“This year, many people’s first introduction to the case that AI will be a big deal was METR’s time horizon plot. It’s a trend that more or less says, newer AI can somewhat reliably do harder and more useful tasks, as measured by how long it would take humans to do the tasks. As of this writing, the best AI can, roughly half the time, do tasks that... See more

When I was jotting down notes for this piece I started listing out all the best observable reasons 2026 could be a bad year.

• the labor market is weakening


• AI capex can’t continue at this rate and is propping up the market


• stocks are expensive and multiples will contract


• private credit boogeyman


• inflation returns