Robotics could scale AI into more of the human labour market, but a lot needs to change.
“For robots to be cost-competitive for 10% of human work today, costs would need to decline about 70%. At a 3% decline per year, that would take around 40 years.“
ODP have a unique dataset that tracks ad spending from a daily panel as well as ad budget plans.
This means a lot of interesting and out-of-consensus forecasts on the digital ad companies like Google, Reddit, Meta, and Snap.
Below is a snapshot of the Q3 digital ad forecasts.
Snippet readers can email to get a free five week trial and the full pdf of forecasts which also includes Q4 guide estimates and the interesting consensus deviating forecasts.
Hedge funds are now the largest cash borrowers in private repo markets, $3tn by late 2025, with much of the borrowing supporting leveraged Treasury basis trades.
This is a table of the S&P 500 companies with the highest percentage buy ratings and sell ratings.
All in a period where, according to FactSet, currently 59.9% of all stock ratings are a Buy ratings – if this holds to month-end, it will be highest (month-end) percentage of Buy ratings for the S&P 500 going back to at least 2010.Â
“Early in 2026 the Census Bureau completed a major update to the methodology of the HBA series. There had been a flood of applications starting in late 2024 with characteristics that made them seem like traditional HBA applications, but which actually ended up having a very low likelihood of transitioning to an employer business. There were “applications associated with internet sales,” and the methodological fix excluded them to make the series more accurate.“
Why it matters: Negates one of the most supportive bits of evidence of a post-pandemic entrepreneurship boom.
Source: Full paper here, Economic Innovation Group, Sep 2026.
Most factor work uses the CRSP database. This paper found a dataset predating CRSP.
Momentum and value survived this “out-of-sample” test.
Why it matters: Many factor studies begin with CRSP in 1926. Finding momentum and value in an earlier market makes data-mining a less convincing explanation.
“Clear visibility on copper supply growth extends only about three to four years out. As such, base-case production capability is always forecasted to decline due to closures and grade attrition. However, the actual trajectory of supply has been very different.“
Why it Matters: A narrative violation that copper is becoming more scarce.
The genuine sceptic is not the person who doubts everything, nor the person who demands proof before believing anything. It is the person who has learned to hold their best views with exactly the right strength of grip: firmly enough to act on them, loosely enough to let go when the evidence turns”
