At DevGrid Summit 2026, Foxglove Systems Makes the Case for Boring Software

The Portland product studio Foxglove Systems brought three engineers and one unfashionable argument to DevGrid Summit, the developer conference that filled the Austin convention center with 4,200 attendees last week. The argument, delivered across two talks and a workshop: the most reliable products of the next decade will be assembled from boring, proven parts, and novelty should be treated as a budget to be spent, not a default.

DevGrid, now in its fifth year, has grown into one of the larger independent developer events in the country, with 90 sessions across three days and tracks covering infrastructure, AI tooling and engineering craft. It is a stage where platform vendors usually set the tone. This year, by the informal measure of hallway conversation, a twelve-person studio held its own.

Three talks, one thesis

Founder Theo Lindgren opened with the session that gave the studio its theme, titled Boring by Design. Its centerpiece was a case study: the rebuild of a logistics client's ordering system on a stack Lindgren cheerfully described as fifteen years old on purpose. The team spent its entire novelty budget on one component, a scheduling engine where the product genuinely needed invention, and shipped everything else on technology with a decade of failure modes already documented.

Every team gets a novelty budget. Spend it on the one thing your product actually needs to be new at, and make everything else boring on purpose.

Theo Lindgren, founder, Foxglove Systems

The studio's second talk went deeper into the one place Foxglove has chosen to be novel. Senior engineer June Akagi walked through Stitch, the local-first synchronization engine the studio built for a field-service client whose crews work days without connectivity. Her live demo, merging three days of offline edits from two simulated crews without a conflict dialog, drew the loudest applause of the session and a hallway line that outlasted the coffee break. A sampler of the engine's merge logic is available on the studio's public repository.

The third slot was practical. Engineer Wes Calloway ran an accessibility audit hour in the workshop track, walking about forty developers through auditing their own products against a checklist the studio uses on client work. By the end of the hour, most of the room had filed at least one fixable issue against their own codebase, which Calloway called the only conference metric he trusts.

The pushback, and the plan

Not everyone in the hallway bought the thesis whole. Several attendees argued that boring can become an alibi for underinvestment, and that the studios preaching restraint are often the ones with the senior talent to make restraint work. Lindgren conceded the point on stage during questions, with a caveat: the failure mode of novelty is outages, and the failure mode of boring is refactoring, and he knows which invoice he would rather explain.

DevGrid's organizers announced next year's edition for Denver in September 2027. Foxglove says it will publish the full Boring by Design playbook, including the novelty budget worksheet from the talk, on its site this fall. Judging by the line at their table, the print run should be generous.