Evening Scan

Markets had a bad day and AI had a complicated one. Stocks sank on Middle East strike news while Anthropic shipped two new models that are pulling real benchmark weight — and a court decided Google should be liable when its AI makes things up.

Stocks Sink and Oil Jumps After Wave of Strikes Across Middle East (paywall)

A two-month Iran-U.S. ceasefire is looking fragile, and energy markets are pricing that in fast. (New York Times)

Anthropic Has Caught Up to OpenAI in Image Understanding

Claude Mythos 5 and Fable 5 dropped Tuesday, and the benchmark story is genuinely competitive now — not a consolation-prize narrative. (Understanding AI)

Breaking: Google Liable for Hallucinations

A court finding that a company is on the hook when its AI fabricates things is the kind of precedent that lands on every legal team's desk by morning. (Marcus on AI)

Tata's Chairman Predicts AI Agents Will Replace Half of TCS's Jobs (paywall)

Asia's largest outsourcer saying the quiet part loud — and following it with an actual hiring slowdown — differs from the usual hedged executive commentary. (Bloomberg)

“AI will boost productivity in the near term, but only two [of sixteen economists surveyed] expect more jobs.” — Wall Street Journal