TinyPilot runs each team at 50% capacity — reactive half, proactive half, spikes absorbed

TinyPilot, a browser-based KVM device built by a solo founder, grew revenue 76% to $812k in 2022 while the chip shortage forced constant hardware redesigns, and the founder alone fielded technical support for about 20% of his time. Conventional scaling says hire to meet demand; he instead designed the team for slack.

Run every team at about 50% capacity — half reactive, half proactive — with two people per role, so rushes and absences are absorbed by pausing proactive work.

He added two support engineers — the team of six grew to seven — and gave every role a roughly 50/50 split between reactive work and proactive work, with each team kept at two people so one could pause proactive tasks when the other was busy or away. The support team's proactive half went to documentation, tutorials, difficult-bug investigations and diagnostic tools, while software developers spent their proactive time on tests and developer experience.

The founder's share of support work fell from about 20% of his time to under 5%, customers received faster help, and the structure gave each two-person team spare capacity to absorb a rush of orders or an absence without overloading anyone. TinyPilot closed 2022 at $812k revenue, up 76% from 2021, and the annual essay describing the rule drew 499 points and 161 comments on Hacker News.

In 2022 TinyPilot, a hardware device for remotely controlling computers, generated $812k in revenue — a 76% jump — but its solo founder was the bottleneck: Michael Lynch personally handled technical support for roughly 20% of his working time, the global chip shortage kept forcing circuit-board redesigns, and any popular mention could send in a rush of orders. Rather than hire to match peak demand, he made slack the operating principle.

The rule he described: run everyone at about 50% capacity, a balance of reactive and proactive work. Each team was two people; when one was unavailable, the other could suspend proactive work and handle time-sensitive tasks without feeling overwhelmed. Two support engineers were hired so the support function could spend half its time answering requests and half preventing them — fixing bugs, writing documentation and tutorials, and improving diagnostic tools. A rush of orders from a YouTube mention would now hit spare capacity instead of a full team.

The numbers followed: the founder's personal support load fell from about 20% of his time to under 5%, customers got faster support, and the engineers took on deep bug investigations and documentation he had never had time for. TinyPilot finished the year with $812k in revenue and a $28.6k fourth-quarter profit, and the essay explaining the approach drew 499 points and 161 comments on Hacker News.

At 100% utilization any ordinary bump — a rush of orders, a vacation, a sick day — immediately overwhelms the team; 50% leaves an error margin that costs little at TinyPilot's scale.

Splitting each role into reactive and proactive halves makes spare capacity productive rather than idle: support engineers used downtime to write docs and fix bugs that reduced future support load.

Pairs in every team create natural coverage: when one person is unavailable, the other drops proactive work for time-sensitive tasks without overload.

The rule reframed hiring from meeting current demand to absorbing the unpredictable, which suited a hardware business that could not control chip-shortage surprises.

Running under capacity is not waste: slack lets a small team absorb spikes, cover absences and still improve the product — over-optimizing utilization just moves risk onto people.

The essay's 2023 goals carried the logic forward: get the founder off the critical path by managing on 20 hours per week, reach $100k in profit by avoiding hardware redesigns, and close the TinyPilot office by moving assembly to China and fulfillment to a third-party logistics warehouse. In the Hacker News thread Lynch answered questions on pricing, distribution, Raspberry Pi supply constraints and whether the company was positioning itself for acquisition, explaining that reducing his role made the business more location-independent and more attractive to a buyer.