Eye On A.I.

The episode contains a handful of genuinely interesting ideas - the evolutionary mystery of consciousness requiring strong emergence, the McClelland 'permanent agnosticism' paper, and Goff's skepticism about LLM sentience - but the density is low. Large stretches repeat the same panpsychism framing, and almost nothing is actionable for any B2B operator, even one thinking seriously about AI adoption.

“I think there's a real mystery of why consciousness evolved”

“an interesting paper by, um, Tom McClelland at Cambridge, it's just come out arguing for permanent agnosticism on a core question here”

Goff's inversion - starting with consciousness to account for physics rather than the reverse - and his evolution-based argument that consciousness must involve strong emergence are genuinely counterintuitive framings. However, panpsychism, the Galileo separation, and philosophical zombies are well-worn philosophy-of-mind fare; the originality is real but bounded by the field's existing literature.

“we're going to turn it upside down. We're going to start with consciousness and account for everything else, account for physics, chemistry, neuroscience, and in terms of consciousness”

“If consciousness brings with it new forms of causation, um, new forms of uh, causal operations that aren't completely reducible to known, understood, um, chemistry, neuroscience, um, if that were the case... that would be wonderful because you could then empirically see from the outside the markers of consciousness”

Philip Goff is a legitimate academic philosopher with real publications and a genuine intellectual research program - not a career podcast guest or pure thought-leader. However, 'seniority and relevance' to a B2B operator audience is essentially zero; he is a metaphysics professor and the rubric explicitly rewards operators who have done the thing at scale.

“I talk about this a lot in my, oh my, my book, modestly titled why the Purpose of the Universe that is just out in paperback”

“I was PhD examiner for a wonderful thesis”

Goff is surprisingly concrete for a philosopher: the Koch-Chalmers wager with its 25-year timeline and crate of fine wine, the named McClelland and Michel papers, the FMRI resolution figure, and multiple named researchers and theories give the episode more texture than typical armchair philosophy. The specificity is academic rather than business-relevant, and no hard empirical data support the central panpsychism claim.

“25 years ago, um, the neuroscientist Christoph Koch bet the, uh, bet the philosopher David Chalmers that this would all be wrapped up by now... a, uh, couple of summers ago, he, he publicly recanted... and gave David Chalmers the crate of fine wine he'd promised him”

“every, every pixel corresponds to what is a 10,000 neurons or something”

The host occasionally lands a productive challenge - pushing back on the ant-consciousness interpretation and the 'universe is not a biological system' objection - but questions are routinely long-winded, self-answering, and leading, giving Goff little to push against. There is no sustained productive disagreement on the central claims, and the host frequently forgets his own question mid-sentence.

“When you talk about um, new forms of, of um. I can't remember the way you phrased it, of causality within the brain.”

“I mean, the universe is not a biological system, right. If consciousness for higher forms of biology is more developed than say, a bird compared to a human, how do you feel about the proposal”