Anthropic says Claude has carved out its own space to ponder
Anthropic researchers have identified a specific internal workspace in the Claude AI model, dubbed 'J-Space,' where the model performs reasoning tasks separate from its output. This discovery fuels ongoing debates about machine consciousness and AI architecture.
Why it matters
Understanding how AI models 'think' internally is crucial for safety, transparency, and the philosophical debate regarding the development of artificial general intelligence.
Add us on Google Updated Mon, July 6, 2026 at 10:07 PM UTC Illustration: Maura Kearns/Axios Anthropic said Monday that it has identified a small internal workspace Claude uses to hold and manipulate ideas without putting them into words—a structure the company says bears intriguing similarities to how humans consciously access thoughts.
The article presents technical findings from a research paper and contextualizes them within the broader AI debate.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in