AI Part 2: Fakery, Friction, and Flaws
As AI infiltrates creative industries, the line between human craft and algorithmic 'fakery' begins to dissolve. This analysis examines the friction between STEM-driven automation and the humanities, questioning whether our cultural output can survive the transition to a machine-led economy without losing its essential human friction.
Compound Engineering: Make Every Unit of Work Compound Into the Next
Software development is shifting from manual coding to 'compound engineering,' where AI agents autonomously execute and review features. By moving the human role from creator to architect and reviewer, this new paradigm promises to collapse the cost of production while fundamentally altering the demand for traditional engineering labor. Can a system that improves itself with every unit of work redefine the very nature of technical contribution?
Mythos, Muse, and the Opportunity Cost of Compute
As AI development costs skyrocket, the industry faces a brutal trade-off between frontier research and enterprise profitability. This analysis explores how the 'opportunity cost of compute' forces a choice between building the next digital muse and serving corporate automation, while open-source models threaten to commoditize the very intelligence frontier labs are trying to monopolize.
Demis Hassabis: Why AGI is Bigger than the Industrial Revolution & Where The Bottlenecks in AI Are
Demis Hassabis argues that AGI represents a shift more profound than the Industrial Revolution, potentially decoupling productivity from human labor. By examining current technical bottlenecks, the discussion explores how a world of abundant intelligence will necessitate a total reimagining of our economic and social structures.
Software engineering reaches its terminal velocity as systems hit one million lines of code with zero human intervention. By processing a billion tokens daily without human review, this 'harness engineering' paradigm signals the total automation of the developer lifecycle. Is the era of the human programmer already over?
How Will AI Change Work? Forecast the Future of Labor Automation — $35,000 in Prizes