February 6, 2026
Markdowns vs marked down
Wall Street just lost $285B because of 13 Markdown files
“13 text files tanked $285B?” Commenters call clickbait, cheer memes, and fear job-eating bots
TLDR: A buzzy post linked a $285B tech selloff to 13 open-source text files, igniting a fight over whether AI helpers can replace paid software. Commenters split between “SaaS saves brain space and handles compliance” and “tiny scripts will eat your apps,” with many calling the headline pure clickbait.
Did 13 tiny text files crash Big Tech? A viral post claimed a $285 billion “SaaSpocalypse,” blaming a simple set of open-source notes for spooking Wall Street. The community promptly flipped the table. One camp howled “clickbait,” with users like krupan saying they expected proof and got none. Pros in legal and enterprise piled on, arguing the author glossed over how real legal work (and compliance) actually happens, calling the take glib.
But the meme squad showed up swinging. When someone linked the actual repo — Anthropic’s knowledge-work-plugins — a commenter dropped the old-school joke: “Go away or I will replace you with a very small shell script,” now upgraded for the AI era. Cue napkin math jokes about “$1 million erased per byte” and an archived copy for maximum popcorn.
Beneath the snark is a real split: Do AI “agents” — smart helpers you talk to — actually replace paid software, or just sit on top of it? Skeptics say companies buy SaaS to reduce brain drain and dodge risk, not because they can’t code. True believers say agents plus a few files can leapfrog clunky dashboards. The original post nods that “systems of record” (the core databases companies run on) still have moats, but commenters are torn between “SaaS survives” and “scripts eat everything.”
Key Points
- •The article claims about $285B in tech market value was wiped out on Feb 3, 2026, amid reports of an Anthropic ‘legal tool.’
- •The supposed ‘legal tool’ was, according to the author, a small set of Markdown files in a GitHub repository (‘knowledge-work-plugin’) totaling around 156KB.
- •The author argues agentic AI workflows can bypass traditional SaaS by executing tasks directly from instructions, rather than UI interactions.
- •An experiment using Claude Code with text files is presented as evidence that agents can deliver detailed tax Q&A beyond typical SaaS process automation.
- •The author posits that systems of record retain moats and, when exposed via interfaces like MCP or APIs/CLI, can be efficiently leveraged by agents for rapid reporting and analysis.