AI Search: The Bitter-Er Lesson
Scaling large language models by adding explicit search over possible solutions—akin to how modern chess engines explore game trees—is being floated as a path toward far more capable AI, even “superintelligence.” Commenters probe whether this is technically and economically viable given enormous branching factors, weak general-purpose value functions, and current architectures, and raise concerns about safety, alignment, and overhyping AGI timelines. Many see search as essential for domains with clear evaluators (games, theorem proving, drug design), but doubt it can straightforwardly generalize to messy real‑world problems or bootstrap AI research itself.
What “search” means here
- Most commenters read “search” as classic AI tree search (minimax, MCTS, breadth/depth-first), not web search or RAG.
- For LLMs this would mean branching over candidate thoughts/solutions, evaluating them, pruning, then answering – akin to “pondering”.
Perceived promise of adding search to LLMs
- Could let models spend more compute on hard problems and less on easy ones.
- Might turn today’s “intuitive oracle” LLMs into explicit problem solvers that can revise and refine plans before replying.
- Some see this as a plausible path to much stronger systems or even “AI foom,” especially in domains with cheap, automated evaluation (games, theorem proving, fuzzing, some science tasks).
Limits: value functions and search spaces
- Strong objection: chess works because there is a well-defined state space and fast, good value function; real-world tasks and “AI research” do not.
- Value functions today are highly domain-specific; general ones are lacking and their feasibility is unclear.
- For broad domains (AI research, curing Alzheimer’s, “cure cancer”), the state space and transitions are themselves unclear.
Compute and practicality
- Tree search over token sequences is computationally enormous (branching factor in the tens of thousands at token level).
- Even coarse-grained idea-level branching could be very expensive; recent papers using search need drastically fewer rollouts than game AIs, suggesting cost pressure.
- Debate over train-time vs inference-time cost tradeoffs; 100–1000× inference cost may be unacceptable for many applications.
Alignment and superintelligence debates
- Some warn: anything that accelerates paths to superintelligence worsens alignment risks; article is criticized for ignoring “what to optimize for” and control.
- Others are skeptical that “superintelligence” is even a coherent or reachable concept, or see AGI as requiring multiple unknown breakthroughs and long timelines.
World models, generalization, and LLM limits
- Repeated concern that current LLMs lack robust world models and generalization; they remix text more than they reason.
- Without reliable internal models, search may just traverse a space of biased, sometimes false beliefs.
- Several argue we still need mechanisms to learn usable world models (e.g., from video, rich simulations, adjustable abstraction levels).
Symbolic vs statistical approaches
- Commenters note that classical search, planning, and theorem-proving already have near-optimal algorithms under known tradeoffs (soundness, completeness, efficiency).
- Some advocate hybrid neuro-symbolic systems where logic, simulators, or ontologies provide structure and evaluation, with LLMs as generators.