Unlike traditional SaaS, AI applications have a unique vulnerability: a step-function improvement in an underlying model could render an app's entire workflow obsolete. What seems defensible today could become a native model feature tomorrow (the 'Jasper' risk).
Related Insights
Overly structured, workflow-based systems that work with today's models will become bottlenecks tomorrow. Engineers must be prepared to shed abstractions and rebuild simpler, more general systems to capture the gains from exponentially improving models.
The ease of building applications on top of powerful LLMs will lead companies to create their own custom software instead of buying third-party SaaS products. This shift, combined with the risk of foundation models moving up the stack, signals the end of the traditional SaaS era.
The enduring moat in the AI stack lies in what is hardest to replicate. Since building foundation models is significantly more difficult than building applications on top of them, the model layer is inherently more defensible and will naturally capture more value over time.
Startups are becoming wary of building on OpenAI's platform due to the significant risk of OpenAI launching competing applications (e.g., Sora for video), rendering their products obsolete. This "platform risk" is pushing developers toward neutral providers like Anthropic or open-source models to protect their businesses.
An AI tool's quality is now almost entirely dependent on its underlying model. The guest notes that 'Windsor', a top-tier agent just three weeks prior, dropped to 'C-tier' simply because it hadn't integrated Claude 4, highlighting the brutal pace of innovation.
