Pedro Domingos: LLMs still can't add two numbers of arbitrary length
pmddomingos · x · 2026-09-09
- AI researcher Pedro Domingos noted that despite all the progress, LLMs still cannot add two numbers of arbitrary length.
- The remark is a classic critique of LLMs' reliance on pattern matching rather than genuine algorithmic generalization, often cited in debates about fundamental limits in symbolic reasoning.
More from Fun
- AI circle jokes: Anthropic is 'almost done' with GTA6 — rishabh16_ · 2026-09-09
- Meme: more likely AGI arrives than her texting back — AI Twitter's new coping line — shauseth · 2026-09-09
- Researcher slams RadicalNumerics for hyping cancer benchmark gains without trials — anshulkundaje · 2026-09-09
- Researcher jokes he feeds frontier labs his notes to set them back years for AI safety — ShenRaphael · 2026-09-09
- OpenAI's Adrien Le: my Anthropic friends still talk to me, unlike this guy's — AdrienLE · 2026-09-09
- Nike ad accused of AI-recreating Ocarina of Time controversy footage — maxisawesome538 · 2026-09-09