Log inSign up
Log inSign up
Frank de Jonge
34.4K posts
@frankdejonge

Frank de Jonge

@frankdejonge
Building duna.com / typo conniosseur / perhaps the most optimistic pessimist ever
Amsterdam, North Holland
frankdejonge.nl
Joined February 2009
493 Following
13.9K Followers
RepliesRepliesRepostsRepostsMediaMedia

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·Ads Info·© 2026 X Corp.
  • @frankdejonge
    Frank de Jonge
    @frankdejonge
    22h
    It has always been difficult to measure impact of SWEs and AI made it a whole lot worse. It’s trivial to prop up a few dozen PRs that make a big splash but deliver little to no business value. A lot of visibility for very little value. I must admit, I struggle with this.
    5
  • @frankdejonge
    Frank de Jonge
    @frankdejonge
    Oct 7
    Move from tactical instructions to communicating values to stay true to, then point towards solid referenced and provide objectives. Form a system around those and see stability increase. Document only what is not expressible in or made obvious by code (this includes test cases).
    @thdxr
    dax
    Anomaly
    @thdxr
    Oct 6
    a weird inversion with LLMs is the models improve faster than the tinkerers when i see people with custom workflows and setups they're all addressing problems that don't exist anymore the person naively using vanilla codex is more likely to be experiencing state of the art
    1
  • @frankdejonge
    Frank de Jonge
    @frankdejonge
    Oct 4
    What’s the verdict? Opinion or sarcasm?
  • @frankdejonge
    Frank de Jonge
    @frankdejonge
    Oct 4
    Imo it’s an incredible time to pay it back and the yield is 🫶❤️ incredible! I have more “time” than ever to do tactical refactors and doing drops token spent on product development. It’s a flywheel babyyyyyyyy
    @GergelyOrosz
    Gergely Orosz
    The Pragmatic Engineer
    @GergelyOrosz
    Oct 4
    The cost of introducing and paying off tech debt has changed very, very much and I don't even know most of us have the time and space to think about what it means Tech debt has been so central to the software craft for decades, and now... how much does it matter?
  • @frankdejonge
    Frank de Jonge
    @frankdejonge
    Oct 3
    Real talk though. At Duna, we primary do behavior driven tests because they explicitly lock in expectations of outcome, not of implementation. This prevents a great deal (if not the majority) of slop lock in. Parts where the behavior is hard to define get more scrutiny.
    @shcallaway
    Sherwood
    Sazabi
    @shcallaway
    Oct 2
    We deleted 800k lines of unit tests from the @sazabi monorepo earlier this week. Three hypotheses: 1- Unit tests "lock-in" slop code by making it hard to change or remove 2- Unit tests slow down dev cycles by making your agent do more work locally and by increasing CI times
    5