951: Context Engineering, Multiplayer AI and Effective Search, with Dropbox's Josh Clemm - Super Data Science: ML & AI Podcast with Jon Krohn Recap

Podcast: Super Data Science: ML & AI Podcast with Jon Krohn

Published: 2025-12-23

Duration: 1 hr 0 min

Guests: Josh Clemm

Summary

Josh Clemm from Dropbox discusses the challenges of enterprise search and how Dropbox Dash aims to solve them with AI-powered universal search. He explains the concept of multiplayer AI and the importance of context engineering in AI systems.

What Happened

Josh Clemm, Vice President of Engineering at Dropbox, introduced Dropbox Dash, an AI-powered search tool designed to address the inefficiencies in enterprise search. He explained how knowledge workers often use up to 20 different search tools daily, which is inefficient and cumbersome. Dash aims to provide a universal search across all work applications, improving productivity and organization by aggregating and indexing content from various sources.

Clemm discussed the concept of 'multiplayer AI,' which goes beyond personal assistants to work as a teammate that understands not just the individual user but the entire team and their work context. This approach allows for better collaboration and more effective use of AI in the workplace. He emphasized the importance of understanding the organizational context and data to make AI tools more effective and integrated into team workflows.

The episode delved into the issue of 'context rot,' where too much information in an AI system's context window degrades quality. Clemm highlighted the importance of context engineering, ensuring that only relevant data is included to maintain the system's accuracy and performance. He stressed that more context isn't always better and that precision in context inclusion is crucial for effective AI performance.

Clemm warned against over-reliance on A/B testing, which can lead to organizational lethargy and inhibit innovation. He advocated for balancing data-driven decision-making with gut instincts and qualitative insights, encouraging leaders to maintain their conviction and adaptability in the face of data.

He shared insights from his side projects, like an open-sourced AI search app and an earthquake app, which help him stay sharp and informed about emerging technologies. These projects allow him to apply hands-on experience to his leadership role, enabling better guidance and understanding of AI's practical applications.

The conversation also touched on the importance of a people-first leadership approach, especially when managing large teams. Clemm highlighted the need to focus on team inputs rather than just outputs, ensuring that the organizational culture and processes are conducive to growth and innovation. He shared techniques for maintaining a cohesive team culture as teams scale beyond Dunbar's number, emphasizing the role of clear communication and structured processes.

Key Insights