Layoffs at Block and Snap Reveal AI Limits in Software Engineering

Recent CEO statements tying Block and Snap layoffs to AI capabilities overlook financial drivers and structural work layers that resist automation. The decide-execute-deliver framework shows AI compresses only the middle execute step while decision and delivery demands persist even in unregulated software engineering. Primary data from employee accounts and company filings confirm no broad productivity threshold was crossed to enable mass replacement. Studies tracking code generation rates, such as those in GitHub's 2023 Octoverse report, indicate AI assists 20-30% of tasks without reducing headcount needs. Economic analyses from Autor (2015) on task complementarity further demonstrate that knowledge work maintains labor demand through coordination barriers absent from pure capability forecasts. Snap's AR division cuts and Intuit's focus on management layers align with cost pressures documented in SEC filings rather than AI exposure metrics.