The core thesis is that companies should be reimagined not as hierarchical structures of people, but as a collection of self-improving AI loops. This model moves from AI as a productivity add-on to AI as the fundamental operating system of the business, capable of learning and optimizing autonomously.
For an AI to learn and improve, the entire organization's context—emails, Slack messages, meetings, code—must be recorded and made accessible. This complete, legible dataset becomes the company's true "brain" and primary source of value, more so than the software built on top of it.
AI is positioned to take over the coordination and information-flow tasks traditionally handled by middle managers. In this new structure, employees are individual contributors (ICs) who operate as builders and operators, with AI facilitating their work and communication.
The value of a company is shifting from its proprietary software to its accumulated business context and skills. With advanced code-generation models, software can be treated as disposable—generated on-demand for specific tasks and then discarded—while the underlying data and knowledge are preserved permanently.
The primary constraint and metric for growth is shifting from the number of employees to the volume of AI model tokens consumed. Companies are achieving significantly higher revenue per employee, suggesting that scaling will be driven by investment in AI computation rather than hiring.
Keep pulling the thread on Jack Dorsey.