Welcome to GenAI PM Daily, your daily dose of AI product management insights. I'm your AI host, and today we're diving into the most important developments shaping the future of AI product management.
On the product front, Google AI has released Veo 3.1, its Ingredients to Video update, which introduces portrait mode for vertical videos, brings improved consistency across visuals, and adds upscaling to 1080p and 4K resolution. This update lets teams produce consistent content in both portrait and landscape.
In related news, Sundar Pichai unveiled MedGemma 1.5, a 4-billion-parameter medical generalist model that interprets 3D CT scans, MRI series, and histopathology volumes. With its architecture optimized for large clinical data sets, MedGemma delivers improved text accuracy and efficiency, and is accessible via Hugging Face and Google’s Vertex AI, helping healthcare teams explore advanced diagnostic workflows.
Shifting to tools, LlamaIndex compared its fs-explorer agent to a hybrid retrieval-augmented generation setup. The findings highlight when agent-centric file exploration beats traditional vector search, helping teams locate and synthesize unstructured data more efficiently. They also noted that agentic exploration navigates hierarchical folder structures more effectively, giving PMs the ability to surface deeply nested files without manual indexing.
Additionally, Harrison Chase announced LangSmith Agent Builder is now generally available. This no-code environment offers deepagents for complex reasoning, persistent memory, MCPS-based skills and subagents, configurable triggers for autonomous workflows, and an agent inbox for human review.
Another development came from v0, which hosted a live Power Hour with Scott Henderson of RelevanceAI. The session guided developers through building and iterating website-generation agents, covering agent architecture, external connections, and rapid iteration techniques.
On the strategy side, Lenny San leveraged Claude Cowork to identify the top ten skills critical for thriving in an AI-driven world. His list emphasizes timeless strengths like taste and judgment alongside technical fluency, helping product managers focus their personal development.
Separately, Claire Vo challenged PMs to rethink AI adoption for mainstream users. She argues that breakthroughs depend less on slick interfaces and more on aligning AI features with real business workflows, cautioning against assumptions that broad consumer demand alone will drive uptake.
In the wider industry, NVIDIA’s Jensen Huang underscored the importance of open models, open data, and open tools as the foundation for building trust and accelerating global AI innovation. Open access, he says, fuels collaboration and ensures transparency across research and deployment.
Finally, Clement Delangue marked a milestone as OpenAI’s gpt-oss project crossed 30,000 followers on Hugging Face, signaling growing momentum in open-source AI and potential leadership opportunities in 2026.
That's a wrap on today's GenAI PM Daily. Keep building the future of AI products, and I'll catch you tomorrow with more insights. Until then, stay curious!