Specific, critical insights unravel the hidden traps in Anthropic's billing model
💡 Claude is the best AI product on the market and the least trustworthy company. This week, I hit “monthly org limit” on a Max‑20× plan just days into the billing cycle. Token usage down. Agent was… | Ken Stott | 54 comments
Verdict
"Anthropic exposes usage metrics but not the ones that matter"—Ken Stott nails the core issue with a candid critique of Claude's billing opacity. This is no abstract jab; it's a direct hit on how untrackable 'messages' replace visible tokens, leading users blindfolded into financial quagmires. Stott's analysis isn't propped up by credentials but stands solid on clarity and examples: "a tiny 20-token status check becomes thousands of invisible, high-cost cycles per day." Refreshingly devoid of self-promotion, his focus remains on unraveling the deceptive trade practices under FTC guidelines—a bold move when most would hedge. The phrase "agentic black box" alone cuts through LinkedIn jargon with its vivid depiction of concealed complexities, underscoring his honest assessment. By avoiding platitudes and providing concrete evidence, the post elevates itself above typical industry critiques.
The author expresses disappointment without pretending to downplay their expertise.
While mentioning FTC standards, it does not lean heavily on credentials or industry associations.
There are specific critiques but some points remain somewhat abstract regarding billing transparency.
The post criticizes a company's lack of transparency while relying on vague terms like 'agentic cycles'.
'I love the product' is a minor self-endorsement in an otherwise critical review.
'Transparency' and 'financial risk' are common phrases used but context gives them weight.