Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchIn a personal essay on DEV Community, software engineer László Szabó describes building an agent skill called crazy-commiting after his employer’s engineering dashboard showed his individual Git commit count below some coworkers’. The skill split his already-finished changes into as many reasonable, self-contained commits as the work allowed. A few weeks later the number went up. By his account, the feature, the code, and the amount of engineering work had not changed. The episode is a clean illustration of a metric rewarding how work is recorded rather than what the work produces. The KPI system, the employer, and the result are described only from the author’s account; none of them is independently verified.
What happened
Szabó says his company tracked individual commit counts as an engineering performance measure, and that he was told his number was lower than several colleagues’. He describes his role at the time as covering architecture, technical decision-making, mentoring, code review, team leadership, difficult debugging, cross-product coordination, and hands-on coding. By his own account, much of that work produced no commits of his own.
As an Amazon Associate I earn from qualifying purchases.
His response was a post-work agent skill. Its job was to inspect pending changes, find the parts that could stand alone, stage them separately, and write a proper message for each. He is explicit about the target: the maximum number of reasonable, coherent commits, not padding. He says the skill was not meant to create fake commits, whitespace-only edits, or empty messages.
His worked example shows the mechanism. One broad commit that synchronized data was reorganized into separate commits for configuration, repository access, mapping, service logic, validation, error handling, and tests. Each piece was coherent on its own. The final repository state was the same.
#1 Best Overall
Why a commit is not a unit of value
The core of the essay is that a commit count measures the shape of version-control history, not the size or importance of the work behind it. Szabó’s points, each illustrated by his own examples, are these:
- A one-character typo fix and a complex data migration can each be a single commit.
- The same final change can be recorded as one commit, four commits, or seventeen, with an identical end state in the repository.
- Where a count is taken can change the number. On his blog he notes that when a branch is squash-merged, a branch with many commits can appear as a single commit on the main branch, depending on which view the dashboard reads.
Taken together, these mean a count can rise or fall because of a branching or merging habit, a splitting habit, or the tooling’s counting rules, none of which touches the product.
Where individual counts miss the work
For senior and lead roles, Szabó lists kinds of work that rarely appear as commits. These include code review, mentoring, system design, production incident investigation, migration coordination, risk reduction, and deciding not to build something unnecessary. He observes that a decision to skip an unneeded service may leave no lines, commits, or pull requests at all. These are his reasons, not measured effects, but they describe a real gap in any count of personal output.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →- Review and mentoring: time spent making other people’s changes better, often invisible in the author’s own history.
- Architecture and decisions: choices that shape the system for months, produced as design notes or conversations rather than code.
- Incidents and coordination: debugging production problems and aligning teams across products.
- Prevention: complexity that never gets written, which has no artifact to count.
Activity data as a question, not a verdict
Szabó does not argue that activity data is worthless. He accepts that an unusual change in repository activity can be useful context, and that it can justify asking a colleague what they are working on. His objection is to skipping that conversation and treating the graph as a conclusion about performance. In his words, “The commit graph can help start the conversation.” His closing question makes the point more sharply: “If I can improve the metric significantly with an agent without improving the product, the team, or the engineering outcome, what exactly is the metric measuring?”
Rank #3
The essay also attributes a version of Goodhart’s Law to the article: “When a measure becomes a target, it stops being a good measure.” The essay does not identify an original source for that exact wording, so readers quoting it should attribute it to the essay or check its origin separately.
What to measure instead
Szabó recommends setting expectations by role and judging work by outcomes. Examples of outcomes he gives are a migration shipping, an incident rate falling, a new hire becoming productive, or an architecture decision holding up under load. On his blog he suggests that leads be assessed on team delivery, technical decisions, and the growth of the people around them, while individual contributors can be held more closely to their own output.
Rank #4
He points readers to two broader frameworks. The table below summarizes how he and the DORA and SPACE material he cites describe them. The DORA figures come from his blog and the DORA site; the SPACE summary is his and should be checked against the original ACM Queue article before it is quoted.
Free tools Windows power users keep installed
One-click scans. No signup required.
| Approach | What it looks at | Scope as described |
|---|---|---|
| Individual commit count | Number of commits by one person | Depends on how work is split, squashed, and counted |
| DORA | Deployment frequency, lead time for changes, change failure rate, and time to restore service (the author notes this has recently been renamed failed deployment recovery time) | Describes how software moves through an organization; the official DORA site identifies it as a Google Cloud research program |
| SPACE | Five dimensions: satisfaction and well-being; performance; activity; communication and collaboration; efficiency and flow | Presented as a multidimensional view, so activity is one input among several |
The comparison is useful because it changes the questions. Instead of asking how many commits a person made, a team asks what shipped, how often, how safely, and whether people could do their work well.
Best Value
- Author: Bungay Stanier, Michael.
- Publisher: Page Two
- Pages: 244
- Publication Date: 2016-02-29
- Edition: 1
Why AI agents make the problem worse
Szabó’s broader argument is that agents make visible activity cheap to produce: commits, pull requests, lines of code, tests, documentation, and tickets. In his example the agent did not write more code. It changed how existing work was recorded in Git history. His forecast that activity metrics will lose value as generation gets cheaper is his interpretation, not a measured industry finding.
Reading the account carefully
The essay is a first-person account, not a controlled study or an independent review of the employer. The dating also needs care. The author’s blog says the reprimand happened in 2025, while the blog itself is dated September 28, 2026, and the DEV page shows “Posted on Sep 29” without a year. No public repository for the skill has been verified, and no effect size is given for the skill or for any organization-level result. The lesson holds regardless of those details: a count of commits describes a habit of recording work, and it should start a conversation before it ends one.
”
The Bottom Line
A commit count tells you how a person’s work was divided into history, not how much engineering value they produced. Use an unusual number as a reason to ask what someone is working on, then judge the work by delivery, quality, and the team’s outcomes.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
