Elevator Pitch

• The author argues that LLMs will primarily make everyday systems—customer support, pricing, policing, and purchasing—more frustrating by adding lying automation and diffusing accountability.

Key Takeaways

• Companies will route more support and “fuzzy” decisions to LLMs because they only need to be cost-effective, not correct, making it harder to reach empowered humans.
• People will spend increasing time “arguing with machines” across pricing, insurance, jobs, and bureaucracy—sometimes using their own LLMs—creating an arms race of drudgery.
• ML systems will further diffuse responsibility: decisions become illegible, multi-organization outputs where “no one is directly responsible,” even when harms are severe.

Most Memorable Quotes

• “The latest crop of machine learning technologies will be used to annoy us and frustrate accountability.”
• “I expect we’ll spend more of our precious lives arguing with machines.”
• “I think ML models will further diffuse responsibility, replacing judgements that used to be made by specific people with illegible, difficult-to-fix machines for which no one is directly responsible.”