Urgent.News

What's breaking now, across thousands of outlets.

Tech

The best engineer on my team ships the least code.

The best engineer on my team almost got a bad review last quarter for shipping too little code. I'm not being cute. I pulled the numbers before his review because I knew the numbers were going to be used against him. Fewest commits on the team. Smallest diffs. Whole sprints where his name was on almost nothing in the changelog. If you ranked the team by anything a dashboard can count, he came out…

For years, many tech companies ranked engineers based on the amount of code they produced, assuming that writing code was the hardest part of the job. However, this assumption is no longer valid. Code is now cheap and easy to produce, while the ability to review and catch potential issues remains scarce. This shift has caused a mismatch between the metrics used to evaluate engineers and their actual value.

In the past, output - such as lines of code, commits, and pull requests - served as a fair proxy for an engineer's contribution. If an engineer shipped a lot of code, it generally meant they had the necessary skills and were making significant progress. However, with the abundance of cheap code, output has become a cheap input, and it no longer accurately reflects an engineer's true value.

The most valuable engineers are now those who go beyond just writing code. They carefully review and catch potential issues that could lead to errors or financial losses. These engineers, although they may ship less code, are the ones who ensure the system remains stable and reliable. Their work often goes unnoticed, as it doesn't produce output that can be measured by traditional metrics.

The review process, which has been built in an era when writing code was the scarce input, now fails to capture the true value of engineers who focus on prevention rather than production. It is difficult to attribute the impact of preventing a bug or financial loss to a specific engineer, as there is no field in the review form to do so. Consequently, these engineers are often undervalued and overlooked.

In summary, the shift from code production to code review as the scarce input has created a mismatch between the metrics used to evaluate engineers and their actual contributions. The engineers who excel in preventing issues, even if they ship less code, are the most valuable members of the team. However, the current review process fails to recognize their true worth, leading to an unfair assessment of their performance.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in Tech

Three citeable security samples from Time Signals

Three items taken from pages that already return full HTML on Time Signals . Not a “trending” listicle — each row is a canonical URL plus the primary sources you should open.

  • Bitget exchange loses $387M in October 2026 theft
  • DTU data breach exposes up to 200,000 users in Oct 2026
  • GoAnywhere MFT vulnerability exploited via license forgery

I tested the OCR switch instead of trusting its label

An OCR checkbox is easy to overinterpret. I see it next to a PDF extractor and assume it will recognize the page again when the existing text is broken. That assumption deserves a test.

  • Author tested OCR switch in PDF tool.
  • Synthetic notice with Chinese and English text used.
  • OCR switch didn't replace missing Chinese text.

Feature Flag SDK Design for Multi-Language Consistency and Performance

You see inconsistent experiment numbers, customers who get different behavior on mobile vs server, and alerts that point to "the flag" — but not which SDK made the wrong call.

  • Enforce deterministic evaluation with canonical JSON and SHA-256 hash function
  • Optimize initialization with non-blocking default path and blocking option
  • Implement reliable updates via streaming with SSE and resilient reconnection

Building a Password Generator in Python

📌 Quick Info Topic: A beginner project using Python's built-in libraries Target Audience: Beginners who want a second project Goal: Build something useful in under 40 lines 1.

  • Python tutorial builds password generator in under 40 lines
  • CLI tool creates robust, random passwords in under a minute
  • Uses random, string modules without additional installation

When I would stop trying PDF extractors and ask for the source

After trying three PDF extractors on the same damaged notice, I still did not have the Chinese text I needed. The outputs looked different, but none contained the original Chinese title.

  • Tried PDF extractors on damaged Chinese text notice with no success
  • Controlled test with synthetic notice showed extractors need source text
  • Requesting source content is key to verifying document accuracy

More from Monday 5 October →