What is move fast and break things?
Engineering culture and software practice
Move fast and break things was Facebook's early engineering motto, set out in Mark Zuckerberg's 2012 letter to investors, that valued shipping speed over caution and accepted breakage as the price. The company publicly retired it in 2014, replacing it with move fast with stable infrastructure. The phrase is now shorthand, usually critical, for a risk posture that ships quickly and pushes the cost of mistakes onto other people.
What this means
The motto captured a genuine engineering idea: if you release small changes quickly and learn from what happens, you build and improve faster than rivals who deliberate endlessly. The willingness to break things was meant as permission to try, fail and fix, rather than to freeze out of fear of mistakes.
The trouble is what breaking things means once a product is large and important. Inside a tiny start-up, a bug affects a few users and is fixed in minutes. At the scale of a platform used by more than a billion people, the same posture means real harm to real people, and fixing what broke can take longer than the speed was ever worth. That tension is why the company itself walked the motto back.
Today the phrase is used mostly as a warning. When someone describes a culture as move fast and break things, they usually mean it treats speed as the highest value and is relaxed about who pays when something goes wrong. It has become a label for externalising risk.
Why it matters
For leaders, the phrase is most useful as a prompt for an explicit conversation about risk appetite. Speed is genuinely valuable in some settings and reckless in others, and the motto's own history shows why: the very company that coined it concluded that, past a certain scale, unbridled speed cost more than it earned. The lesson of the retraction is that speed depends on stability, not the other way round.
The useful frame is reversibility. Where a change is reversible, low in impact, and gives you fast feedback, moving quickly is often exactly right, because a mistake is cheap to undo. Where a decision touches money, safety, personal data or regulated activity, mistakes can be irreversible or land on people who never agreed to bear them, and there caution earns its keep. Sorting decisions into reversible and irreversible is more practical than arguing about speed in the abstract.
The phrase also matters as a cultural signal when you assess a supplier or partner. If a vendor proudly describes itself as moving fast and breaking things, it is worth asking what exactly they are willing to break, and whether it is your data, your customers or your compliance obligations that will absorb the breakage.
The debate has sharpened with AI. Commentators have revived the motto, sometimes reworked as move fast and fix things, in arguments about how quickly AI capabilities are being shipped relative to their safeguards. That makes it a live question for any organisation deploying AI, not a piece of history.
How it works
Where the term came from
Move fast and break things was Facebook's internal engineering motto in the late 2000s. It was put on the public record in the letter Mark Zuckerberg included in the company's 2012 registration statement, the S-1 filing lodged with the US Securities and Exchange Commission ahead of the flotation. The letter, describing what it called the Hacker Way, states: "We have a saying: Move fast and break things." It goes on to argue that if you never break anything you are probably not moving fast enough.
The company retired the motto at its F8 developer conference on 30 April 2014. As reported at the time, Zuckerberg told developers the mantra had changed to move fast with stable infrastructure, and announced a two-year stability guarantee for the core interfaces other developers depended on, noting the platform then had 1.28 billion users. His explanation was that breaking things had been fine when Facebook was a small company rapidly prototyping, but that once it was large, the time spent fixing what broke outweighed the speed it bought. This is the point of origin, stated as fact; it is not a comment on the company's products.
What it meant inside the company
Within the company, the motto had a fairly specific meaning: make small changes, get them in front of users quickly, watch what happens, and tolerate the occasional failure as the cost of learning fast. It was paired with other slogans such as done is better than perfect and code wins arguments, and with a testing culture that ran many experimental versions at once. Understood this way, breaking things did not mean not caring; it meant not being paralysed by the fear of small, recoverable mistakes.
Why it was retired
The retirement was about scale and dependency. When millions of other applications and more than a billion users relied on the platform, a broken interface did not just inconvenience one team; it broke other people's products, and users blamed the platform, not the developer. The cost of breakage had grown faster than the benefit of speed. Replacing break things with stable infrastructure was an admission that, past a certain size, reliability is what lets you keep moving fast, rather than an obstacle to it.
Its public afterlife
The phrase escaped the company and became a cultural shorthand. Jonathan Taplin used it as the title of his 2017 book, published by Little, Brown, a critical history of how a handful of large platform companies came to dominate culture and, in his argument, undermine the livelihoods of creators. In technology policy debate the phrase now stands for externalising the human costs of speed. From 2024 onward it has been revived in AI coverage, including reworkings such as move fast and fix things, in arguments about whether AI products are being shipped faster than their safeguards.
Examples
A software team of eight building an internal tool decides that small, reversible changes to a low-risk dashboard can ship daily, because any mistake is caught quickly and undone in minutes. For that surface, a move-fast posture is appropriate, and the team sets it deliberately rather than inheriting it as a slogan.
