Nate B. Jones challenges the dominant narrative that AI's primary value lies in reducing headcount. Drawing on Jevons' Paradox, historical technology parallels, and real-world examples from AI-native companies, he argues that the dramatic drop in execution cost creates a vastly larger opportunity for companies willing to raise their ambitions rather than simply cut costs. He presents six practical "unlocks" that are available today, none of which require AGI or speculative breakthroughs.
The video opens with WHOOP's decision to nearly double its workforce to 1,400 people while simultaneously investing in AI. Jones frames this as the most important strategic bet of 2026: the choice between cutting headcount (a fixed-pie mindset) and expanding what is possible (a growing-pie mindset).
The wrong question, Jones argues, is "How many fewer people do we need?" The right question is: "Given that execution cost just dropped by an order of magnitude, what can we do right now that was previously impossible?"
When efficiency increases, total consumption of a resource goes up, not down, because cheaper resources make new applications viable. Jones applies this to AI: if the cost of intelligence-driven work plummets, demand for insight, judgment, creativity, and domain expertise is about to explode. Historical parallels include cheap steel enabling skyscrapers and railroads, cheap computing enabling the internet and mobile, and cheap distribution reshaping the media industry.
When product iteration compresses from months to days, strategy itself changes. Instead of two to four big bets per year, teams can run 200 learning cycles annually. Jones cites Cursor's February 2026 update, which lets developers run up to 20 parallel cloud agents simultaneously, each working on separate branches and opening pull requests. About a third of Cursor's own code and PRs are now written by autonomous agents.
The bottleneck shifts from "Can we build it?" to "Should we build it?", which is a fundamentally human question. Leaders must empower their people to move quickly, accept that old paradigms are shifting, and give teams the courage to be entrepreneurial.
There are roughly 35-40 million developers worldwide, but hundreds of millions of domain experts who know exactly what software their field needs: the doctor who understands her patient panel, the logistics manager who can draw a routing algorithm on a whiteboard, the teacher who knows what adaptive learning her students need.
These experts have been locked out of building by the "translation layer," the expensive, lossy gap between knowing what should exist and making it exist as software. That layer is dissolving. Platforms like Lovable, Bolt, and Replit are putting production-quality development in the hands of non-coders. Jones predicts we will move from tens of millions of builders to hundreds of millions, expanding the total surface area of human problems addressed by custom tools by one to three orders of magnitude.
Most software has been mediocre not because engineers lack talent, but because teams lack the execution capacity to deliver great testing, documentation, security review, performance optimization, accessibility, and visual polish all at once. These are all agent-verifiable tasks that are simply labor-intensive.
When agents handle testing, security, and documentation as standard procedure rather than expensive add-ons, the baseline quality of all software rises dramatically. Differentiation shifts from "can we ship it?" to "what is the amazing customer experience we bring?"
Building and maintaining integrations has traditionally been a nightmare. Jones points to the open protocol movement (referencing MCP and similar standards) as a signal that the world has shifted. Instead of thinking of systems as closed with expensive bridges, companies should think of their systems as fundamentally open, because agents will interact with them regardless.
The cost of building integrations has dropped so far that every company is now a platform by default. The strategic question becomes whether your platform is sticky and delivers real value. Jones argues this kind of platform-level strategic thinking, previously reserved for VP-level conversations, now needs to be socialized across the entire organization because individual contributors can ship two or three integrations in an afternoon.
Companies routinely leave money on the table: passing on a $10 million market because the engineering team costs $3 million, or killing an R&D project with a 20% chance of success because failure costs two quarters of roadmap. When execution costs drop 10x to 100x, all those calculations flip. The $10 million market becomes viable. You can run five experiments instead of agonizing over one.
Capturing these expanded opportunities requires people with vision, domain expertise, and creative insight. The human challenge is getting people to dream big again after years of operating under cost constraints that trained them to think small.
When a team gains a reliable insight into what a customer wants, the default response should be to get it into code, not to write up documentation, raise it to leadership, and wait for approval cycles. This is a cultural shift: for decades, organizations have been told that code is scary and should be gated behind process. In the AI-native world, the instinct to "default to code" is what separates fast organizations from slow ones.