Another way to intelligence
A dog learns to fetch by repetition.
A human learns calculus once, from a book, on about 20 watts.
Today's AI runs almost entirely on the first kind of learning.
The big labs teach machines by repetition at enormous scale. Nature shows it is not the only route to intelligence. A short scroll, by Semantic Reasoning Labs.
scroll
> 01
How you train a dog
Training a dog to fetch, in a few steps:
- The dog does something.
- If it gets closer to the desired behavior, it gets a reward (for example a cultivated meat treat).
- Repeat many, many times, and the desired behavior emerges.
That is the essence of reinforcement learning. The deeper pattern (guess, get feedback, adjust a little, repeat at enormous scale) is the basic training loop behind much of modern AI. An LLM learns by guessing the next word, comparing the guess with the real text, and nudging billions of weights by a tiny amount, over trillions of words. (Formally, that stage is called self-supervised learning, but the core loop holds: a positive response leads to a small adjustment.)
And notice what this kind of learning builds: intuition. Fast, automatic, hard to explain.
References: Wikipedia: Large language model · Hugging Face: Illustrating RLHF
> 02
Humans learn this way too
Humans learn this way too.
It is how we learn for example motor skills (riding a bicycle), speech (especially when you learn intuitively as a child), and many forms of innate knowledge we call intuition.
The thinking it produces is fast and effortless. You cannot explain how you keep balance on a bicycle. The knowledge is in the wiring, not in words.
But it is not the only way we learn as humans.
> 03
But calculus is learned once
Humans do not learn mathematics the same way dogs learn to fetch.
Once we understand calculus, a new piece of knowledge does not need to be repeated many times.
If the rule fits into our existing mental model, we can use it immediately. If you already understand symbolic integration, learning the integral of a particular function often takes a single exposure.
Practice still matters. But to figure out all the implications of the new piece of knowledge, not to learn it in the first place.
This is a different kind of learning. One exposure, not millions.
And it powers a different kind of thinking: deliberate, step by step, every step something you can write down, check, and defend.
Neuroscience sees the same split. Repetition slowly trains procedural circuits, while a new insight can be encoded by the hippocampus in a single exposure.
References: Learning, Fast and Slow in the Hippocampus
> 04
The hard problems live on the calculus side
Math, science, law, strategy, and many hard problems are much closer to this way of thinking and learning than to learning how to fetch.
Our mental model of the domain is built from explicit concepts, rules, and relationships we can fully articulate. If a legal contract is wrong, it usually does not "intuitively feel wrong". It is a particular clause, or a single word, that implies undesirable outcomes or uncertainties.
These are also the domains where 95% accuracy is usually not enough.
References: Wikipedia: Dual process theory · Thinking Fast and Slow in AI
> 05
The brain is the existence proof
Today's leading models learn even the calculus side the way the dog learns to fetch, at a scale no dog has ever seen. Meta trained Llama 4 on a mix of more than 30 trillion tokens.
A student learns calculus from one textbook. She brings years of schooling and a brain that evolution took millions of years to shape, and that whole brain runs on about 20 watts. Over an entire semester it uses the energy a single Nvidia B200 draws in about two days.
That a brain can do this means that learning structured knowledge from little data, on little energy, is physically possible. Humankind does not yet know how to do it in a machine. But it can be done.
References: Brain power, PNAS 2021 · Meta: The Llama 4 herd, 2025 · Epoch AI: Frontier data center power, 2026
> 06
The other route was tried before
Building intelligence from explicit knowledge is not a new idea. Symbolic AI, which writes concepts, rules, and relationships down directly, was the mainstream bet of the 1980s. Expert systems boomed, and two thirds of Fortune 500 companies came to apply them in daily business.
But symbolic AI alone was brittle. The world is messy. Symbols are clean. Going directly from the real world to symbols broke outside of very narrow domains.
What was missing was a bridge. Embeddings and modern neural networks, especially LLMs, now provide one: a robust way from the messy world into clean symbols.
DeepMind's AlphaGeometry pairs a neural model with a symbolic deduction engine and approaches the average score of IMO gold medallists on olympiad geometry problems. Its successors reached silver-medal standard on the live 2024 olympiad.
References: Wikipedia: Expert system · Wikipedia: AI winter · AlphaGeometry, Nature 2024 · DeepMind: silver-medal standard at IMO 2024
> 07
Why another kind of AI must be possible
That is how humans think: they combine both. Intuition reads the messy world. Deliberate reasoning builds on what we understand.
An AI built on both kinds of learning would keep what intuition does well, and add knowledge you can inspect, reasoning you can check and repeat, learning from single examples on far less compute.
An LLM can follow a new rule too, if you paste it into the prompt. However, the rule lives in the context window, not in the model, and is forgotten the moment it falls out. Learning a rule once means the model itself changes. Permanently, and the same mistake does not come back.
References: IBM Research: Neuro-symbolic AI · MIT 6.S191, Lecture 7: Neurosymbolic Hybrid AI
Further reading
- Learning, Fast and Slow: Single- and Many-Shot Learning in the Hippocampus (Annual Review of Neuroscience, 2024): the brain itself uses two learning mechanisms; the hippocampus can encode a new experience in a single exposure, while procedural circuits need many repetitions.
- Thinking Fast and Slow in AI (AAAI, 2021): IBM researchers propose dual-process cognition, System 1 and System 2, as a research agenda for AI.
- CYC: Using Common Sense Knowledge to Overcome Brittleness (AI Magazine, 1985): the original bet of the symbolic era, a large explicit body of world knowledge so that AI could reason instead of breaking outside narrow domains.
- AlphaGeometry (Nature, 2024): a neural language model guiding a symbolic deduction engine solves 25 of 30 olympiad geometry problems, approaching the average score of IMO gold medallists.
- AlphaProof (Nature, 2025): reinforcement learning with formal proof search reaches olympiad-level theorem proving; with AlphaGeometry 2 it scored at silver-medal standard on the 2024 IMO.
- Logic-LM (EMNLP, 2023): an LLM translates a problem into symbolic form and a deterministic solver does the reasoning; on five logical reasoning benchmarks this beats chain-of-thought prompting by 18.4% on average.
- Think-on-Graph (ICLR, 2024): an LLM explores a knowledge graph step by step instead of answering from memory; every answer is traceable to a graph path, and wrong knowledge can be corrected without retraining.
- Reasoning on Graphs (ICLR, 2024): relation paths grounded in a knowledge graph serve as faithful, interpretable reasoning plans for an LLM.
- Automatic Ontology Construction Using LLMs as an External Layer of Memory (2026): an LLM extended with a persistent, verifiable knowledge graph as its memory instead of relying on model weights and vector retrieval.
- ReKG-MCTS (ACL, 2025): Monte Carlo tree search explores a knowledge graph while an LLM guides the search, with no training required.
- Neuro-Symbolic AI in 2024: A Systematic Review (2025): the field is growing rapidly; explainability and trustworthiness remain open research areas.
- IBM Research: Neuro-symbolic AI: frames neurosymbolic AI as the path to combining statistical learning with humanlike symbolic reasoning.
- MIT 6.S191, Lecture 7: Neurosymbolic Hybrid AI (2020, course schedule): why hybrid systems improve sample efficiency, interpretability, and robustness.