Applied AI vs Foundation Models: Know Where You Win

Where to place your bets: prioritize applied AI that leverages your existing distribution and data, and partner on foundation models instead of competing to build them.

Sep 29, 2025 0m 19s

James LePage

Expert Insights

James LePage emphatically identifies Applied AI, not foundation models, as a winning strategy in AI-assisted software development for companies by leveraging their user data and distribution. He underscores the consequential focus should be on enhancing user experience, rather than diving into creating foundation models for smaller businesses with limited resources.

Hear James explain:

Why Applied AI offers significant gains by harnessing existing user data and distribution.
The importance of focusing resources on enhancing the user experience of their product ecosystem.
Why attempting to develop proprietary foundation models might not be a winning strategy for smaller or resource-strapped companies.

Yeah, so we win in applied ai. That's where we win because we have the distribution, we have the users, we have the data. How can we make our experience better? How can we make it more futuristic across all of our ecosystem? We don't win on trying to train our own foundation models, given the size and the staffing and the resources of the company.

— James LePage, Director of Engineering AI, Automattic

THE NEW DEFAULT angle

Here's how to leverage Applied AI effectively within your software development ecosystem:

Focus on your strength: Instead of dedicating limited resources to developing foundation models, understand and capitalize on your company's unique advantages, such as user data and distribution.
Enhance user experience: Use available data to better understand user behaviors and preferences. Apply AI in designing personalized, user-focused experiences.
Be resource-efficient: Larger companies may have the resources to train foundation models, but smaller teams should strategically invest in areas where they can make the most significant impact.
Evaluate AI Implementation: Regularly assess the effects of AI on your service offerings. Use metrics such as user engagement and feedback to measure results and inform your AI strategies.
Foster continuous learning: Stay aware of AI trends and advancements. Orient your team toward learning and implementing new AI techniques that can optimize your software development process, even without owning proprietary models.