Google DeepMind's Logan Kilpatrick: Why the Model Eats the Harness

By Sequoia Capital

Categories: VC, Startup

Summary

Google is shifting from the Gemini API as its product through-line to an agent harness that powers agentic properties across 50+ products. This fundamental architecture change enables AI Studio, Search, and Gmail to function as specialized agents taking real action on behalf of users, with coding proven as a general-purpose agent harness.

Key Takeaways

  1. Agent harness is becoming the universal through-line across Google's ecosystem (replacing Gemini API), with ~80% shared base infrastructure that gets specialized for specific use cases like coding vs. consumer-facing agents.
  2. Coding has emerged as the general-purpose agent harness beyond just software development—it's the fundamental primitive for building agentic systems across different products.
  3. Anti-gravity operates as a multi-layered ecosystem (IDE, web agent-first experience, CLI, SDK) meeting developers where they are, with the same harness powering Search, Gemini app, and Cloud products.
  4. Model capability is enabling hyper-realistic content synthesis—real-time video editing can inject contextual details (animals interacting with speakers) that require world understanding and subtle behavioral modeling from the AI.
  5. The shift from passive AI (search, summarization) to active agents (email replies, task execution) creates new product cannibalization challenges requiring strategic product positioning.

Related topics

Transcript Excerpt

So, we could edit this set so it looks like we're >> Yes. Yeah. Yeah. I I want this where we were talking off camera like we should do that for the intro because I think it just like makes all this stuff more capable. I've seen these examples of like such subtle nuance that like make me appreciate that it's like the world understanding playing out. I was I was giving a talk um and was on stage with with my friend Tulsi who leads the model team. I had mentioned to someone in the crowd to like edit the video and they had literally like took the picture, edit it with Omni in real time and this like dog came on the stage. In the edited version, the other guests sort of like look down, see the dog, they like chuckle a little bit. This is while I'm like opining about whatever AI >> laugh about y…

More from Sequoia Capital