Untrained networks already tell cats from dogs: architectures as evolutionary selection
joannejang · x · 2026-09-03
A blog post argues that today's neural network architectures can already distinguish cats from dogs even before training, suggesting that modern architectures may have been selected through an evolutionary process rather than being purely accidental designs.
More from Research
- Skeptic dissects Astra's rumored recurrent architecture: likely just looping each layer twice — mike64_t · 2026-09-03
- Cosmos grantee to spend 3 months probing LLM reasoning faithfulness — hunarbatra · 2026-09-03
- Video Delta Net speeds up open video generation 75-90x: 14s video in 11s — xiuyu_l · 2026-09-03
- CoreAuto team publishes new blog post on neural architecture discovery — _arohan_ · 2026-09-03
- Researcher claims they may have 'cracked RL' for their task — andrew_n_carr · 2026-09-03
- Judea Pearl fires back in causal inference debate: RCT's new language deserves open scrutiny, not hiding — yudapearl · 2026-09-03