Why I built this tracker
This week has seen the release of the GPT6-Astra LLM model. Over the last couple of days I’ve been experimenting with it, and I have been very impressed by the increased quality of its output compared with its predecessors.
Like many others, I am both fascinated and somewhat nervous about what this technology will mean for knowledge work (I’m an accountant and data analyst - two professions that are clearly candidates for AI disruption). I haven’t seen any attempts to build a simple tracker that seeks to quantify the potential impact of AI.
I also thought it would be good to have timestamped projections; if the next three years see these metrics track along the “High” projections then this could be a clear signal that AI is impacting the labour market for white-collar work. Similarly, if they all track along the “Low” dashed lines then would be a rebuttal to claims about a material disruption of knowledge work due to AI. The data is sourced for the United States and Canada as these countries provide the richest data for constructing these metrics.
I intend to keep tracking these metrics through until the end of 2029.