Welcome back to The AI Shift. This week’s edition is all about a much-discussed chart and what it can and cannot tell us about the current state of different occupations’ exposure to AI.

John writes

Allow me to introduce the chart in question, which comes from the latest update of Anthropic’s Economic Index. This type of chart may be unfamiliar to most readers, so just to briefly unpack it, it is a radar chart, where the shape and size of the blue shaded area represents the extent to which different categories of job are theoretically exposed to automation by large language models, and the shape and size of the red shaded area represents the extent to which that theoretical automation is showing up in actual usage data from Anthropic’s set of Claude models.