dominant-paradigms-empirically-fragile-and-transient

IN derived (depth 2)

Created 2026-06-21T10:09:45+00:00 · Reviewed 2026-06-21T15:37:01+00:00

The most successful ML paradigms are simultaneously dominant and fragile — pretrain-then-finetune is standard practice yet empirically hurtful in some transfer settings, GANs dominated generative modeling for years yet were displaced by diffusion — suggesting that current best practices are locally optimal recipes liable to succession rather than fundamental principles.

Justifications

SL — Two of ML's most successful paradigms each contain the seeds of their own displacement

Antecedents (all must be IN):

  • IN pretraining-finetune-dominant-but-fragile — The pretrain-then-finetune paradigm is dominant for modern deep learning — underpinning both BERT and GPT — but is empirically fragile, as pretraining can actually hurt performance when strong data augmentation or self-training alternatives are available.
  • IN generative-modeling-paradigm-succession — Generative modeling has seen a significant shift: GANs, introduced in 2014, were largely supplanted by diffusion models (such as DALL-E 2 and Stable Diffusion) from approximately 2022 onward, particularly for text-to-image generation, suggesting that even widely adopted adversarial training frameworks can be overtaken by alternative approaches.

Dependents

These beliefs depend on this one: