.jpg)
Recently I walked into a different future of work. I was invited into Y Combinator (YC) the legendary startup accelerator that has birthed companies like Airbnb and DoorDash. At YC, they review 20,000 applications each quarter and fund about 200 founders. Each gets $500K and 90 days to build something real.
The founders and their ideas were impressive. But that’s not all that blew my mind. It was how these founders build their organizations. When I asked how their companies were structured, none described functions, roles, or org charts. They weren’t thinking: Who do we hire? They were thinking: What agents do we build?
In their world, work is designed for AI first. In many situations, humans are only brought in with the understanding that their role is to coach the agents. Once the agent performs, the human moves on. Compare that to most Fortune 500 conversations: “We have 5,000 people in this function, how do we augment them with AI?” That question starts in absolutely the wrong place.
Most large organizations benchmark themselves against peers. Each function compares itself to that function at other companies of similar scale and the functional leader pats themselves on the back for being ahead or at least not behind. The problem with this as I've come to see is that almost everyone doing AI transformation is benchmarking against people who are similarly blind to what's possible. It's like asking people lost in a jungle to compare notes on the best way through without anyone climbing high enough to see where the jungle ends. What’s missing is a different reference point.
What's needed is a fundamentally different reference point. I'm calling it the YC Benchmark: the discipline of deliberately seeking out AI-native thinkers, who have never been burdened by what's been built before, to understand what's truly possible when you start with a clean sheet of paper. This isn’t about copying startups. It’s about exposing how limited our current thinking is.
I'm not actually suggesting that YC itself should become a corporate field trip destination for every company like a Bay Area founders zoo. Although if you have that access you will be equally as blown away as I was. I just got back from Guatemala speaking to a group of large business CEOs and I did suggest they fund their own local incubator just to see how AI Native thinkers would solve their biggest problems. But short of such an endeavor, within every organization's ecosystem, there are founders, researchers, and builders who see the world from an entirely different vantage point. Spending time with them isn't just for inspirational fun. It's a strategic necessity.
If you are running a large organization today, here is the honest question the YC Benchmark forces you to ask: Are we truly redesigning work, or are we just adding AI tools to work that was designed for a different era?
The founders I met at YC weren't asking how to make their existing structures more efficient. They were asking what the work actually needed to be and then designing the human-agent system to accomplish that, from scratch with the principle of AI first and humans augmenting agents as needed.
This doesn't mean large organizations need to tear themselves down and rebuild overnight. But it does mean the benchmarking needs to change. The standard can no longer be "how are similar organizations of our size managing this transition?" That question compares you only to people navigating the same jungle you're already lost in.
The new standard has to include the unencumbered view from people building companies who have never once asked themselves, "But what will we do with all these people?"
That question, it turns out, might be the very thing holding us back.