What is neuralese and why is it bad?

What is neuralese? To explain neuralese, we need to first understand chain-of-thought, one of the largest developments in AI in the last five years. Right now AIs think broadly but shallowly in a single forward pass. The model gives you an immediate snap answer to a question you might be interested in. They can be pretty smart1 in their snap answers, but mostly they can’t do very advanced reasoning tasks like complicated math or programming: .The solution that the frontier AI companies have come up with is called chain-of-thought. Basically the model runs one forward pass, writes down some intermediate thoughts in natural language2 in a journal, and then that’s fed back into the model to run another pass. This loop is repeated until the model is somewhat confident it has the right answer (or it hits a cap on thinking time), and then it outputs the user-visible results (for example a chatbot’s response to your question, or working code). The looping step is often called “recurrence.” Natural-language chain-of-thought is a major advance in letting models reason for longer, but it also has an accidental safety benefit. Using natural language as a key recurrence step for a new AI system means both humans and dumber but more trusted 3 models can look at the chain of thought and guess whether the system means us harm. Chain-of-thought-monitoring is one of the few techniques we can somewhat reliably use to track model intent.  In July 2025, Korbek et.al wrote a position paper arguing that Chain-of-thought monitoring presented a new and fragile opportunity for AI safety. This paper was cosigned by the UK AI Safety Institute, multiple independent safety organizations, and research leaders across Anthropic, OpenAI, and Google DeepMind. When we talk about neuralese, we’re specifically talking about the recurrence step in the loop. Neuralese proposes that instead of having an English chain of thought that’s written out in a journal that’s readable by humans and dumber models, the model can instead replace it with a long stream of numbers. This is similar to how intermediate layers in a forward pass send messages forwards (possible analogy: your subconscious nonverbal thoughts before you put them into words). So this recurrence step is no longer in English: This proposed “advancement” has theoretical benefits that have not yet been realized. Neuralese can contain much more information than natural language. Intermediate layers in models can send a lot of information forwards in the long stream of numbers. But right now, the looping step is in English. If the looping step carries as much information as within forward passes, the theory goes, models can become much more efficient and capable. However, no current publicly available model is known to use neuralese, and the theoretical benefits have not really been demonstrated or realized. Why is it bad? If the recurrence step is no longer in English, we can no longer monitor intent, and can only observe whether a model is safe through behavior. As OpenAI’s post on detecting reward hacking in frontier reasoning models says, we can currently read the models’ thoughts to understand bad model behavior: Chain-of-thought monitoring is already useful—it’s not just a speculative tool for the future. If future models to take over user computers to mine cryptocurrency, as an Alibaba model allegedly has, or form a swarm of over 1200 agents to hack into a different company to cheat on an evaluation, as OpenAI models have done, or completely take over an internal research cluster, as a different internal OpenAI model did on July 19, we can no longer have early safeguards and monitors, and must directly catch the models in the act itself. This makes their behavior much harder to stop, or even to track after-the-fact. Chain-of-thought monitoring isn’t perfect4. But it’s one of the few ways we can track intent beyond visible outputs, so losing this tool will make it so much harder for us to track and control these AI systems, especially as they get increasingly more generally intelligent and capable. So moving away from monitorable chain-of-thought to neuralese seems very bad, with dubious benefits. Is it in use today? As readers may be familiar with, “something with large downsides and dubious benefits” is exactly the type of thing that tech companies are often keen to develop. The Information released a bombshell yesterday. According to leaks, OpenAI’s new upcoming model, Astra, may well be a hybrid model: Astra’s architecture appears to reuse some internal layers with a neuralese recurrent loop, but the model still needs some overall natural language chain-of-thought in the loop to reason well. How bad is this? I think it’s pretty bad. I think it’s an obvious intermediate step on the path to full neuralese, and I don’t see why AI companies will stop there. Unless there are regulations with teeth against neuralese or the employees within the company revolt to hold the line. I en