Andrew Cull and the trouble with getting bigger

Andrew Cull thought five million dollars would be the ceiling. When he started RMI in 2003, that was his estimate of the entire available market. By 2010, the company was already doing $6.8 million in annual revenue. The business had outgrown his arithmetic.

The beginning was smaller and considerably less comfortable. A bank would not extend credit to a young founder without a business record. An uncle lent him $18,000. Cull ran the company from a bedroom in the Alki home he shared with his wife, whose salary helped support them. For five years he took no salary from the business.

That origin invites the familiar founder story: a loan, a bedroom, an expanding company. Cull’s subsequent choices make the story more interesting. He has helped build ocean rowing boats, co-founded a backcountry motorcycle organization, flown tailwheel airplanes and written at length about the contents of his computer. The common interest is how an undertaking actually works once enthusiasm has done its part.

“What I didn’t foresee was the size,” he said in 2011. Size would become a recurring problem to solve: how to distribute decisions, how to keep equipment useful, how to make an organization function without requiring its founder to be everywhere. His career makes a good case for paying attention to the practical questions that arrive after a good idea.

When the plan runs out of room

An expanding business can be a curious reward for an outdoors person. Cull had been climbing since fourteen. Later, he found himself running an organization that kept growing while he was away. In the Seattle office, new hires sometimes mistook the soft-spoken young founder for another employee. The org chart was advancing faster than introductions.

By 2008, RMI described operations and training on seven continents. The geography matters to his story. A distributed company has to work across distance; it cannot wait for every question to travel back to a single desk. Preparation has to be carried by people and processes that will be somewhere else when the decision comes.

Cull’s later description of his career includes startups, turnarounds and mature organizations whose growth had outrun their processes. The sequence is telling. Winning more work creates its own work. Someone must decide how information moves, which decisions belong where, and whether the structure that served ten people can still serve the next ten.

His current interests include the way incentives and organizational design shape behavior. Those are useful questions for someone who has already watched a company escape its first forecast. A larger business may have more resources and still ask too much of the people inside it.

Other people’s adventures need builders

In 2010, Cull was among the founders of Backcountry Discovery Routes, alongside Bryce Stevens, filmmaker Sterling Noren, Tom Myers and Paul Guillien. The organization developed mapped routes for motorcycle travelers and worked to preserve access to public land. An adventure, in this instance, began with someone doing the route work.

That distinction gives his outdoor interests a useful texture. Going somewhere and making it possible for other people to go there are related pleasures, but they involve different jobs. A route requires maps, coordination and attention to access. A good afternoon on a motorcycle rests on rather more administration than the photograph suggests.

At SpinDrift Ocean Rowing in Port Townsend, Cull became a co-founder and adviser with Sonya Baumstein. She brought long experience in rowing and expeditions and managed the business day to day. The company built custom ocean rowing vessels using carbon composites, with attention to storage, steering, power and the ability to right themselves.

An ocean boat is an unusually persuasive setting for a discussion about systems. Its occupants have to live with the consequences of small design choices. A compartment, a power arrangement or a steering mechanism has a job to do long after the launch photograph. The attraction, in Cull’s case, fits an established interest in preparation that survives contact with distance.

A runway is a fairly direct performance review

Flying offers another view of that temperament. In September 2015, Cull wrote that he had been a pilot for more than a decade. After earning his private certificate in a Cessna 172, he learned in Piper Cub and Super Cub aircraft. He credited those early tailwheel hours with sharpening his basic flying skills.

He went on to fly Cessna 180 and 185 aircraft almost exclusively at that time, including cross-country trips and short-field landings. His account of landing technique was specific, but so was its caution: readers should work with a competent instructor. The interest was in learning a demanding task properly, with no claim that reading about it could replace instruction.

One earlier trip was recorded with the economy of someone who had simply enjoyed it. After dropping friends in the San Juan Islands, he returned in clear weather and detoured over the glaciers of Mount Olympus. The outward night flight had been turbulent. The return gave him something quite different to look at.

His 2017 note about a 1977 Cessna 180K Skywagon announced that the airplane had been sold. He was pleased it was going to a pilot who would appreciate it. Even the sale notice contained advice about getting training from someone who knew the aircraft. Ownership had ended; the interest in how people used the machine continued.

The founder underneath the operating system

His computer writing brings that same curiosity indoors. Cull began programming and building websites in sixth grade, later converted the home desktop to Red Hat Linux and took a Linux machine to college. He did not make programming his initial profession. He kept the habit of investigating the machinery anyway.

Running companies complicated his personal preferences. He liked tinkering; employees needed dependable tools. In 2016 he described balancing Linux servers and his own experiments with the Apple computers used across his businesses. It is an ordinary management dilemma, made unusually concrete: the founder’s favorite setup does not automatically become a reasonable request for everyone else.

Later that year, he installed Arch Linux after experimenting with other distributions. He wanted to understand what was on his machine and add only what he needed. The installation took four to six hours. He treated the time as an education, while acknowledging that the same arrangement was unsuitable for a broad deployment to employees who were not developers.

By March 2026, the practical irritation had changed. Different client projects required different software dependencies. He rebuilt his workstation around NixOS and published its configuration. His account included a keyboard with blank keycaps and an admission that some special characters still challenged his accuracy. The detail is pleasantly deflating. A carefully configured workspace can still contain a person hunting for punctuation.

The work after the first company

Agema Labs, founded in 2013 and associated with Cull’s current work, takes on strategy and operational problems. Its projects span technology, energy and infrastructure. The appeal of the assignment is recognizable: a company has an ambition, and somebody has to assemble the agreements, processes and technical work that will make it possible.

One published Agema case describes a company spending $338,000 a year on labor-intensive audit processes. Agema says it designed and deployed a rules-based automation system that removed those annual costs. The figure is the firm’s account of the engagement. Its relevance here is the kind of problem being addressed: repeated work embedded in an organization’s routine.

The shift from building one company to working on other organizations is substantial. It asks a founder to look closely at arrangements he did not create. Cull describes stepping in when systems are strained, helping teams regain clarity or correcting structural problems. The unit of attention becomes the decision, the handoff, the recurring piece of friction.

His March 2026 focus list includes agentic coding for reliable software and infrastructure, alongside changes to Agema’s tools and processes. These are stated areas of work, rather than promises of a finished transformation. They also show someone still willing to revise his own setup while advising others on theirs.

“I write here to think clearly, test ideas, document, and share what I’ve learned.”

Andrew Cull / on his personal website

A calendar has an operating limit, too

There is a personal cost to making yourself indispensable. In a 2016 interview, Cull described years of traveling roughly a hundred days annually and working seventy-hour weeks. His advice to entrepreneurs put time with children ahead of differences in income. He also argued for giving one’s life a mission statement, as deliberately as one would give a business one.

In August 2023, he returned to the subject in a LinkedIn article challenging entrepreneurs’ pride in overwork. That position sits usefully beside the bedroom years and the long travel schedule. A founder can recognize the effort it took to start a company without treating the same schedule as a permanent measure of worth.

His 2026 reading project adds another kind of preparation. He arranged books about institutions and power alongside psychology, philosophy and literature. He hoped to read for an hour a day, expected the program to continue into 2027, and favored physical books and independent booksellers. Even the reading list had a sequence: he wanted one book to help him understand the next.

Among the smaller entries in his outdoor writing is a photograph of a backcountry kitchen. The accompanying note says they had finally worked out a setup that made meals easy, after a summer spent in the mountains that still felt too short. It is a modest ending for a career involving global operations and complex projects. Get the arrangements right, and there may be more time for the place you came to see.