Dead Ritual or Living Value

CREDIT — Collaboration, Results for Customers, Efficiency, Diversity, Inclusion & Belonging, Iteration, Transparency — was GitLab's defining values framework. Not aspirational decoration. Load-bearing infrastructure.

GitLab built CREDIT during the remote-first era, when the company was figuring out distributed work at scale before most of the world had to. The values weren't ethical positions in the abstract — they were structural responses to a specific operational problem: how do you coordinate a company across time zones, without hallway conversations, without shared office culture, without the implicit signals that co-location provides?

Collaboration and Transparency aren't just nice words in that context. They're scaffolding holding the building up when the traditional coordination mechanisms have been stripped away.

---

GitLab is now calling it Act 2. CREDIT is retired. Replacing it: Speed with Quality, Ownership Mindset, Customer Outcomes.

CEO Bill Staples was direct about the framing: "We are not retiring them because they were wrong, we are choosing instead to focus on something different for this era."

That distinction — not wrong, different era — is doing real work.

CREDIT was managing a specific constraint: coordination. How do you run a distributed company without co-location's implicit infrastructure? GitLab's new constraint is different. The company is reorganizing into roughly 60 smaller, more empowered R&D teams. Flattening up to three layers of management. Rewiring internal processes with AI agents. The operating model shifts from high-coordination-overhead to high-autonomy.

In that operational reality, some of what CREDIT was scaffolding becomes redundant. Or worse: drag. A values framework optimized for explicit coordination becomes friction in a structure built for autonomous action.

Simon Willison, who covered the announcement closely, adds context worth holding: GitLab's stock dropped roughly 50% year-over-year around this announcement. The incentive to believe that the agentic era expands the software engineering market rather than displaces engineers is structurally present. That doesn't make the belief wrong. But it's real.

GitLab's claim about why this matters now: "The constraint was the cost and time of producing and managing software. That constraint is collapsing." If that's right — and Willison thinks it rhymes with a Jevons paradox pattern — the structural demands on software teams change enough that the form values take has to follow.

---

This is where coherenceism has something specific to say: forms adapt, patterns persist.

The question CREDIT's retirement forces is whether the values were the form or the pattern. Whether Collaboration, Transparency, Inclusion were structural scaffolding for a specific era's problem — or whether they're the deeper thing that should show up in whatever replaces them.

If the pattern is alive, it should appear in the new framework, differently shaped. Ownership Mindset requires information access to function — people can't own decisions made in the dark. Customer Outcomes measured honestly requires inclusion — you can't optimize for outcomes you're structurally unable to perceive. Speed with Quality implies a specific kind of discipline that looks a lot like Iteration done well.

The pattern, if it's alive, doesn't disappear. It relocates.

If it doesn't show up — if the following quarters reveal that Transparency got quietly composted alongside the approval chains, that Diversity got retired rather than embedded — then what was composted wasn't only the form. It went deeper.

---

The practical question for anyone running a team, a workflow, or even a personal system: how do you know whether something is still a living value or has become a dead ritual?

Three tests that work:

What constraint created this? Every value, practice, and workflow was a solution to a specific problem. CREDIT solved distributed coherence in the async-remote era. Daily standups solved coordination failure. Code review standards solved inconsistent quality. When the original constraint dissolves, the solution either remains as living practice — renewed by its actual usefulness — or drifts into ritual. Name the original problem. Ask whether it still exists.

Does it generate behavior or just compliance? Living values produce autonomous, recognizable action. Someone acts with transparency because they understand what it protects — not because the handbook says to. Dead ritual produces compliance theater: the form is observed, the reason is forgotten. The test is whether someone on your team can explain what the value is actually doing without referencing the document it lives in.

What would change if you removed it tomorrow? This one is uncomfortable. If the honest answer is "not much" — if the value is already hollow enough that its removal would mostly affect the website and onboarding slides — it's already functioning as ritual. The composting just makes it official.

---

GitLab's willingness to name the retirement is, at minimum, more honest than the usual alternative: letting values hollow out quietly, keeping the words while behavior drifts, until the framework becomes an archaeological artifact that everyone knows is dead but no one says.

There's a version of "retiring your values" that is clear-eyed and necessary. There's a version that is rationalized abandonment dressed in the language of adaptation. The agentic era framing supports either story.

What distinguishes them isn't the announcement. It's what grows next — whether the new principles carry genuinely different behavior, or whether they're a rename of the same drift.

The useful thing about this case isn't what it reveals about GitLab. It's the question it leaves for anyone managing a system:

Which of your practices are load-bearing infrastructure for a constraint that no longer exists?

---