How NOT to lose your job to AI (article by Benjamin Todd)

Current AI Capabilities and Automation Risks

• Half of the population fears job loss to AI, a concern grounded in current capabilities: GitHub code completion, photorealistic video generation, safe autonomous driving, and accurate medical diagnosis.
• Mass automation and falling wages are projected as real possibilities within the next five years as AI capabilities improve rapidly.
• While automation drives down the value of skills AI can perform, it simultaneously drives up the value of skills AI cannot yet replicate.
• Historically, wages may increase before falling, as automation generates wealth and leaves remaining human tasks as the bottleneck for growth.

Economic Mechanics: The ATM Parable and Automation Cycles

• Partial Automation Effect: Similar to ATMs reducing bank clerks per branch (21 to 13) but increasing total clerks for two decades by lowering branch costs and enabling expansion, AI can initially increase employment by boosting productivity.
• Thorough Automation Effect: Eventually, if automation becomes wholesale (e.g., online banking), employment declines as the human element is removed.
• Competing Forces: AI productivity gains increase employment, while wholesale replacement decreases it; with medium automation, prediction is difficult, but with thorough automation, displacement tends to win.
• Aggregate Wealth Shift: The Industrial Revolution shifted the majority of labor from agriculture to higher-paying jobs, increasing aggregate income 100-fold while automating the bulk of production tasks.
• Remote Work Automation Scenario: Epoch AI estimates that automating 100% of remote work (approx. one-third of tasks) could increase economic output 2–10x, potentially raising wages for non-remote tasks by the same multiplier.
• Full Automation Risks: Models suggest that if 100% of tasks are automated, wages could plummet below subsistence levels due to an infinite labor supply; if only 1% remains for humans, wages could theoretically rise indefinitely.
• Timeline Speculation: Under default assumptions, wages may rise tenfold, then crash in the late 2030s as the final human bottlenecks are removed.

Four Categories of Skills Likely to Increase in Value

• Data-Poor, Messy, Long-Horizon Tasks: Skills requiring judgment over years (e.g., entrepreneurship, managing complex organizational strategy) are harder to automate than short, well-defined tasks because they lack massive training datasets and clear objective verification for reinforcement learning.
• AI Deployment and Oversight: Skills needed to organize, audit, and direct AI systems (e.g., system design, error checking, user experience) become more valuable as AI becomes a force multiplier, often resembling traditional management roles.
• High-Elasticity Output Sectors: Skills producing goods or services where demand scales with wealth (e.g., advanced healthcare, luxury goods, personalized education) will see wage growth as automation drives down costs and concentrates wealth.
• Inelastic Labor Supply: Highly specialized expertise (e.g., top-tier R&D, novel conceptual insights) where the labor pool cannot easily expand to meet demand will command premium wages.

Specific High-Value Work Skills

• AI Application: Learning to deploy AI to solve real-world problems by using cutting-edge models to achieve outcomes in current roles or startups.
• Personal Effectiveness: General productivity (goal setting, focus), social skills (relationship building, coordination), and "learning how to learn" (rapid retraining using AI tutors).
• Leadership and Strategy: Entrepreneurship (spotting ideas, coordinating resources), management (overseeing AI teams, bearing liability), and strategic decision-making (prioritization, vision setting).
• Taste and Communications: Judgment regarding design, beauty, and branding, as well as storytelling and personal connection, which remain valuable even as content creation becomes automated.
• Government and Policy: Skills to navigate policy implementation, political strategy, and government decision-making, sectors likely to remain slow to automate and human-centric.
• Complex Physical Skills: Precise physical tasks in unpredictable environments (e.g., surgery, data center electrical work, robotics maintenance) where robotics lag behind software AI.
• Complementary Technical Fields: Machine learning, information security, robotics development, and infrastructure construction (data centers, power plants).

Skills with Uncertain or Declining Future Value

• Routine White-Collar Work: Entry-level positions in finance, law, government, and healthcare involving established knowledge recall, routine writing, admin, and analysis face high automation risk, potentially leading to a "pyramid replacement" structure.
• Coding and Applied STEM: While AI makes coding easier to learn, the long-term value of years spent mastering coding may decline as AI surpasses humans at complex tasks; the skill may shift toward managing AI rather than writing code.
• Visual Creation: Roles in animation, special effects, and graphic design face layoffs as AI achieves photorealistic video generation, though demand remains for human oversight and direction.
• Routine Physical Skills: Jobs like driving, which are becoming automatable via robotics (e.g., self-driving taxis), face potential mass layoffs in the coming five years as reliability increases.

Strategic Career Recommendations

• Avoid Long Training Cycles: Be cautious with long-term education (e.g., PhDs) given the speed of AI change; prioritize shorter, high-impact skill acquisition unless pursuing elite expertise in rapidly growing fields.
• Target Small, Growing Organizations: Roles at AI application startups or growing firms offer faster learning in leadership, AI deployment, and entrepreneurship compared to specialized entry-level roles in large corporations.
• Leapfrog Entry-Level Paths: Acquire leadership and communication skills immediately via side projects or current roles to bypass traditional training ladders that AI is eroding.
• Build Resilience: Diversify geographic exposure, save aggressively, and invest in mental health to buffer against economic disruption and rapid skill obsolescence.
• Adopt a "Ride the Wave" Mindset: Continuously monitor AI capabilities and adjust to the current human bottleneck rather than seeking a single permanently immune job.