Skip to main content
AI Engineer World's Fair (2024)

Deep Dive: The AI emperor has no DAUs: why most devs still don't use code AI

— PromptPanel

← Head back to all of our AI Engineer World's Fair recaps

frame_0061.jpg

Quinn Slack @sqs / Sourcegraph
Watch it on YouTube | AI.Engineer Talk Details

Quinn Slack, CEO and co-founder of Sourcegraph, gave a thought-provoking talk on the current state of Code AI adoption and the challenges facing the industry. His main argument was that despite the hype, most developers still don't use Code AI tools regularly.

Key Statistics

frame_0165.jpg

  • Only about 5% of professional developers use Code AI tools
  • Total recurring revenue from Code AI usage is estimated at around $300 million ARR
  • This represents only about 1/120th of Salesforce's annual revenue

The Reality of Code AI Adoption

frame_0291.jpg

Quinn presented data showing that while there's a lot of excitement around Code AI, actual usage is much lower than many people realize. He emphasized the gap between perception and reality in the industry.

Building Successful Code AI Products

frame_0775.jpg

Drawing from Sourcegraph's experience with their Cody product, Quinn shared several insights:

  • The importance of dogfooding: If you're not using your own product daily, it won't succeed
  • Avoiding hype-driven development: Focus on real user needs, not buzzwords
  • The "freakish" success of AI code completion and why other features are harder to get right

Product Development Framework

frame_0913.jpg

Quinn introduced a 4-box model for evaluating AI features:

He stressed the importance of aiming for features in the top-right quadrant: high frequency of use and high accuracy/ease of verification.

Future Directions

frame_1024.jpg

  • Searching for the "next autocomplete" - a highly impactful Code AI modality
  • Exploring chat-oriented programming (CHOP) as a potential paradigm shift
  • Building manual, explicit features first before adding "magic"

Resources

This talk provided a sobering but insightful look at the current state of Code AI adoption. Quinn's focus on real-world usage and practical product development strategies offers valuable lessons for anyone working in the AI tooling space.