"Work in progress" gets used like it explains everything. Sometimes it's accurate: life keeps changing, and the phrase has mercy built into it. It lets unfinished stay unfinished without turning it into failure. But it also has a second use that is harder to admit. It can mean learning, or it can mean "don't judge this yet," or it can mean "I don't want to decide." Those versions sound identical in the moment, which is why the phrase works so well as an escape hatch.

If the phrase is going to be more than comfort, it needs terms. A useful contract has two clauses: expect to learn, and ship something real.

That second clause is the hard one, because it requires a definition of "real." The definition that holds up in practice is simple: create freeze points. Freeze does not mean finish. Freeze means a reference version gets created and pushed into reality in some form: deployed, published, printed, shared with a real reader, sent as the poem it is today, not the poem it might become. Iteration can continue afterward, but the current version has been made visible.

That is the shipping-loop version of Act Like It’s Impossible to Fail: make the next attempt small enough that reality can answer quickly.

Software culture made unfinished feel normal. In open source, frequent releases were never a confession of low standards. They were the mechanism of learning, often summarized as "release early, release often" (and listen to customers).¹ Then the web took that norm and removed the finish line. Tim O'Reilly called it "the perpetual beta," a world where software is delivered as a service and features ship as continuous improvement rather than scheduled releases.²Once the release cycle disappears, "finished" stops being a real state. Things are always getting tweaked, patched, improved, undone, rebuilt, and that mindset leaks into how people talk about themselves.

The trouble is that a person is not a web app. People build on what gets done, not on what gets promised. A beta label looks mostly technical until the social role becomes obvious. Gmail wore the label for more than five years, and when Google removed it (July 7, 2009), the official explanation explicitly mentioned large enterprises that were not keen to run their business on something that sounded like it was still in a trial phase.³ "In progress" is not only about the state of the thing. It's also about what expectations people are allowed to have about the thing.

Heraclitus is the cliché source, but the cliché is popular because it points at a paradox that feels true: some things remain the same only by changing. The river stays the river because the water turns over.⁴ Interests rotate. Projects come and go. Obsession flips to boredom and back again. Forcing a single finished identity tends to produce a brittle costume, so the goal isn't to stop changing. The goal is to stop confusing motion with progress.

Progress implies direction. Motion is just movement. In code it's possible to refactor forever and never ship. In photography it's possible to edit forever and never let the image exist outside the tool. In poetry it's possible to revise forever and end up with something technically cleaner and emotionally dead. The only reliable test is whether the work touches reality, and freeze points force contact.

There's also a version of unfinished that isn't a problem at all. Japanese aesthetics has language that treats imperfection and impermanence as part of what makes something real. Wabi and sabi are often summarized as "beauty in imperfection," but the philosophical version is subtler. It points at austere beauty, patina, and the texture of time.⁵ That matters because it provides a standard other than perfection: unfinished can be a texture, a living edge, the visible evidence of use.

But unfinished can also be a lie. The lie is using "in progress" to avoid the moment where commitment has to happen, to an artifact, a boundary, or a decision. When a project is alive, "in progress" has pull, the work makes the next move obvious, the thing gets clearer with contact. When a project is dead, "in progress" has drag, the same moves repeat, the same adjustments happen, the work doesn't reveal itself, it just absorbs time. Unfinished because it's growing is different from unfinished because it's a hiding place.

At the same time, freezing can become output theater, and in some domains "shipping" partial versions can cause real harm. So the move is not "ship constantly." The move is "pick a freeze that matches the risk." Share a private draft, not a public manifesto. Set a boundary, not a brand new identity. Make a reference point that reality can push back on.

If "work in progress" is a contract, those two clauses make it usable.

The first clause is psychological: expect to learn. This is close to what people gesture at with "growth mindset," but the evidence is more modest than the slogans. Meta-analytic work finds weak overall effects in both (a) the relationship between mind-set and achievement and (b) mind-set interventions on achievement, while still supporting some benefits in specific circumstances (for example, for students who are academically at risk).⁶ Large experiments also find benefits for lower-achieving students and show that context matters, the environment has to support the message for it to land.⁷ Expecting to learn is not magic. It's permission to iterate without turning every bug into identity.

The second clause is behavioral: ship something real. If nothing ships, progress isn't happening. It's rehearsal.

A person is similar to a codebase. Some parts of a life should be allowed to evolve constantly: taste, interests, experiments, the questions that keep returning. Some parts need to be stable enough to build on: basic reliability, boundaries, the kind of "yes" and "no" people can count on. Stability here doesn't mean rigidity. It means treating those parts like interfaces, not like drafts that might be rewritten at any moment.

Freeze points are a cheap way to respect both clauses. They make work discrete and visible and they remove the ability to hide behind the word "progress" forever.

Once something is frozen, it becomes possible to tell whether the work was developing or rearranging anxiety. It also does one more thing that's easy to miss: it turns identity into a trail of artifacts instead of a speech. If everything is always in progress, the only honest evidence of what's being built is what actually gets released.

So when "work in progress" shows up as a reflex, a second sentence helps: it's in progress, and here's what's getting frozen next.

If that next freeze is fuzzy because the work itself is misaligned, the Discipline Alignment Diagnosis - Free Tool is a useful pressure test.

Rabbit Hole

  1. Wikipedia, "Release early, release often." The article quotes the full phrase ("Release early. Release often. And listen to your customers") and attributes its popularization to Eric S. Raymond's 1997 essay The Cathedral and the Bazaar. This matters because it anchors the slogan as a real, named idea in software culture rather than a vague vibe.
  2. Tim O'Reilly, "What Is Web 2.0," section "4. End of the Software Release Cycle." O'Reilly describes "the perpetual beta" and argues that web-era software shifts from scheduled releases to continuous improvement, with features slipstreamed in frequently. It matters because it explains how unfinished became a normal product state, and how that can leak into life language.
  3. Official Google Blog, "Google Apps is out of beta (yes, really)" (July 7, 2009). The post states Gmail "has worn the beta tag more than five years" and says the beta tag "doesn't fit for large enterprises that aren't keen to run their business on software that sounds like it's still in the trial phase." It matters because it shows "beta" functioning as expectation management and trust signaling, not only as technical status.
  4. Stanford Encyclopedia of Philosophy, "Heraclitus" (Spring 2011 archive edition). The SEP frames the river fragments in a way that emphasizes constancy through change, which fits the identity metaphor here. It matters because it keeps the essay from leaning on a sloppy meme quote.
  5. Stanford Encyclopedia of Philosophy, "Japanese Aesthetics" (Fall 2024 archive edition). This entry situates wabi and sabi inside a tradition that treats impermanence as fundamental, and glosses them as austere beauty and patina. It matters because it lets unfinished be discussed as texture without celebrating avoidance.
  6. Sisk VF, Burgoyne AP, Sun J, Butler JL, Macnamara BN. "To What Extent and Under Which Circumstances Are Growth Mind-Sets Important to Academic Achievement? Two Meta-Analyses." Psychological Science (2018). The abstract reports weak overall effects in both the mind-set/achievement relationship and mind-set interventions, while noting some support for specific contexts (for example, academically at-risk students). It matters because it prevents over-selling "expect to learn" as a cure-all.
  7. Yeager DS, Hanselman P, Walton GM, et al. "A national experiment reveals where a growth mindset improves achievement." Nature (2019). The study reports improved grades among lower-achieving students and finds the effects depend on school context (peer norms aligned with the intervention message). It matters because it grounds the contract in reality: learning is real, but environment and behavior still matter.