The same team is also building a feature that adjusts customers' billing. Here they apply the opposite rule, because errors touch money and are hard to reverse. Changes go through review, testing and a staged release, and the team explicitly refuses to move fast and break things where the breakage would land on customers' invoices.
A charity evaluating an AI vendor hears the supplier describe its culture as move fast and break things. The charity treats this as a prompt rather than a reassurance, and asks what the vendor is willing to break, how it protects personal data, and what happens when something fails. The answers, not the slogan, decide whether the charity trusts them with sensitive information.
Common misunderstandings
Break things meant being careless. Not as intended. Inside the company it meant tolerating small, recoverable failures rather than being paralysed, and pairing speed with heavy testing. The criticism the phrase later attracted is about what happens when that posture meets scale and irreversible harm.
The motto is still Facebook's slogan. It is not. The company publicly retired it in 2014 at its F8 conference, replacing it with move fast with stable infrastructure, precisely because breakage had become too costly at scale.
Moving fast is always reckless. Not necessarily. Where a change is reversible, low impact and gives fast feedback, moving quickly is often the right call because mistakes are cheap to undo. The posture becomes dangerous mainly around money, safety, personal data and regulated activity.
This is the same as cowboy coding. It is not. The existing Cowboy coding article describes an individual's undisciplined, ad hoc habits, and Death march describes crushing schedule pressure. Move fast and break things is a named institutional philosophy that was deliberately adopted and then publicly retracted, and the phrase now signals a whole risk culture rather than one person's sloppiness.
Saying move fast and break things is a compliment. Rarely, now. Since its afterlife in books and policy debate, the phrase is used mostly as a criticism, implying that a group values speed above the people who bear the cost of its mistakes.
Risks and boundaries
The main misapplication is using the motto to justify a blanket high-speed posture regardless of what is at stake. Its own history warns against that: the company that coined it retreated once the stakes rose. Treating the phrase as a strategy rather than a warning ignores the lesson that speed depends on stability. Equally, dismissing all speed as reckless overcorrects; plenty of change is genuinely reversible and low-risk, and slowing it down wastes effort and opportunity.
The boundaries follow the reversibility frame. The motto does not tell you which of your decisions are reversible, and that judgement is the whole point; misclassifying an irreversible decision as a cheap experiment is where real damage occurs. It also offers no guidance on who bears the cost of breakage, which is exactly the ethical gap critics highlight. The live debate, now centred on AI, is whether the speed of shipping capabilities is outrunning the safeguards, and whether reworkings such as move fast and fix things genuinely change the posture or merely rebrand it.
What to do next
Set an explicit speed policy. Rather than inheriting a slogan, decide deliberately where your organisation will move fast and where it will not, and write it down so teams are not guessing.
Sort decisions by reversibility. Classify changes as reversible and low impact, where fast iteration is encouraged, or as irreversible or high stakes, where review and staged release apply. Anything touching money, safety, personal data or regulated activity belongs in the second group.
Interrogate the phrase in suppliers. If a vendor markets a move-fast culture, ask specifically what they are willing to break and how they protect your data and customers, and judge them on the answers rather than the slogan.
Pair speed with the ability to recover. Invest in the testing, monitoring and rollback that let you move quickly safely, taking the lesson of the 2014 retraction that stability is what makes speed sustainable.
Apply the same discipline to AI deployment. Treat the speed of shipping AI features against the maturity of your safeguards as an explicit decision, especially where errors would fall on people who did not choose to bear them.
FAQs
Where did move fast and break things come from?
It was Facebook's internal engineering motto, put on the public record in Mark Zuckerberg's letter within the company's 2012 S-1 registration filed with the US Securities and Exchange Commission.
When was it retired?
The company retired it at its F8 developer conference on 30 April 2014, replacing it with move fast with stable infrastructure and adding a two-year stability guarantee for the interfaces other developers relied on.
Why did they change it?
Because at scale the cost of breakage outgrew the benefit of speed. With more than a billion users depending on the platform, fixing what broke took longer than the speed was worth, so reliability became the priority.
What did break things originally mean?
It meant tolerating small, recoverable failures as the price of shipping and learning quickly, rather than being careless. It was paired with heavy testing and slogans such as done is better than perfect.
Is moving fast always a bad idea?
No. For reversible, low-impact changes with fast feedback, moving quickly is often right because mistakes are cheap to undo. The posture is risky mainly where decisions touch money, safety, personal data or regulated activity.
How is this different from cowboy coding?
Cowboy coding is an individual's undisciplined habit. Move fast and break things is a named institutional philosophy that was deliberately adopted and then publicly retracted, and the phrase now describes a whole risk culture.
Why is the phrase back in AI debates?
Since 2024 commentators have revived it, sometimes as move fast and fix things, to question whether AI products are being shipped faster than their safeguards, making it a live concern for anyone deploying AI.