AI isn't a productivity tool — it's the new operating system for how startups should be built and run
Diana's central argument is that AI represents a categorical shift, not a productivity improvement. The right framing isn't "AI makes engineers faster" but rather "AI makes previously impossible things possible." Every company process should be redesigned as an intelligent closed loop — one that captures information, feeds it into an AI layer, and continuously improves outcomes — rather than the open-loop, execute-and-forget model most companies still run on.
The practical foundation for this is making the entire organization queryable: record meetings, minimize dark-channel communication (DMs, email), embed agents across tools like Slack, Linear, Notion, and GitHub, and build custom dashboards across every function. With sufficient context, AI agents can analyze sprint output against customer needs and business goals — Diana claims teams doing this have cut sprint time in half and increased throughput roughly 10x, though these figures are unverified.
On the engineering side, Diana describes the emergence of AI software factories where humans write specs and tests, and agents generate and iterate on code until a satisfaction threshold is met. Some teams have reportedly reached the point where no hand-written code exists in their repos. This is framed as the path to the "1000X engineer" — a single person orchestrating a system of agents to build what would previously have required a full team.
Organizationally, Diana argues (drawing on Jack Dorsey's framing) that middle management as information routing becomes obsolete when an intelligence layer handles that function. The new archetypes are: individual builder-operators who make things directly; DRIs owning a single outcome end-to-end; and AI-founder types who build, coach, and lead by example. Everyone ships working prototypes — not decks.
Diana argues that founders are misframing AI as a workflow booster when it actually enables entirely new capabilities and organizational structures. Companies that rebuild themselves around intelligent closed loops, software factories, and token-maximizing teams — rather than headcount — will operate orders of magnitude faster than incumbents.
AI is not just going to change how quickly software gets built or what workflows get automated. It's going to fundamentally change the way startups should be run from what roles it will exist to what products are possible to build.
The right person with AI tools can now build features that used to require an entire team or were just impossible.
In the old world, companies basically ran as open loops. You made a decision, executed it, and didn't always systematically measure the outcome, and adjust the process.
An agent with access to Linear tickets, Slack engineering channels, customer feedback from emails or tools like Pylon and GitHub, high-level plans in Notion or Google Docs, sales calls, and daily stand-up recordings can analyze what was actually shipped in a previous sprint and how well it met customers' needs.
I've seen teams that do this cut their engineering sprint time in half and get close to 10x more done in that time.
Some companies have already pushed this to the point where the repos contain no hand-written code, just specs and test harnesses.
Strong DM's AI team built their own software factory where specs and scenario-based validations drive agents to write, test, and iterate on code until it meets a probabilistic satisfaction threshold, with an end goal of a system that essentially eliminated the need for a human to write or review code.
The classic management hierarchy no longer makes sense with AI loops everywhere, a queryable organization, and software factories. In the old world, you needed middle managers and coordinators to route information up and down an organization; in the new world, the intelligence layer serves that purpose.
Your company's velocity is only as fast as its information flow. Every layer of human routing you can remove is a direct speed gain.
Jack Dorsey's view is that if you keep the same org chart and management structure, you'd miss the shift entirely. The company itself has to be rebuilt as an intelligence layer with humans at the edge guiding it rather than routing information through it.
Maximizing token usage, not headcount, will be the critical shift. The best companies will be the ones that are token maxing.
One person with AI tools can be the equivalent of what used to take a large engineering team at a pre-AI company, meaning dramatically leaner engineering, design, HR, and admin teams.
Early-stage founders have a huge advantage: no legacy systems, entrenched org charts, or thousands of people to retrain, and can build their company right from day one.
Existing companies have to maintain and grow a live product while unwinding years of standard operating procedures, and by their nature will have a much harder time going AI native.
Mutiny is a great example of an existing company spinning up a small internal skunkworks team to build AI native systems from scratch, separate from the core business.
Most people talk about AI in terms of productivity — making engineers more productive or adding copilot to existing workflows — but this framing misses the shift, which is less about productivity boost than entirely new capabilities.
AI should not be a tool your company just uses. It should be the operating system your company runs on. Every workflow, every decision, and every process should flow through an intelligent layer that is constantly learning and improving.
Every important process in your company should be captured by an intelligent closed loop — a closed loop captures information, feeds it back into an intelligent system, and improves the process over time.
To get their full capabilities, you need to provide models with as much context as you would provide an employee.
AI software factories: humans write a spec and a set of tests that define success, AI agents generate the implementation and iterate until the tests pass. The human defines what to build and judges the output; the actual code is the agent's job.
This is how you achieve the 1000X engineer that Steve Yegge talked about — by surrounding a single engineer with a system of agents that enable them to build things they would have never been able to build before.
Jack Dorsey suggests every company will have three employee archetypes: (1) the individual contributor / builder-operator who directly makes and runs things; (2) the DRI (directly responsible individual) focused on strategy and customer outcomes — one person, one outcome; (3) the AI founder type who still builds, coaches, and leads by example.
To build closed loops, make your entire company queryable: record meetings with an AI note-taker, minimize DMs and emails, embed agents throughout communication channels, and build custom dashboards covering revenue, sales, engineering, hiring, and ops.
You should be willing to run an uncomfortably high API bill, because it's replacing what would have taken a far more expensive and inflated headcount.
You cannot outsource your conviction on the power of these tools. You need to develop it yourself by actually sitting with coding agents and using them until you start to break your own priors about what is now possible to build.
Shared with Earmark · earmark-ai.com