The software development vertical has become the primary arena for AI competition. Companies like Cursor and Anthropic (with Claude Code) are showing massive, consumption-driven revenue growth, but also face intense scrutiny over pricing, model dependency, and long-term strategy.
The episode introduces a framework for evaluating AI companies based on two axes: 'time-to-value' (how quickly users benefit) and 'durability of value' (how sustainable the benefit is). While early 'vibe coding' apps failed on durability, the coding vertical excels on both, justifying high valuations.
A core investment thesis is the potential for an AI-native company to become the definitive platform for the entire engineering function. Unlike sales (Salesforce) or security (CrowdStrike), engineering has remained fragmented, representing a massive, winner-take-all opportunity.
The discussion reveals how VC firms are adapting to a market of high valuations and extended private timelines. The strategy of "laddering up" to a target ownership percentage (e.g., 20%) through multiple funding rounds has replaced the traditional single-round, high-ownership model.
The conversation delves into the internal culture of a top-tier VC firm, emphasizing rigorous self-assessment through formal reviews of investment misses (like Rippling and ElevenLabs). It highlights the importance of institutional learning and confronting failures to refine future strategy.
Keep pulling the thread on Miles Clements.