TREMOR News: Jev Rank Cuts My Daily Headline Scan to a Third

TREMOR News: Jev Rank Cuts My Daily Headline Scan to a Third

Published: October 8, 2026

The daily scan of the News headlines on the TREMOR app was a bit of a pain. It pulls about 800 headlines from 60+ sources. Now I hit Run Jev Rank, and in 2 seconds it scores and ranks the headlines based on my interest areas.

The feeds come from news sites like Reuters, CNBC and Business Standard, central banks and bodies like the Fed, IMF and BIS, and tech and cyber security sources.

Jev, the decision model from TypeSafe, reads only the headline and scores it from 0 to 100 against your interest areas. After a week of use, the sorting is uncannily what I would have done myself, and my daily scan time is down to about one third.

The default interest areas are mine:

You can modify your standing interests in Edit interests, or type a few words into the focus box for a quick check. It goes by intent, so it pulls out related stories even when the headline does not have your words.

Live tool at tigzig.com/tremor/news

How it works

Jev is TypeSafe's (typesafe.ai) System One model. It returns a probability for each question it is asked, and no text. Each headline goes in as one yes or no question, would this reader want to read it, and the probability of yes becomes the 0 to 100 score. A full page of about 800 headlines goes out as four calls that run at the same time.

An earlier post covers the first release: Jev Rank on TREMOR News.


Working on something similar? How I work covers the rates, the availability and what I take on.