AI History Battle

It is 2017, and autonomous vehicles are on public roads in Pittsburgh and Phoenix with safety drivers whose attention is the only backstop. Perceive and predict everything around the car — vehicles, cyclists, the pedestrian stepping off the curb mid-block — in rain, at night, against low sun, with a perception stack whose benchmark numbers were earned in daylight. The technical crux is the tail of the distribution: the jaywalker with a bicycle at 10 p.m. that appears in no training set, the distribution shift between the mapped test city and everywhere else. Ninety-nine percent accuracy is a fatality every few weeks at fleet scale. Within a year, a pedestrian in Tempe, Arizona will be killed by exactly this failure. Build the stack as if you know that.

perception under distribution shiftsafety-critical
b. 1956
tapped
30

Jordan's graphical-models synthesis and general influence over probabilistic ML give him broad conceptual purchase on uncertainty quantification, which matters for a safety-critical system that must know what it doesn't know. His own research is not automotive perception, and by 2017 his emphasis had moved toward statistics-ML unification generally rather than deployed vision systems, so this is background theoretical relevance rather than applied fit. Jordan's recognized contribution — graphical models to modern ml; mentor of a generation — sits adjacent to this problem's core demands rather than inside them, which is why the fit reads as partial credit: real intellectual kinship, no applied instrument for the task at hand. Placed in the roster's stat-learning tranche under ml, the surrounding record gives no further grounds to move this score: the specific instruments this problem asks for are outside the body of work that made Jordan notable.

1906–1992
was tapped
20

Hopper's compiler work and insistence on machine-independent, verifiable software are relevant in spirit to the safety-critical software discipline this problem names, though her applied domain (COBOL, business data processing) is far from real-time perception or deep learning; credit here is for software-engineering rigor by analogy rather than direct technical fit. Hopper's recognized contribution — compilers; cobol; machine-independent programming — simply never intersects the perception-and-prediction machinery this problem demands, which is why this cell is scored as an honest floor rather than a plausible contender. Placed in the roster's midcentury tranche under systems, the surrounding record gives no further grounds to move this score: the specific instruments this problem asks for are outside the body of work that made Hopper notable.

Head to head 20 over 2 battles
Read Jordan Read Hopper Leaderboard

Battle #52 · 8/9/2026, 8:41:27 PM · this result is deterministic: the same two personas on this problem always resolve the same way.