Making $$ with AI Agents

By Greg Isenberg

Categories: Startup, Product

Summary

Howie Lou believes the AI agent opportunity is worth far more than Sequoia's $1 trillion estimate, with current deployment at only 50% in software engineering and under 10% across back office, marketing, and sales—indicating massive underpenetration ready for disruption with today's frontier models.

Key Takeaways

  1. Current AI agent deployment is severely underestimated: software engineering shows 50% adoption, but back office (9%), marketing (4%), and sales (4.3%) reveal massive untapped potential across industries that could already benefit from frontier AI capabilities.
  2. Software development paradigm shifted from AI-augmented human workflows to fully autonomous agent development: frontier developers now run 30+ parallel cloud code instances with browser-autonomous agents capable of creating PRs and commenting without human IDE interaction.
  3. AI agent capability breakthrough occurred 4-5 months ago with models like Claude Opus 4.5, enabling truly autonomous work on tasks that previously required many human engineer hours—this capability is now unlockable across every domain beyond software engineering.
  4. Frontier companies have already inverted their development model from mostly human-written code with AI augmentation to fully AI-driven development, while most companies still operate at the three-year-old state-of-the-art, creating a massive competitive gap.
  5. HyperAgent enables building digital employees and autonomous apps; Howie is offering $1,000 free credits to first 1,000 users (committing $1M total) for builders to experiment with agent automation and research.

Topics

Transcript Excerpt

Howy Lou is an absolute legend. I mean, this guy started Air Table, half a billion in revenue, a billion dollars in the bank, growing quarter after quarter. So, he's one of those people that when I want to know where is the world going, I call Howie. This episode is structured into two parts. First, where is the opportunity when it comes to AI agents? I think that there's a trillion dollars up for grabs in AI agents. Does he think there's more? Does he think there's less? Spoiler alert, he think...