newsfilter.io
Fireside Chat, Interview

Why AI Moats Still Matter (And How They've Changed)

  • Software market opportunity is expanding beyond IT spend to replace labor directly, with per-outcome pricing models potentially quintupling revenue despite fewer seats sold.
  • Barriers to software creation have dropped significantly, leading to an anticipated increase in supply, though defensibility will rely on end-to-end workflows, context, network effects, and deep customer embedding.
  • New entrants face high hurdles to reach the "gravitational scale" required for data network effects, particularly in markets with limited new system creation like electronic health records or payroll, where a high rate of new company creation is necessary to challenge incumbents.
  • Incumbents like Salesforce, Zendesk, NetSuite, and ADP are expected to retain market share despite the ability to "vibe code" competitors, as they often overshoot market needs with complex features and edge cases, while some business process outsourcing firms face a bear case of clients bypassing them.
  • The competitive landscape for software is projected to undergo consolidation, with markets of roughly 20 companies likely seeing the bottom 15 go bankrupt, allowing top players to charge higher prices for improved quality.
  • AI is not expected to eliminate jobs but to enable tasks for a low cost that were previously too expensive for human labor, with potential for adoption in areas previously ignored due to training or cost barriers if service costs drop to zero.
  • Major model providers like OpenAI are not anticipated to build specific vertical applications (e.g., orthodontic clinic management) until approximately 2045, though they will likely act as a backend for everyone, build horizontal applications like IDEs, and engage in forward-deployed sales similar to Palantir.
  • AI is distinguished from previous tech cycles by near-total consensus, with expectations that big models could reach 5 billion weekly active users and that markets previously deemed too small may eventually represent trillions of dollars in value.
  • Companies in the "Goldilocks zone of irrelevance," such as those providing janitorial services, are expected to remain stuck in their positions as customers do not expend mental energy on them.
  • Public software companies have faced market pressure due to per-seat pricing models that appear unfair when software replaces labor, whereas the current era is characterized by a "Cambrian explosion" of interesting markets driven by labor substitution.