Tony Stubblebine once thought he would become a corporate tax attorney. He liked role-playing games, especially their rule books, and took this as evidence that his professional future ought to involve reading a great many rules. It was an admirably literal piece of career planning. Fortunately for everyone who wanted a habit-tracking app, the diagnosis did not survive adulthood.
What stayed with him was the possibility of leveling up. A character could acquire skills, get better, move forward. Years later, when he founded the company that became Coach.me, he was still working on that question: what helps a person make progress? The rule books had supplied a useful obsession, even if they had recommended the wrong occupation.
His parents offered a more practical example. He describes his mother as both meticulous and artistic. His father worked in music, mastering a craft before turning it into a business. Stubblebine followed a similar sequence with software: learn to build things, then try to make a company out of them. He also credits executive coach Jonathan Rosenfeld as an influence on his understanding of coaching.
That interest eventually took him into a rather larger assignment. On July 20, 2022, he became Medium’s second chief executive, succeeding Evan Williams. The Coach.me founder was now responsible for a writing platform whose readers, writers, employees and investors all needed a reason to keep showing up. The business had plenty of activity. It needed a different result.
An engineer learns to want the whole job
Stubblebine studied at Grinnell College, attending from 1996 to 2000. Programming gave him a way into technology; writing widened the possibilities. Much later, he sent a thank-you note to a humanities professor, Professor Cummins, for making him work so hard at his prose. He had arrived at college expecting to become a programmer. Learning to write for an audience gave him another direction to travel.
There is something pleasingly specific about an executive who can trace part of his career to difficult feedback on an essay. Code has to run. Writing has to reach another person. Stubblebine’s working life would keep returning to both demands, although the proportion of each changed considerably.
At O’Reilly Media he worked in engineering and online publishing. He wrote the Regular Expression Pocket Reference, a compact guide to the patterns programmers use to search and manipulate text. His later enthusiasm for personal development did not arrive with a complete wardrobe change: here was someone already interested in making complicated tasks more manageable.
He then worked at Odeo, the podcasting company from which Twitter emerged, as an engineering leader. That brought him into Williams’s orbit long before Medium existed. Stubblebine also worked at Wesabe. These were different products, but they put him close to people trying to work out what the internet should make possible next.
Eventually he built CrowdVine, a bootstrapped social-networking business that found customers among conferences. It made money. It also left him asking an awkward founder’s question: why had he chosen this particular business? He did not especially want to become an authority on conferences. Being able to build a product had turned out to be a separate qualification from wanting to spend years with its customers’ problems.
A checkmark acquires a coach
Lift, which became Coach.me, brought the subject closer to home. Stubblebine was interested in performance across activities, from programming to entrepreneurship. Even watching The Great British Baking Show could send him thinking about how people get better at a craft. The attraction was the learning, with considerable flexibility about the craft.
The early product changed shape several times: more game-like, then more social, then more clearly suited to a phone. In early 2015, the business moved into coaching. The continuity mattered to him. He distinguished working through versions of one idea from repeatedly abandoning an idea for an unrelated one.
There was also a commercial education underneath those changes. The habit tracker had begun with an appealing possibility and an uncertain route to customers and revenue. Coaching gave the business a clearer account of both. His advice to founders now spends considerable time on distribution: who will use this thing, how will they find it, and what will keep the business going?
Coach.me connected tracking with actual people who could help users follow through. It also offered a setting for coaches to develop their own businesses. The nickname “Coach Tony” came from this work and his involvement in Better Humans, the self-improvement publication he built on Medium. For someone who began with software, the work had become increasingly conversational.
The reader gets the keys
Medium was already familiar territory. Stubblebine wrote there, ran publications there and advised the company. Better Humans, Better Programming and Better Marketing reflected his recurring interests: improving a skill, understanding a practice, making useful work travel further. By his own account, the publications he ran once accounted for roughly two percent of Medium’s traffic.
He had also shared office space with the company. When Williams wanted to step down, Stubblebine could make a case that combined years of running smaller businesses with unusually detailed experience of the product. He understood it as a customer and publishing partner, including the parts that frustrated the people using it.
