David Silver's Ineffable Bets on RL Over LLMs — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- David Silver co-founded Ineffable to build an AI 'super learner' using reinforcement learning, aiming to create systems that can teach themselves without relying on large language model architectures
- Ineffable is pursuing a path distinct from dominant LLM-based AI development, drawing on Silver's work at DeepMind where he helped create AlphaGo, the first AI to beat a top human Go player
- David Silver emphasizes self-teaching AI systems, noting that reinforcement learning enables machines to learn from experience rather than from vast datasets of human-generated text
- Ineffable's approach positions reinforcement learning as a viable alternative to LLMs, with potential advantages in efficiency and generalization, according to Silver's vision outlined in the interview
Why it matters: This shift toward reinforcement learning as a core AI paradigm could reduce reliance on data-heavy models, affecting how startups and labs allocate resources; if scalable, Ineffable's method might lower barriers for AI development where labeled data is scarce, challenging the current dominance of LLM-focused firms.
Ask SkimNews
