Behind the seamless chatbots, predictive diagnostics, and "smart" decisions shaping our world lurks an unseen force: a global army of annotators, linguists, and ethicists whose nuanced judgments train machines to mimic intelligence. Yet their stories-of burnout, breakthroughs, and ethical tightropes-remain buried under billion-dollar valuations.
The Cognitive Middle unveils this invisible infrastructure through the reconstructed saga of Merkor, a Bay Area startup that scaled from a midnight server meltdown to rejecting a $50M defense deal on principle. Synthesized from public patents, workforce surveys, and industry filings-no leaks, just pattern-revealing narrative-this book exposes:
Part field manual, part philosophical probe, this is essential reading for AI builders, policymakers, and anyone grappling with intelligence's true cost. In a 2025 world where AI ethics tops bestseller lists (think Atlas of AI meets The Alignment Problem), Crane delivers unvarnished truth: Machines learn from us, but only if we safeguard the middle ground.
Advance praise: "A vital human lens on AI's algorithmic facade." -Echoing Kate Crawford's planetary critique.