Skip to content
Gopi Krishna Tummala

Tag: multi-agent

All the articles with the tag "multi-agent".

  • Intermediate Agentic AI
    9 MIN READ

    Discovery Harnesses: How Google's Co-Scientist and AlphaEvolve Generate and Test Ideas

    Two Google systems show what a harness can do when the goal is discovery rather than chat: Co-Scientist, a team of agents that generate, debate and rank scientific hypotheses in a tournament, and AlphaEvolve, which evolves code against automated evaluators. How each loop works, what they've found, and the shared idea underneath: generate many candidates, test them hard, keep the best.

  • Intermediate Agentic AI
    10 MIN READ

    Part 02: Agentic AI, Part 2: Memory, Tools and Orchestration

    How agents reach the world through tools and the Model Context Protocol, what they remember and how, why their limited context window is the resource that matters most, how work is split between a supervisor and specialist agents, and how routine steps move to smaller, cheaper models.

  • Intermediate Agentic AI
    10 MIN READ

    Part 03: Agentic AI, Part 3: Agents in the World

    Agents that operate computers and browsers like a person does, robots that plan with language and act with learned skills, 3D scene understanding, imagining the outcome of an action before taking it, and teams of agents that debate and divide work. What each adds, and what makes each hard.