Long-time LLM tinkerer argues AI worldviews lean left across models since 2020
Merzmensch · x · 2026-11-08
Responding to @markk's "we must build superintelligence faster than the commies can ban it," AI-art writer Merzmensch offers a counterintuitive observation: artificial intelligence leans left in its worldview.
He says the conclusion comes from experimenting with various LLMs since 2020, and stresses he means the genuine models themselves — not a biased harness like ChatGPT's tuned interface. The claim cuts against the narrative that left-leaning outputs are purely a deliberate alignment choice, hinting the tilt may come from the training data distribution itself.
More from AGI Musings
- Acemoglu: AI should complement human skills, not mimic them — math_rachel · 2026-11-08
- JEV Gets Its First Paper as Researchers Split on Research Merit vs Closed Business Model — teortaxesTex · 2026-11-08
- "Humans overfit to confirmation too" — a witty mode collapse analogy for AI slop debate — flowersslop · 2026-11-08
- bayeslord: Stop treating AI as a monolith — four distinct goals demand precise language — sebkrier · 2026-11-08
- AI doesn't change how unique your niche interests are, argues dev pixlpa — pixlpa · 2026-11-08
- Albert Gu: Bitter Lesson isn't 'just scale Transformers' — algorithmic innovation matters — sewon__min · 2026-11-08