April 12, 2026
Code bros vs hammock pros
The Peril of Laziness Lost
From 'be lazy' to '37k lines a day'—coders feud over AI speed and what counts
TLDR: A veteran engineer blasted AI-fueled line-count bragging after a 37k-lines-a-day boast produced a messy codebase audit. The comments erupted: some backed “write it twice” restraint, others said stop shaming and judge results, while a few called the critique sour grapes—raising the core question: what really counts as good software now?
Old-school wisdom meets new-school swagger, and the comments are on fire. A veteran engineer invoked the classic “laziness” creed from Programming Perl—meaning: think hard, make things simple—then slammed AI-fueled hustle culture after investor Garry Tan bragged about cranking out 37,000 lines a day. A Polish dev’s teardown of Tan’s “newsletter-blog-thingy” found test files, a “Hello World” starter app, a random text editor, and eight logos (one literally empty), igniting memes about “gym bros for code” and judging software by the pound. One commenter cracked that DTrace, a famous tool, is barely 60k lines total; so what are we even counting here?
That’s when the thread split. Nostalgia lit up from gnerd00, who loved the “be lazy, think smart” era. Meanwhile, suzzer99 waved the flag for WET—“Write Everything Twice,” then maybe abstract later—an anti-overengineering mood that got loud cheers. On the other side, johnfn scolded both the bragger and the roasters: don’t worship line counts, but don’t dunk on messy code either—measure value, not volume. Then simianwords dropped the spicy take: the author is “grieving” as AI (those large language models that spit out code) devalues hard-earned chops. One more voice, pityJuke, applauded leadership that “gets it.” The big fight: is smart laziness dying in the age of AI, or are we just changing what “good” looks like? Links were thrown, egos flexed, and somewhere a zero-byte logo shed a silent tear
Key Points
- •The article emphasizes Larry Wall’s “laziness” as a virtue that motivates building strong abstractions to simplify software.
- •It argues modern abstractions have enabled productivity but also encouraged a superficial focus on visible output (“false industriousness”).
- •LLMs are described as amplifying tendencies toward high code volume, not necessarily higher value.
- •Garry Tan’s claim of producing 37,000 LOC/day is contrasted with DTrace’s ~60,000 LOC size to question LOC as a metric.
- •A teardown of Tan’s project found unnecessary inclusions (test harnesses, a Hello World Rails app, a text editor, and multiple logos), highlighting risks of bloated outputs.