20VC: Anthropic Wipes Billions Off Markets | Citrini Research: The Ultimate Breakdown: Agents, "Ghost GDP", Consumer Spend etc. | Figma Earnings Beat & Four Public Stocks to Buy | Jack Altman Joins Benchmark

Jack Altman has data from over 10,000 restaurants showing that customers already want AI agents to order their food. While most investors debate whether artificial intelligence will disrupt commerce, Altman is watching it happen in real time through his portfolio company's metrics. His prediction follows logically: if people trust agents with their pizza orders, they'll trust them with much bigger decisions.

Altman's thesis cuts deeper than consumer preferences. He argues that entire categories of software are becoming "dumb databases" while AI agents capture the incremental value that once justified SaaS valuations. "If all these SaaS apps become dumb databases and Toast just becomes a POS system, they're not going away, but they become more and more commodified," Altman explains. "The real issue is they don't capture enough of the incremental value."

This shift explains why markets wiped billions off software company valuations after Anthropic launched its security product. Investors suddenly grasped that their recurring revenue darlings might become low-margin infrastructure plays. The math is brutal: if an AI agent can perform 80% of what a $50-per-seat SaaS tool does, why pay the premium? Companies will pay for the database functionality but route the high-value work through cheaper, more capable AI systems.

The restaurant data reveals how this transition actually works. Altman's investment gives him access to ordering patterns across thousands of establishments, and the results contradict conventional wisdom about human preference for control. "If your agent has all the historical context on every pizza order you and your wife have made and you know the price point, the location, the delivery time that you're estimating, and it can also analyze every TikTok to Instagram reviews latest food trends," he argues, the agent's recommendations become superior to human judgment. The 5% improvement in outcomes matters more than the psychological comfort of making the choice yourself.

DoorDash's own CTO validates this trajectory. Andy Fang believes "agents and AI will be transformative to our industry," according to Altman. This isn't speculative positioning from a venture investor. It's the technical leader of a $50 billion commerce platform acknowledging that his business model faces fundamental disruption. When platform operators start talking about transformation, the change is already underway.

The employment implications extend far beyond food delivery. Altman projects that "30 to 40 million" US jobs could disappear within four years, focusing on customer support, legal work, bookkeeping, and transportation once Waymo scales. These aren't distant possibilities. Customer service chatbots already handle routine queries better than human agents. Legal research tools outperform junior associates on document review. Autonomous vehicles operate in multiple cities today.

The speed matters more than the scale. Previous technological disruptions unfolded over decades, giving workers time to retrain and economies time to adapt. AI agents compress this timeline into years. A customer service representative can't easily transition to a role that doesn't exist yet. A bookkeeper can't retrain for work that an AI already performs more accurately.

Investors should watch for the commoditization signal Altman identifies. When SaaS companies start emphasizing their data moats over their user experience advantages, they're acknowledging the shift. When they acquire AI companies instead of building competing features, they're admitting their interfaces have become replaceable. When they pivot from per-seat pricing to usage-based models, they're preparing for agent-driven demand that doesn't map to human headcount.

The restaurant data suggests this transformation will happen faster than most expect. If customers already trust AI with personal decisions like food orders, they'll quickly extend that trust to professional decisions like software purchases, investment choices, and hiring decisions. The incremental value will flow to whoever builds the best agents, not whoever built the best dashboards.