Meta Superintelligence Labs Head Says Vision Will Set Startups Apart

Meta Superintelligence Labs head Alexandr Wang said AI agents could change what gives startups an edge: As execution gets easier, founders’ choice of goals and willingness to aim high could matter more.

Speaking with Y Combinator CEO Garry Tan at Startup School 2026 in July, Wang said intelligence and the ability to act would become more widely available. A recording of the conversation was published July 29. “I really think it’s going to be vision and ambition,” Wang said, describing what he sees as scarce resources for founders.

Wang estimated that AI agents could make execution 10 or 100 times easier than it was a decade ago. He said the change gives founders more room to pursue ambitious projects.

Wang contrasted the earlier startup landscape, where smaller teams had to find narrow openings against larger companies, with one in which agents can give a young company more capacity. He described an agent-enabled startup as an enhanced giant competing with a more traditional corporate giant. Startups that embrace agents ambitiously can challenge established firms, he said.

He also pointed to agent groups inside Meta as an example of how organizations might use the tools. Wang described companies as feedback loops: They attract customers, improve the customer experience, then use growth to hire and reach more customers. Much of that cycle is handled by people, he said. Meta has seen groups of agents produce more work than a 100-person engineering team when given a suitable feedback loop and a metric to optimize, Wang said.

Wang described the main bottleneck as spreading AI through the wider world and helping people adapt to it. Separately, he said builders should help businesses and governments adjust to the technology and manage risks involving biosecurity and cybersecurity.

As work shifts from people writing code to coordinating agents, Wang said systematic thinking remains important. He described a progression toward coordinating larger groups of agents, while the need to reason rigorously persists. Wang also urged young people to develop a clear sense of how the world may change and identify durable exponential trends. He pointed to AI progress as the current example, following an earlier era shaped by Moore’s law.