Software Stocks Implode, Claude's Hit List, State of the Union Reactions, Trump's Tariff Pivot

Legal, cybersecurity, and legacy infrastructure sectors hit sequentially

Claude's legal plugin announcement (Feb 3) sent Thomson Reuters and LexisNexis down 10%, followed by security tools (Feb 20) hitting CrowdStrike and Okta, and Cobol modernization (Feb 23) causing IBM's worst day since 2000 (-13%, $31B loss).

Cobol modernization threatens critical infrastructure monopolies

Claude's ability to modernize Cobol—which runs 95% of US ATMs, Social Security, and banking systems—directly threatened IBM's mainframe dominance (85% of Cobol runs on IBM machines).

Market psychology flips from 'when' to 'if'

Investors no longer debate when cash flows will decline but whether they'll exist at all, forcing massive multiple compression (PE ratios halving, revenue multiples dropping from 10x to 3x) and higher weighted average cost of capital.

AI eliminates the 'growth annuity' model

SaaS companies previously traded on predictable metrics like 120% net dollar retention and ARR multiples, but AI introduces binary event risk that makes traditional valuation models obsolete.

Fictional 2028 'death spiral' post moved markets

A Substack scenario predicting AI-driven unemployment and stablecoin disruption went viral with 28M views, tanking financial stocks including Amex (-8%), Capital One (-8%), and Mastercard (-6%) on Monday.

Authorship linked to short-selling hedge fund

The Catrini report's attribution was amended post-publication to include a managing partner of a $262M hedge fund who confirmed short positions in the named companies, suggesting potential market manipulation.

Dueling science fiction narratives replace analysis

As Derek Thompson noted, high uncertainty has turned AI macroeconomic discussions into 'competing science fiction' rather than data-driven analysis, with prediction markets showing only 12% belief in the doomer scenario.

Knowledge work may be a transitory phenomenon

Freeberg argues knowledge work existed only between the computing era and AI era, and that 100x productivity gains could exceed human consumptive capacity, breaking traditional economic models.

The 10% lifestyle improvement imperative

Humans historically require roughly 10% annual improvement in purchasing power to remain satisfied, but AI may create production capacity that outstrips this consumptive ceiling for the first time in history.