Clean Code Revisited in the Age of AI
summary
Revisits Uncle Bob rules against the backdrop of LLMs and shows which rules still hold and which need tweaks: named arguments in Kotlin invalidate the “max 3 args” rule, small single-responsibility functions lower entropy and reduce hallucination, and buried commented code or duplicated generic helpers amplify LM errors. Proposes a single-context principle (put all tokens an LLM needs in one file), explains the loss in the middle effect, and recommends mutation testing plus the 20‑year‑old CRAP metric to force meaningful tests and refactoring.
AI-written summary · about
chapters (15)
- 0:00Clean Code and AI relevance
- 3:13Clean Code still matters
- 7:10LLMs and function design
- 11:02Commands confuse AI
- 14:31Dangers of comments
- 15:34Vertical formatting loss
- 17:39Command-query separation
- 19:47DRY vs semantic duplication
- 21:52Law of Demeter limits
- 22:57New AI-native clean code rule
- 26:49Code to systems engineering
- 27:30Testing with AI tools
- 30:18AI readability and hallucinations
- 31:20Future of software engineering
- 32:43Q&A
related talks
56mClean Code, Two Decades Later
57mCore Software Design Principles for Programmers
11mClean Architecture and AI
16mThe Modular Imperative: Rethinking LLMs for Maintainable Software
45mWhere AI Meets Code
38mNobody Pages the LLM: Engineering Rigour for Vibe Coding
02h 42mClean Code: The Next Chapter
57mThe Art Of Clean Code