Deep learning 'never tried to explain neural networks', piling up conceptual debt
gerardsans · x · 2026-11-19
In an X exchange, data-visualization researcher Gerard Sans argues that public perception of deep learning leans too heavily on chatbots and off-the-shelf generative media models, while the field's real capability is far broader than scaling LLMs into chat/agent interfaces.
He frames deep learning as a new way of engaging with data and urges everyone to build simple networks to understand it. He also criticizes the field for never seriously trying to explain neural networks, leaving misconceptions unchallenged and accumulating 'conceptual debt.'
More from AGI Musings
- The case for externalized memory as the real fix for continual learning — daniel_mac8 · 2026-11-19
- Stanford professor: AI takes students from problem to working prototype in 80 minutes — HariSubramonyam · 2026-11-19
- Berggruet Winner Thomas Nagel Dismisses AI Consciousness Talk as 'Just Ridiculous' — Dr_Atoosa · 2026-11-19
- Dwarkesh Claims China Is Only 12 Hours — Not 12 Months — Behind Frontier AI Labs — teortaxesTex · 2026-11-19
- Researcher calls METR/Anthropic '2x speedup' RSI analysis statistically dubious — jessi_cata · 2026-11-19
- Physician: AI competition now measured in weeks, GPT-6 versions 7 days apart — ramos_casals · 2026-11-19