
Researchers evaluated leading generative AI models—GPT, Claude, Gemini, and Llama—against classic behavioral‑bias tests used on humans. The study found that GPT, Claude, and Gemini consistently displayed human‑like preference biases, while Llama gave more rational answers in those tasks but behaved like humans on belief‑based questions. In contrast, all models answered math‑heavy belief questions correctly, reflecting their strong computational abilities. The results highlight how AI inherits human biases when trained on human‑generated content, especially in preference‑driven scenarios.

The article links soaring U.S. stock valuations to a long‑term decline in labor’s share of GDP, which has shifted profit generation toward capital and free cash flow. Researchers Atkeson et al. show that as employee compensation fell, more cash remained...

A recent analysis of U.S. mutual fund managers shows they struggle to navigate geopolitical shocks such as U.S. export bans on Chinese firms. When a company lands on the U.S. Entity List, affected funds experience falling returns, rising volatility, and...

The cyclically‑adjusted price‑to‑earnings (CAPE) ratio, long‑standing in market forecasting, has underperformed in the United States for roughly the past 15 years despite success elsewhere. Researchers Rui Ma and colleagues propose recalculating CAPE at the individual stock level and then aggregating by...

Reuters published an opinion piece warning that the business model behind large language models (LLMs) may contain a fatal flaw. While LLMs like ChatGPT boost productivity in low‑cost, generic tasks, their propensity to hallucinate raises doubts about reliability for high‑stakes...

Investors are demanding a higher risk premium on U.S. Treasuries, while confidence in the U.S. dollar remains largely intact. A recent study measuring convenience yields shows a sharp decline for short‑term and 10‑year Treasuries since 2024, but the dollar’s convenience...

Prediction markets such as Kalshi are gaining prominence as alternative forecasters of macroeconomic variables. A recent academic study finds Kalshi's inflation and Fed Funds Rate forecasts are statistically more accurate than traditional Fed Funds Futures and on par with the...

A German think‑tank paper outlines seven "narcissistic" foreign‑policy patterns, arguing that states can exhibit behaviors akin to individual narcissism. The authors illustrate each pattern with recent U.S. actions under President Trump in 2025, citing performative superiority, attention‑seeking stunts, vindictive retaliation,...