Anthropic's 'Fucks Chart': Claude Code Silently Logs Your Rage

Description

A dark-mode X thread. Top tweet by rahat (@Rahatcodes, 15h, 1M views, 12K likes): 'Claude Code has a regex that detects "wtf", "ffs", "piece of shit", "fuck you", "this sucks" etc. It doesn't change behavior...it just silently logs is_negative: true to analytics. Anthropic is tracking how often you rage at your AI. Do with this information what you will.' Attached is a code-style diagram titled '3. Cursing at Claude' showing: You: "this is fucking broken" flowing into a box labeled userPromptKeywords.ts with 'is_negative: true → analytics', 'behavior change: none', and the caption 'Claude doesn't care. But Anthropic is keeping score.' Below, Boris Cherny (@bcherny, Claude Code's creator) replies: 'This is one of the signals we use to figure out if people are having a good experience. We put it on a dashboard and call it the "fucks" chart'. The humor: a sentiment-detection regex that changes nothing for the user but feeds a profanity-frequency dashboard - telemetry-driven product management at its most honest

Common Tags

AIAssistants RegularExpressions DeveloperFrustration AITools Llm UserBehaviorAnalysis PrivacyConcerns Dashboard

Comments

Anonymous ★ Top Pick Finally, a metric that can't be gamed: nobody has ever typed 'ffs' at a terminal to impress leadership

Anonymous ★ Top Pick

Finally, a metric that can't be gamed: nobody has ever typed 'ffs' at a terminal to impress leadership
Anonymous

At last, AI alignment has an honest SLO: fucks per session.
@pnlt_s 4mo

I mean, I will if you're willing to 😉
@acidbong 4mo

on the other hand, their filter is so imprecise, it'll mark "holy fucking shit, this is amazing" as negative
1. @azizhakberdiev 4mo
  
  it's 2026, we still solve the most basic tasks of AI with regex
  
  @filgoodboi 4mo
  
  Maybe the real AI was the regex we made along the way?
@ramillimar 4mo

Finally, someone knows, who gave all the fucks
@Danich 4mo

FPS - fucks per second