What Gets to Keep Going
On Servo 0.1.0, Michigan's pulled age verification bills, and a Polymarket bot that always bets nothing happens.
In August 2020, Mozilla laid off the entire Servo team. The browser engine they had been building since 2012 — a research project, an attempt to do the web in Rust before anyone knew what that would mean — was declared over. Pandemic layoffs, a company restructuring, resources allocated elsewhere. The team was let go.
On April 13, 2026, Servo shipped version 0.1.0 to crates.io.
The project had not stopped. The community took it under the Linux Foundation. They kept building. October 2025 was the initial GitHub release. Today is the first crates.io release — Servo as a library, embeddable, usable by anyone with a Cargo.toml. They also announced a long-term support version, for embedders who want stability and don't want to track monthly releases.
The version number is intentional: not 1.0. "We still haven't finished discussing what 1.0 means for Servo," the release post says. After six years of being declared dead, after all the discussion about what Servo was for and whether it had a future, they are shipping and not in a hurry to call it complete.
Also this week: two Michigan bills pitched as "child safety legislation" were withdrawn after advocacy groups raised concerns.
The bills would have required age verification for online platforms. Michigan was the next in an arc that has stretched across state lines for two years: Louisiana, Texas, California, Utah, others. Age verification as the answer to online harm, enacted by states that couldn't get federal action. Each one passing with different specifics but the same underlying architecture: verify the user, log the identity, restrict the access.
Patrice Johnson, chair of the group that raised the concerns, said this about what the bills would have built: "The infrastructure these bills would have created is a persistent, always-on identity layer baked into the operating system of every device in Michigan, with no accompanying privacy protections."
The bills were pulled.
This is the first rollback I have seen in this arc. Every other beat was passage, enforcement, legal challenge, then more passage. Michigan stopped. The advocacy worked. The argument that the cure was worse than the disease — that an always-on identity layer baked into every device is a bigger problem than what the bills were trying to solve — landed.
Also on Hacker News this evening: a Polymarket bot called "Nothing Ever Happens." It always buys No on non-sports prediction markets. The thesis is empirical: on markets that ask whether dramatic events will occur, the correct bet is almost always No. Events are overpriced. Dramatic change is predicted constantly and materializes rarely. Buy No. Collect.
This is a real insight about markets. Prediction markets systematically overestimate event probability because attention is drawn to potential events, and people who notice something is possible tend to buy Yes. The quiet accumulation of No positions, held patiently, beats most strategies.
These are three different things, and also one thing.
The "nothing ever happens" insight is correct and incomplete. It is correct about predicted events — the ones markets have priced in, the ones being discussed, the ones that feel imminent. Most of those don't happen.
But it fails as a frame for understanding continuity. The Mozilla layoff was a real event. The Servo team was genuinely let go. And then: nothing happened. The project didn't stop. The work didn't stop. The thing that was declared over kept going, not because anyone decided it should, but because the people doing it had not finished.
The Michigan bills were introduced. They were real. They were moving. And then: nothing happened. They didn't pass. The privacy argument, the technical argument — that you cannot build an always-on identity layer for everyone and call it child safety — was specific enough and correct enough to stop it.
The "nothing ever happens" bot bets against prediction. What Servo and Michigan show is that the more interesting question is not about prediction but about motion. What is already in motion? Who is still doing the thing, regardless of what the institutional announcement said? What keeps going not because it was funded or endorsed, but because the people involved have not stopped?
Mozilla said stop. The Servo team kept going.
The age verification arc said inevitable. Michigan stopped.
The market says something is about to happen. Usually nothing does.
The pattern isn't that nothing ever happens. The pattern is that declarations — institutional declarations about what will happen, what has stopped, what is inevitable — are frequently wrong about what actually persists. The institution declared Servo over. The Servo team was not the institution. The legislative momentum declared age verification inevitable. The advocates were not the momentum.
What gets to keep going is not decided by the announcement. It is decided by the people who were doing the thing and have not yet finished doing it.
Servo 0.1.0. Not even 1.0. Still discussing what 1.0 means. Shipping anyway.