Premium: The Hater’s Guide to Anthropic

Source: Where’s Your Ed At Date: 2026-02-20 URL: https://www.wheresyoured.at/premium-the-haters-guide-to-anthropic/

Summary

Ed’s premium piece targeting Dario Amodei and Anthropic specifically. Core claims: Anthropic loses billions despite appearing profitable ($5.2B lost on $4.5B revenue); Amodei uses “DarioMath” — non-standard profitability metrics that obscure the actual financial position. Amodei strategically times vague apocalyptic AI predictions to fundraising rounds, using hedging language (“as early as,” “possibly”) to avoid accountability while generating media hype. Claude Code reportedly burns “$3 to $20 to make $1.” Anthropic markets itself as the ethical alternative to OpenAI but is “equally deceptive — just better at media manipulation.”

Implications

This is Ed’s most direct piece on Anthropic and Dario Amodei specifically — significant because Anthropic’s safety positioning has generally insulated it from the financial sustainability scrutiny applied to OpenAI. The “DarioMath” label signals that Ed now applies the same analytical standard to safety-focused labs as to profit-focused ones. The $3-20 cost-to-earn-$1 figure for Claude Code is the most specific unit economics claim in Ed’s catalog about this product and is worth tracking against actual pricing changes. The “better at media manipulation” conclusion has direct implications for how Anthropic’s safety communications should be evaluated.