Taking the job meant giving up an easier arrangement. Before Medium, his businesses were running well enough that he described himself as half retired. He remembered taking a photograph during the interview process: himself floating in a pool on a weekday afternoon, a reminder of the leisure he was considering surrendering. He wanted a goal large enough to test him.
That choice gives the familiar startup career story a less familiar shape. He was volunteering for more difficulty after acquiring some freedom from it. The attraction was a platform he already cared about, and the possibility that better writing could do something useful for its readers. A comfortable afternoon had acquired a formidable competitor.
The turnaround came in small increments
In his July 2025 account of the turnaround, Stubblebine put Medium’s monthly losses in 2022 at $2.6 million. Subscribers were leaving too. Cutting expenses would have addressed only part of the problem: a subscription company also needed readers to believe the next month would be worth paying for.
The work included reducing costs, renegotiating loans and restructuring the company’s investment arrangements. Those are consequential changes with human costs, including layoffs and the dilution of existing shareholders. His description of the process was unusually plain about how uncomfortable the negotiations were. Running a business that survived required decisions he had not previously imagined making.
Inside the company, he described a “ladder to the moon.” Each $10,000 improvement, whether a saving or new revenue, counted as another step. The image made a distant financial target legible to people working on individual pieces of the business. A server saving and a better conversion rate could contribute to the same goal.
Medium reached its first profitable month in August 2024. In a June 2025 interview, Stubblebine put that initial profit at about $7,000. For a company of its size, it was a modest amount. As a change of direction, it meant considerably more. He reported in July 2025 that the business had remained profitable since that August.
He has emphasized the tension between cutting costs and keeping enough capacity to improve the product. A smaller team still needed confidence and room to invent. That helps explain why his account of the rescue keeps returning to roles, trust and the purpose of the work. The arithmetic mattered, and people had to carry it out.
What happens when you reward a different story?
The editorial work ran alongside the financial work. Medium introduced Boost, bringing human judgment into recommendations to help thoughtful, informed stories reach readers. A system that distributes attention also teaches writers what is worth making. Change the rewards, and the next batch of submissions can change with them.
Stubblebine’s preferred contributors often have another profession. An engineer explains a hard problem. Someone writes from experience acquired in a job or a life. Their authority comes from doing the thing, and publishing makes that knowledge accessible beyond the room where it would otherwise remain.
He talks about this as an expert economy. The value of an essay can extend beyond the money earned by that essay: an explanation can become evidence of how its author thinks. His enthusiasm for these writers fits his own history. A programmer who learned to write eventually found himself running a writing company.
His stance on AI follows the same concern for whose work creates value. In 2024 he argued that writers deserved consent, compensation and credit when their work was used for training. He also described human curation as useful in dealing with low-quality submissions. The practical question remains what readers will find worth their time.
“I want to be challenged and I want to have an impact on my life, on the world.”Tony Stubblebine, in conversation with Amy Jo Kim
Leave room in the margin
For all his interest in software, Stubblebine still likes a paper notebook in meetings. He favors a landscape format, wide enough for three columns: notes, things to return to, and space for insights. The margin gets promoted to part of the working area. It is a small design preference with the attention to detail of a programmer.
His own writing tools keep evolving. In September 2026, he described trying to make writing the dominant activity of his day, using TK, Medium’s writing app, for journals, plans and a daily scratchpad. He reported producing about 120 notes a month and 2,000 words a day. These were his own working figures, rather than a prescription for everyone else.
The private writing and the published writing serve different moments of the same process. One helps him work out what he thinks; another gives those thoughts a reader. His experiments include briefings prepared from work documents and quick voice notes captured on his phone. The old question about making progress has acquired a new set of buttons.
Some of those experiments are pleasingly domestic. In June 2026 he described putting family pie recipes into TK, then having the assistant repeat ingredient amounts inside the instructions. A recipe that makes a cook scroll constantly between quantities and directions offends his sense of order. Even dessert, apparently, has a user interface.
The child with the rule books had been looking for a way to advance. The adult keeps adjusting the conditions around the work: a coach, a recommendation system, a financial target, a wider notebook. His career has involved several starts, and plenty of revisions. There is still room in that third column.
Keep the conversation going
Read Tony’s writing, explore Coach.me or try TK.