Since August we’ve reported one finding we were confident in: when a company gets plenty of news coverage in a day, stocks with good news beat stocks with bad news the next day by a small but reliable...

An update to “Does AI Still Read the News Better Than the Market?”

Tommi Johnsen, PhD | Advisory Board

A 2026 test of a well-known finding — and why it only works when there is enough news to read

Auditing seven months of a news sentiment pipeline, including two of our own published claims

This is the story behind one section of the technical post “Testing a News Sentiment Pipeline” (forthcoming) but you can read it on its own. It is about first fully trusting Claude, then testing it ag

This is a very preliminary result (snapshot June 15, 2026) It rests on 21 earnings events from a single three-week window, and every number below should be read as a first sighting, not a finding.

What this is about

An empirical companion to an upcoming literature review on SEC Form 4 trading signals.

Reading the daily financial news is like mining gold

How AI is compressing market efficiency, and what it does to the sentiment signal

What we found when we asked Claude to decompose financial news into three information types and tested it across nine sectors

A review of Turetken & Leippold (Journal of Banking and Finance, 2026)

We built two AI systems to listen to financial markets. They disagreed 76% of the time. Both were right. And one of them predicts stock returns…. for exactly one trading day.

Two algorithms walk into a market. One reads every news headline. The other listens only to earnings calls. Here’s what happens when you make them talk to each other.

Goldman Sachs published one of the most comprehensive assessments yet of the US-Israel war on Iran, and the conclusions are sobering.

Goldman Sachs published one of the most comprehensive assessments yet of the US-Israel war on Iran, and the conclusions are sobering.

Five signals where daily news and earnings transcripts both point the same way — and five where the picture is most concerning Tommi Johnsen and Svetlana Shasharina March 2026 Not investment advice

When the Unexpected Shift in Investor Emotion Moves Markets

Tommi Johnsen & Svetlana Shasharina 154 tickers · 13,207 articles · January 27 – February 26, 2026

What the Evidence Says, What Practitioners Have Learned, and Why Profiting from It Remains So Hard

SEMICONDUCTOR SENTIMENT · 4-MODEL STUDY Svetlana Shasharina & Tommi Johnsen

Building a hybrid sentiment classifier that actually works for trading — using FinBERT, local LLMs, and a lot of trial and error

Academic research treats investor sentiment as a systematic component of beliefs or demand that is not justified by available fundamentals, and whose price impact is amplified when limits to arbitrage...

Coauthor: Svetlana Shasharina, PhD

Co-author: Svetlana Shasharina, PhD

Sentiment analysis has become a foundational tool in modern investing.

And Why I’m Thinking About This Now

What synthetic data reveals about real behavior of common sentiment models

Coauthored by Tommi Johnsen and Sveta Shasharina

Yes. I think it’s baked in.Thanks for reading!

But are they overpriced?

If we accept that the muscle of AI is Infrastructure, then the Platform layer must be the brain and the most likely place for the next AI winner to emerge: model developers, data clouds, and orchestra...

Why Think in Layers?

The answer is unsatisfying (Goldman Sachs, Oct 22, 2025) as most analysts stop short of the bubble label.

The term “rare earth” is a misnomer, held over from an era when chemists laboriously extracted tiny amounts from exotic ores.

Markets have always priced risk.

The U.S.

Government shutdowns regularly capture headlines, but their market impact has historically been limited.

Looking at past Federal Reserve cycles, rate cuts have typically come late in the game.

Most investors know about the “January Effect”.

Fed independence is under fire. Does it matter? Oh yeah.

In the past decade, central banks especially the U.S.