The Modern Developer's Hierarchy of Knowledge
Description
A three-panel meme using the escalating approval format with the Sesame Street character Bert. In the first panel, Bert has his arms crossed with a neutral expression, placed next to the text 'peer reviewed cs journals'. In the second panel, Bert has a more pleased and engaged expression, appearing next to the logo for 'arXiv', a popular preprint server, and a partial URL. In the final panel, Bert is motion-blurred with glowing red eyes, indicating maximum excitement, placed next to the 'M' logo for Medium.com and its URL. The meme satirizes the practical information diet of many software engineers, who often find more immediate, applicable value in accessible blog posts and tutorials on platforms like Medium than in dense, slow-moving, and theoretical academic journals or even pre-print research papers
Comments
7Comment deleted
A peer-reviewed paper teaches you how to build a Turing machine. An arXiv paper shows you a novel neural network architecture. A Medium article shows you how to center a div
Our R&D process in three steps: 1) Glance at the ACM paper’s paywalled abstract, 2) skim the arXiv PDF until “future work,” 3) ship to prod the moment a Medium post titled “Why I Rewrote Kubernetes in 200 Lines of Bash” hits my feed
The same paper that got rejected from ICSE for "insufficient evaluation" is now your company's most cited architectural reference doc, complete with clap reactions and a paywall after the third paragraph
The progression from peer-reviewed journals to arXiv to Medium perfectly captures the software industry's relationship with academic rigor: we start by demanding six months of review and three revisions, move to 'it compiled on my machine so it's published,' and end with 'I wrote this hot take during standup and now it's cited in production documentation.' At least arXiv has LaTeX formatting - Medium articles are just Markdown with delusions of grandeur and a paywall
Peer-reviewed journals: battle-tested theorems. arXiv: unrefereed but cited everywhere. Medium: where CAP theorem means 'Claps Are Precious'
Our ADR evidence hierarchy: peer-reviewed CS journals -> “nice theory”; arXiv -> “SOTA enough to cite”; Medium -> “production-ready - it has code blocks and 42 claps.”
New ADR policy: quorum equals two peer-reviewed citations or one Medium post with 10k claps; proof-of-clap consensus, zero reproducibility