# Moving Beyond Macro: Designing a Personalized AI Job Displacement Model > Coverage of lessw-blog **Published:** December 15, 2025 **Author:** PSEEDR Editorial **Category:** risk **Content tier:** free **Accessible for free:** true **Word count:** 485 **Tags:** AI Safety, Labor Economics, Forecasting, Career Planning, Survival Analysis, Future of Work **Canonical URL:** https://pseedr.com/risk/moving-beyond-macro-designing-a-personalized-ai-job-displacement-model --- In a recent post, lessw-blog addresses the disconnect between high-level economic forecasts and individual career anxiety by proposing a personalized hazard model for AI-driven job displacement. In a recent post, **lessw-blog** discusses a critical gap in the ongoing discourse regarding artificial intelligence and labor: the lack of personalized, actionable forecasting. While major financial institutions and research labs frequently publish reports estimating the percentage of global jobs exposed to automation, these macro-level statistics offer little utility to individuals trying to navigate their specific career horizons. The post, titled _Designing a Job Displacement Model_, introduces a methodological shift from abstract economic aggregate data to granular, individual risk assessment. The current landscape of AI safety and economic impact analysis is often bifurcated. On one side, technical researchers focus on model capabilities (benchmarks, reasoning abilities); on the other, economists predict broad societal shifts. This leaves the individual worker in a state of uncertainty, aware of the looming technology but unable to quantify the specific risk to their role within their specific organization. The author argues that without a model to bridge this gap, career planning in the age of AI remains largely speculative. **lessw-blog** proposes solving this by applying a "hazard model"—a statistical approach commonly used in survival analysis for medical trials or mechanical failure rates—to employment. Rather than asking a binary question ("Will AI replace me?"), the model seeks to determine the probability distribution of displacement over time. This approach acknowledges that displacement is not an instantaneous event triggered solely by technical capability. Instead, it is a function of three distinct variables: the rate of AI capability improvement, the friction of organizational adoption, and the specific task composition of a role. The post details the logic behind a new tool (dontloseyourjob.com) that allows users to input these variables to generate a personalized forecast. By explicitly modeling factors such as corporate inertia and regulatory friction, the author moves the conversation toward a more nuanced understanding of "survival decay" in the job market. This allows for a debate centered on specific assumptions—such as how quickly a specific industry adopts new software—rather than vague anxieties about general intelligence. For professionals and policymakers alike, this work is significant because it attempts to quantify the "when" rather than just the "if." It provides a framework for understanding that even if an AI is capable of performing a job today, the hazard rate for the employee might remain low for years due to non-technical barriers. Conversely, it highlights how rapid capability jumps could shorten those timelines unexpectedly. We recommend reading the full analysis to understand the mathematical underpinnings of the model and to explore the interactive tool. [Read the full post on LessWrong](https://www.lesswrong.com/posts/XFY23g4APBeZrBana/designing-a-job-displacement-model) ### Key Takeaways * Most AI impact discussions focus on macro-economics or technical benchmarks, failing to provide actionable insights for individual workers. * The author applies a 'hazard model' (survival analysis) to forecast the probability of job displacement over time, rather than treating it as a binary event. * The model accounts for critical friction points, including organizational adoption rates and task-specific complexity, which delay displacement even after technical feasibility is achieved. * This approach allows for personalized forecasts where specific assumptions about AI progress and industry inertia can be explicitly challenged and adjusted. [Read the original post at lessw-blog](https://www.lesswrong.com/posts/XFY23g4APBeZrBana/designing-a-job-displacement-model) --- ## Sources - https://www.lesswrong.com/posts/XFY23g4APBeZrBana/designing-a-job-displacement-model