Anthropic finds a global-workspace analogue in Claude: J-space
On July 6, 2026, Anthropic published research reporting a analogue of a "global workspace" in Claude, which it calls J-space. Classical consciousness theory describes the structure of the finding, but it says nothing about the key point: Claude's workspace is made of words, not abstract signals. The finding is described on the company's blog and has already prompted discussion among researchers who study the internals of large language models.
AI-processed from Habr AI; edited by Hamidun News
On July 6, 2026, Anthropic researchers published the results of new research in which they claimed to have found a structure in the Claude language model that resembles a "global workspace": J-space. The finding is described on the company's blog and has already sparked discussion in the community that studies the internals of large language models.
What is J-space
According to the researchers, J-space is a kind of internal working area in Claude through which information passes as the model forms a response. Structurally, it resembles Global Workspace Theory, one of the best-known concepts in the science of consciousness, which describes how scattered processes in the brain come together into a single stream of information processing.
- Anthropic's research was published on July 6, 2026
- The finding is called J-space
- Claimed analogy: the global workspace (Global Workspace Theory)
- Key feature: the workspace is made of words, not abstract signals
Why classical theory does not explain the finding
Global Workspace Theory describes the structure well: there is a central hub where information from different sources converges, and from there it spreads further. But the theory, developed to describe the human brain, says nothing about why a language model's workspace is built from words rather than from more abstract internal representations.
According to the author of an analysis on Habr, this gap was closed a year earlier by the Soviet psychologist Lev Vygotsky. In the first half of the 20th century he described how human speech and thought intertwine, and how inner speech becomes a tool for organizing thinking. By this logic, Vygotsky's work on the role of the word in thought explains Anthropic's finding better than later neuroscientific theories of consciousness, including the model of Bernard Baars, on which Global Workspace Theory is built.
What this means
The discovery of J-space is one more step toward understanding how the "thinking" of large language models works from the inside, and not only from their outputs. If a model's workspace really is organized around words, that may explain both the strengths and the weaknesses of current LLMs, and may also suggest directions for the interpretability and control of future systems.
- Nov 1, 2026OpenAI Rewrites GPT Behavior Rules: Five Values and a Chain of Command
- Nov 1, 2026Google admits AI Mode cuts search revenue, shares fall 15%
- Oct 31, 2026Author puts his script before an AI jury: the verdict cost 850 thousand tokens
- Oct 31, 2026Anthropic: Fold Subagents Back Into a Single Loop — One Customer's Tasks Ran Twice as Fast
Key Questions Answered
What is J-space?
According to the researchers, J-space is a kind of internal working area in Claude through which information passes as the model forms a response. Structurally, it resembles Global Workspace Theory, one of the best-known concepts in the science of consciousness, which describes how scattered processes in the brain come together into a single stream of information processing. Anthropic's research was published on July 6, 2026 The finding is called J-space Claimed analogy: the global workspace (Global Workspace Theory) Key feature: the workspace is made of words, not abstract
Why classical theory does not explain the finding?
Global Workspace Theory describes the structure well: there is a central hub where information from different sources converges, and from there it spreads further. But the theory, developed to describe the human brain, says nothing about why a language model's workspace is built from words rather than from more abstract internal representations. According to the author of an analysis on Habr, this gap was closed a year earlier by the Soviet psychologist Lev Vygotsky. In the first half of the 20th century he described how human speech and thought intertwine, and how inner sp
What this means?
The discovery of J-space is one more step toward understanding how the "thinking" of large language models works from the inside, and not only from their outputs. If a model's workspace really is organized around words, that may explain both the strengths and the weaknesses of current LLMs, and may also suggest directions for the interpretability and control of future systems.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.