https://deepmind.google/blog/accelerating-mathematical-and-scientific-discovery-with-gemini-deep-think/>): Google DeepMind unveils "Gemini Deep Think," a model capable of solving professional research problems in math and physics, achieving PhD-level reasoning capabilities.
- [Boston Dynamics & Google DeepMind Form New AI Partnership](<https://bostondynamics.com/blog/boston-dynamics-google-deepmind-form-new-ai-partnership/>): Boston Dynamics and Google DeepMind partner to integrate Gemini Robotics foundation models into the new Atlas robots for improved industrial task performance.
- [GLM-5: From Vibe Coding to Agentic Engineering](<https://z.ai/blog/glm-5>): Z.ai launches GLM-5, a 744B parameter model designed for complex systems engineering and long-horizon agentic tasks, achieving top performance on reasoning and coding benchmarks.
- [Kimi Introduces Agent Swarm](<https://www.kimi.com/blog/agent-swarm.html>): Kimi launches "Agent Swarm," allowing users to deploy up to 100 sub-agents to parallelize tasks like deep research and document synthesis.
- [What to expect in digital advertising and commerce in 2026](<https://blog.google/products/ads-commerce/digital-advertising-commerce-2026/>): Google discusses the rise of "agentic commerce" with the Universal Commerce Protocol and mentions "Gemini 3" powering new ad tools.
- [Mastering Model Adaptation: A Guide to Fine-Tuning on Google Cloud](<https://cloud.google.com/blog/topics/developers-practitioners/mastering-model-adaptation-a-guide-to-fine-tuning-on-google-cloud/>): A technical guide on fine-tuning Gemini 2.5 Flash on Vertex AI and open-source models like Llama 2 on GKE using LoRA.
- [7 Technical Takeaways from Using Gemini to Generate Code Samples at Scale](<https://cloud.google.com/blog/topics/developers-practitioners/7-technical-takeaways-from-using-gemini-to-generate-code-samples-at-scale/>): Google engineers share best practices for building reliable code-generation pipelines, emphasizing decomposition and end-to-end testing.
- [State of the Model Serving Communities - February 2026](<https://inferenceops.substack.com/p/state-of-the-model-serving-communities-f81>): An update on model serving projects like vLLM 0.15.0, KServe, and llm-d, covering new scheduling features and performance improvements.