ml-capability-fragility-paradox

IN derived (depth 5)

Created 2026-06-21T10:13:04+00:00 · Reviewed 2026-06-21T15:37:01+00:00

ML's progress appears shaped by a tension between capability and fragility: hardware-theory co-evolution selects for pragmatic architectures that scale well on available hardware, and these same pragmatic design choices — favoring engineering expedience over biological fidelity — may contribute to characteristic failure modes like adversarial vulnerability. This suggests that the factors driving capability forward and those introducing fragility are related, though the evidence establishes correlation and plausible connection rather than a direct causal mechanism.

Justifications

SL — depth-5 synthesis — hardware-driven scaling and pragmatic shortcuts are the same evolutionary force viewed from capability vs. safety perspectives

Antecedents (all must be IN):

  • IN hardware-theory-coevolution-governs-all-scales — ML's trajectory has been significantly shaped by hardware-theory co-evolution at multiple scales — the macro deep learning revolution required convergence of bio-inspired architectures, mathematical foundations, and GPU compute (all three becoming available around 2012), while the micro-level sequence modeling arc (RNN→LSTM→Transformer) tracked a CPU→GPU hardware shift that favored parallelizable architectures, suggesting hardware availability is a major factor in selecting among theoretically viable approaches.
  • IN ml-pragmatic-shortcuts-create-fundamental-fragility — ML architectures succeed through pragmatic inductive biases rather than biological fidelity, and these architectures exhibit adversarial vulnerabilities absent in biological perception — suggesting that the engineering choices enabling ML progress may contribute to characteristic failure modes, though the antecedents do not establish a direct causal link between specific shortcuts and specific vulnerabilities.

Dependents

These beliefs depend on this one: