Alexey Tolchinsky

This is not good.

From complementarity to non-duality: Seeing objects and processes as pragmatic constructs

Replies to the comments on
www.sciencedirect.com/science/arti...

is now out here:
www.sciencedirect.com/science/arti...

"From complementarity to non-duality: Seeing objects and processes as pragmatic constructs"

Source

Hoel, E. (2025). A Disproof of Large Language Model Consciousness: The Necessity of Continual Learning for Consciousness. arXiv preprint arXiv:2512.12802.

arxiv.org/pdf/2512.12802

#LLM
#continuous
#learning
#zero-shot
#consciousness

This is good news actually and it is related to zero-shot learning. I do think that systems that @doctorjosh.bsky.social 's lab and @drmichaellevin.bsky.social's lab are building are better in that regard. Fun things will come up from this exciting collaboration.

LLMs are NOT continuously learning systems, unlike a human child.Erik does not say that if you are a continuously learning system you have consciousness. What he is saying is that if you are not a continuously learning system, you can't be conscious.

There is a related function that Erik Hoel writes about. Continuous learning. LLMs - once you train them - have fixed weights. When we hear these rumors of "theft" of LLMs - this is what we are actually talking about - the weights are stolen.

How living organisms outperform robots in zero-shot learning | Alexey Tolchinsky, Psy.D. posted on the topic | LinkedIn Zero-shot learning as one of the main differences between a living organism and a robot. A human child is capable of zero shot learning (deciding what to do and ultimately completing the task in cir...

www.linkedin.com/posts/alexey...

Secret Sauce

This is a follow-up from a previous post related to Joshua Bongard's work that human babies have zero-shot learning abilities and no LLM is capable of that - in areas where it hasn't been trained it's quite helpless.

Hoel, E. (2025). A Disproof of Large Language Model Consciousness: The Necessity of Continual Learning for Consciousness. arXiv preprint arXiv:2512.12802.

arxiv.org/pdf/2512.12802

#AI
#LLM
#prompt

we interpret the prompt and if changes in the prompts are minor, then "it means" the same thing to us - it does not for LLMs. one of of the sources:

When you adjust a prompt with small, semantically irrelevant change and get a very different answer from the LLM, sometimes diametrically opposite answer - that's prompt sensitivity. This is not so with a human -

A lesser known flaw - prompt sensitivity

We talk much about the "hallucinations" of AI, which is unfortunate term that humanizes them (@anilseth.bsky.social's point, a better one would be confabulation.

What we talk less about is - prompt sensitivity.

Meet three scientists who said no to Epstein The warning signs included a web search, a mother’s doubts, and inklings of a “sexist attitude”

these 3 said No.

www.science.org/content/arti...

New preprint with Léo Pio-Lopez:
www.preprints.org/manuscript/2...

"Multi-Scale Longevity: Defeating Aging from Cells to Embodied Human Minds, and the Future of the Species"

a broader view of longevity research.

𝗪𝗵𝗮𝘁 𝗶𝘀 𝗰𝗼𝗺𝗽𝘂𝘁𝗮𝘁𝗶𝗼𝗻 𝗶𝗻 𝗱𝘆𝗻𝗮𝗺𝗶𝗰𝗮𝗹 𝘀𝘆𝘀𝘁𝗲𝗺𝘀?
Interesting paper tackling this difficult question.
Answer (in part): it's complicated!
The same fixed system can be interpreted as implementing many (even infinitely many) computations depending on the chosen decoding map.
doi.org/10.1088/2632...

A perspective, in 2 parts:

link.springer.com/article/10.1...

link.springer.com/article/10.1...

Mind Everywhere: A Framework for Conceptualizing Goal-Directedness in Biology and Other Domains—Parts One and Two

🧪

so sorry to hear you were hurt

If you listen to the brain, it can tell you how to do energy-efficient computing.
Brain-Inspired Energy Efficient Technologies for Next-Generation Artificial Intelligence
par.nsf.gov/biblio/10666...
#neuroscience

Why does time make sense to us? Chris Fields made an important point about how we measure things.

How we measure things and why does time make sense to us

alexeytolchinsky.substack.com/p/why-does-t...

My New Org to Solve Consciousness (or Die Trying); A Rogue AI Community That Wasn't; David Foster Wallace Being Weirdly Prescient; Cow Tools, & more Desiderata #38: links and commentary

Erik Hoel's new lab.
(I am a fan of his work)

open.substack.com/pub/erikhoel...

Join the Lab | Beatricebeebe | Beatrice Beebe I am currently recruiting new volunteers for Fall 2026! A volunteer position in our lab is ideal for students who are between college and graduate school, who are interested in research experience, ...

Beatrice Beebe is recruiting volunteers

www.linkedin.com/posts/beatri...

Eric Spiegel, Ph.D. is a kind, compassionate human being, who is an astute clinician and a renowned expert in hypnosis and trauma therapy. A former president of ASCH.

I recommend.

#trance
#hypnosis
#relationships
#trauma

YouTube Share your videos with friends, family, and the world

New psychology podcast

Your time will be well spent; and this is high quality.

www.youtube.com/watch?v=37yh...

Eric Spiegel, Ph.D. is a kind, compassionate human being, who is an astute clinician and a renowned expert in hypnosis and trauma therapy. I recommend.

#hypnosis
#trauma
#relationships

Victory! The Immigrant Safety Act was just signed into law by Governor Michelle Lujan Grisham.

New Mexico is sending a message: ICE can't be allowed to continue terrorizing our communities.

ActInf GuestStream 126.1 ~ Distributed Information and Computation in Generic Quantum Systems YouTube video by Active Inference Institute

A profound thought from @chrisfields38.bsky.social: Your environment is a cognitive agent.

He shared it here (minute 18) in the context of quantum information theory

www.youtube.com/watch?v=jm70...

I wrote comments on how I understand Chris here:

www.linkedin.com/posts/activi...

Distributed Information and Computation in Generic Quantum Systems This book provides insight into modern constructs of AI and quantum information and includes tools for the characterization of physical interactions.

a useful book by @chrisfields38.bsky.social and James Glazebrook

link.springer.com/book/10.1007...

Presentation about it
www.youtube.com/watch?v=jm70...

Cognition spaces: natural, artificial, and hybrid Cognitive processes are realized across an extraordinary range of natural, artificial, and hybrid systems, yet there is no unified framework for comparing their forms, limits, and unrealized possibili...

I like this paper, and not just for the obvious reason. A great summary of where we stand in thinking about types of mind. Nice also to see this made explicit: "All living systems share a fundamental property: agency."
arxiv.org/abs/2601.12837

Alexey Tolchinsky (@alexeytolchinsky) “No-thing” and “every-thing” are nearly the same under the Free Energy Principle (FEP). “Some-thing” only exists when connections between the components are sparse. All things, objects and agents alik...

Every-thing is no-thing under FEP.

substack.com/@alexeytolch...

Platonic Space discussion 3 YouTube video by Michael Levin's Academic Content

Karl Friston's useful connection of two concepts: Epistemic affordance/foraging, which in some ways is close to play is only possible in deep/hierarchical generative models, not in flat ones. minute 48 here

www.youtube.com/watch?v=E6_X...