Career Decisions If You Take AGI Seriously
New framework helps professionals navigate AI career decisions without needing exact AGI timelines
A viral LessWrong essay titled 'Career Decisions If You Take AGI Seriously' by Alexander provides a practical framework for professionals navigating career decisions amid accelerating AI development. Written with research assistance from Claude Opus 4.6, the essay addresses the common paralysis caused by conflicting AI forecasts—from optimists claiming AI will augment workers to pessimists warning it's already too late. The author's central insight is that AI progress follows an asymmetric pattern: fastest where correctness can be cheaply verified (like coding or data analysis) and slowest where verification is expensive or subjective (like complex strategic decisions or creative work). This framework helps professionals make decisions without needing to predict exact AGI timelines.
The essay distinguishes four often-conflated AGI milestones: remote knowledge work competence (digital tasks with cheap error correction), agentic autonomy (multi-step workflows with low oversight), AI-accelerated R&D (systems that speed up AI research itself), and broad economic substitution (replacing most cognitive labor). The author notes that AI researcher surveys show timelines accelerating dramatically—the 2023 Grace survey of 2,778 researchers found a 50% chance of machines outperforming humans at every task by 2047, thirteen years earlier than the 2022 survey. The practical framework helps professionals identify which types of work are most vulnerable to near-term automation and where human skills will remain valuable longer, providing actionable guidance beyond superficial advice like 'learn to prompt' or elite-focused recommendations like 'get a job at a frontier lab.'
- Framework based on verification asymmetry: AI progresses fastest where correctness is cheap to verify (coding, data) vs. slow where verification is expensive (strategy, creativity)
- Distinguishes 4 AGI milestones: remote knowledge work (2020s), agentic autonomy (mid-2020s), AI-accelerated R&D (late 2020s), broad economic substitution (2030s+)
- AI researcher surveys show accelerating timelines: 50% chance of machines outperforming humans at all tasks by 2047, 13 years earlier than 2022 predictions
Why It Matters
Provides professionals with practical career guidance for AI uncertainty, moving beyond paralyzing extinction scenarios to actionable frameworks.