Guardian analysis questions imminent AI jobs apocalypse narrative

Porter’s column captures a backlash against the idea that AGI will imminently erase most jobs. Using Anthropic’s own data, he points out that highly ‘exposed’ workers have not yet seen systematic unemployment spikes, and that headline productivity statistics look underwhelming compared with the 1990s IT boom. That is a useful corrective for Race to AGI readers: the narrative that “AI will instantly eat all work” has been great for valuations and funding, but the deployment reality has been slower, messier and bottlenecked by infrastructure and organizational change.([theguardian.com](https://www.theguardian.com/technology/2026/jul/25/ai-jobs-apocalypse-human-labor))

This does not mean AGI is a mirage, but it does suggest the path is more incremental than some forecasts imply. If it takes a decade for businesses to reorganize around current‑generation systems, the payoff from each new model might look more like a series of compounding productivity bumps than a single phase change. Strategically, that favours players with durable balance sheets and distribution over those betting on a very short, explosive window. For policy, the column highlights a different risk: we could end up dedicating huge chunks of GDP and energy to AI infrastructure without securing matching social gains, which might trigger a political backlash long before AGI‑level capabilities mature.