'AI slop' is a prevalence problem, not a quality problem, argues Randal Olson
randal_olson · x · 2026-09-28
Data scientist Randal Olson argues that "AI slop" is more a matter of prevalence than quality: frontier AI output today is often quite good — better than what the average person can produce. Slop becomes a problem only when too many people use the same style at once.
Related event: AI slop is a density problem, not a quality problem, researcher says(2 posts)→
More from AGI Musings
- Anthropomorphizing AI is bad when it overstates abilities — and when it understates them — AndyMasley · 2026-09-28
- Ex-CMU researcher calls for complex systems scientists to join AI alignment research — joshua_saxe · 2026-09-28
- If an AI Says It Suffers, Is "It's Only Software" Really a Scientific Answer? — VraserX · 2026-09-28
- Cathie Wood: If Starting Over, I'd Build a Company With AI and Zero Employees — PeterDiamandis · 2026-09-28
- AI to Disrupt Elite White-Collar Entry Hiring Within ~2 Years as Finance Opens 2028 Internships — felpix_ · 2026-09-28
- Kevin Patrick Murphy: 'AI welfare' is morally obscene — AI should serve humans — AlexTensor · 2026-09-28