Episode Description
Jared interviews CS and cybersecurity grad student Daniel Fichtinger about “slopware” in open source. These are projects, often boosted by AI, that perform legitimacy with buzzwords, emoji feature lists, templates, donation links, and sweeping claims, while the underlying code is messy, over-scoped, or not actually delivering what the README promises. Daniel argues the issue is not simply “bad code” or “used AI,” but honesty, scope, and whether the maintainer can explain and maintain the work. Good projects make a strong first impression through humility, clear boundaries, and sometimes explicit limitations or alternatives. They reframe “gatekeeping” as community maintenance, a social contract of not wasting others’ time, using a gardening metaphor where slop spreads like weeds and harms beginners most by teaching bad patterns. Daniel describes stopslopware.net as a linkable educational response to repeated spammy posts and offers rehab steps: rewrite your README yourself, then incrementally replace AI-generated parts until you genuinely understand and can stand behind the whole project.
Links:
- stopslopware.net
- ficd.sh
- Daniel’s blog
- Lobsters
- The XY Problem
- Motherfucking Website
- Crafting Interpreters
- Codeberg
Dead Code Podcast Links:
Jared’s Links:
Hosted on Acast. See acast.com/privacy for more information.