Open vs Proprietary Models and Future of Models

As CTO of Block, Dhanji Prasanna has overseen a dramatic enterprise AI transformation, with engineers saving 8-10 hours a week through AI automation. Block’s open-source agent goose connects to existing enterprise tools through MCP, enabling everyone from engineers to sales teams to build custom applications without coding. Dhanji shares how Block reorganized from business unit silos to functional teams to accelerate AI adoption, why they chose to open-source their most valuable AI tool and why he believes swarms of smaller AI models will outperform monolithic LLMs.

Hosted by: Sonya Huang and Roelof Botha, Sequoia Capital

Mentioned in the episode:

• goose: Block’s open-source, general-purpose AI agent used across the company to orchestrate workflows via tools and APIs. 

• Model Context Protocol (MCP): Open protocol (spearheaded by Anthropic) for connecting AI agents to tools; goose was an early adopter and helped shape.

• bitchat: Decentralized chat app written by Jack Dorsey

• Swarm intelligence: Research direction Dhanji highlights for AI’s future where many agents (geese) collaborate to build complex software beyond a single-agent copilot.

• Travelling Salesman Problem: Classic optimization problem cited by Dhanji in the context of a non-technical user of goose solving a practical optimization task.

• Amara’s Law: The idea, originated by futurist Roy Amara in 1978, that we overestimate tech impact short term and underestimate long term.

  
  

  00:00 Introduction

  01:48 AI: Friend or Foe?

  03:13 Block's Journey with AI and Technology

  04:47 Block's Diverse Product Range

  07:04 Driving AI at Block

  14:28 The Evolution of Goose

  27:45 Integrating Goose with Existing Systems

  28:23 Goose's Learning and Recipe Feature

  29:41 Tool Use and LLM Providers

  31:40 Impact of AI on Developer Productivity

  34:37 Block's Commitment to Open Source

  39:09 Future of AI and Swarm Intelligence

  43:05 Remote Work at Block

  45:15 Vibe Coding and AI in Development

  48:43 Making Goose More Accessible

  51:28 Generative AI in Customer-Facing Products

  54:09 Design and Engineering at Block

  55:38 Predictions for the Future of AI