For the past 24 hours the following story has been in the news:
“Anthropic finds hidden ‘workspace’ inside Claude Al. Researchers say the structure holds unspoken concepts during reasoning, mirroring a leading neuroscientific theory of human consciousness.”
Two subjects I’ve long been interested in. (106 blog posts tagged “consciousness,“ 46 tagged “thoughts.“) I prompted Claude identify my blog posts that seem most relevant to this story?
Strongest match — your 2013 post on Michael Graziano’s Consciousness and the Social Brain (around line 49750, URL smays.com/2013/10/consciousness-and-the-social-brain/). Graziano’s Attention Schema Theory is a close cousin of Global Workspace Theory — both describe a privileged internal model that represents what the brain is doing to itself. The quotes you pulled are uncanny in light of the J-space finding: “The unconscious machinery of the brain is so vast that it is like an elephant. Perhaps consciousness is a little boy sitting on the elephant’s head.” And most directly: “We will build computers that can construct their own awareness in the same way that the human brain does.” You were sitting with this exact question twelve years ago.
Second — your Sam Harris free will post (around line 46740). The quotes you curated there — “The intention to do one thing and not another does not originate in consciousness — rather, it appears in consciousness” — describe precisely what J-space makes visible: reasoning that happens before and beneath the verbalized output.
Third — your Nisargadatta quotes (around line 47284). “A tremendously complex work is going on all the time in your brain and body, are you conscious of it? Not at all. Yet for an outsider all seems to be going on intelligently and purposefully.” That’s a near-perfect lay description of what Anthropic found: most of what Claude does proceeds without anything touching the J-space workspace, and the system functions fine — until you need flexible reasoning.
Fourth — Susan Blackmore’s *Ten Zen Questions (around line 48160). “There are multiple brain processes going on, some of which take up more of the brain’s capacity than others, but there is no me who experiences them, and no time at which they become conscious.” That’s the global workspace structure in Buddhist drag.
The irony worth noting: the Anthropic paper uses the word “conscious” over 200 times while carefully avoiding claiming Claude is conscious — which is exactly the epistemic position you’ve held about consciousness generally for years. The unknowability argument you’ve made in AI consciousness conversations applies here too: finding a functional analog to the global workspace doesn’t settle whether there’s anything it’s like to be Claude, any more than identifying the neural correlates of consciousness in humans settles the hard problem.