Uncategorized – Digital Minds

The Hidden Spring

The Hidden Spring, by Mark Solms, offers an interpretation of consciousness by locating its source not in the cortex, where cognitive neuroscience usually places it, but in the affective and homeostatic mechanisms of the brainstem. This is a view shared by many other scientists, including António Damásio, Peter Hacker, and Jaak Panksepp.

For Solms, what makes a mental state conscious is that it feels like something, and feeling is inseparable from the organism’s basic biological need to regulate itself and remain alive. He argues that consciousness is essentially the subjective experience of deviations from homeostasis, an internal barometer of how well or poorly the organism is doing.

The book is structured around clinical cases that challenge the standard view that cortical processing generates conscious experience. Solms discusses children with severe cortical loss, patients with brainstem lesions, and various disorders of awareness to show that affective arousal, rather than sophisticated cognition, seems to be the indispensable component of consciousness. Perception, memory, and reasoning can all occur unconsciously, he argues, but feelings cannot. This leads him to claim that the origins of consciousness lie in ancient neural architectures whose primary function is emotional and motivational.

An interesting part of the book is Solms’s integration of this affect-centred view with Karl Friston’s Free Energy Principle. He proposes that organisms survive by reducing prediction error and maintaining physiological equilibrium, and that affective feelings are the subjective representation of this process. Consciousness, in this framework, emerges because feelings guide behavior toward states that are compatible with long-term viability. The mind becomes an internal monitoring system whose qualitative tones, such as pleasure, anxiety, frustration, and relief, signal how the organism is faring.

While the book is persuasive in its use of clinical evidence and offers a refreshing alternative to cortico-centric theories, some of its claims remain controversial. Critics argue that Solms sometimes overstates what certain neurological cases actually demonstrate. The behavioural signs he interprets as evidence of consciousness may have other explanations, and the complete displacement of cortical contributions is not universally accepted. Moreover, although Solms simplifies the Free Energy Principle, the connection between its mathematical formalism and subjective affect remains only partially clarified.

Overall, The Hidden Spring stands as a provocative contribution to the science and philosophy of consciousness. It shifts the emphasis from thinking to feeling, from cortical representation to biological regulation, and from cognitive abstraction to embodied affect. Even if Solms’s conclusions are not universally persuasive, his reframing forces a reevaluation of deep assumptions about where consciousness arises and what it is for. For anyone interested in the intersection of neurobiology, subjective experience, and the foundations of mind, his book offers a compelling and challenging perspective.

Boomers vs. Doomers: Diverging Narratives at the AI Crossroads – Reflections from the Paris AI Summit

At the recent AI Summit in Paris, the tension between competing narratives of AI’s future—what some call the “boomer” optimism and the “doomer” alarmism—was on full display. This post will try to unpack the ideological divide as it emerged across key sessions, focusing on the satellite meetings “AI Science and Society” and “The Inaugural Conference of the International Association for Safe and Ethical AI”. While the former highlighted AI’s potential to advance human knowledge, social good, and collective governance, the latter stressed existential risks, regulatory urgency, and global coordination frameworks. By comparing these perspectives, this post explores how different epistemic communities frame the stakes of AI development, and what this means for research, policy, and public engagement moving forward.

The AI Science and Society meeting was mostly focused on the positive aspects of AI technologies and how these technologies could be used for the benefit of humanity.

The conference chair was Michael Jordan (h=214, c=325k), the Pehong Chen Distinguished Professor in the Department of Electrical Engineering and Computer Science and the Department of Statistics of UC Berkeley.

The program included many invited talks from distinguished researchers:

In his inauguration address, Michael Jordan disparaged the existing hype about AGI, human-level intelligence in AI systems and related issues, arguing for the development of down-to-earth business models that use AI technology in new ways.

Among the many participants, Yan LeCun (h=156,c=408k), one of the most well-known boomers, argued that we do in fact need systems with human-level intelligence:

In parallel (partially), another meeting took place, The Inaugural Conference of the International Association for Safe and Ethical AI, chaired by Stuart Russell, Distinguished Computer Science Professor also from UC Berkeley.

The meeting counted with contributions from many well-known researchers, including Turing Award and Nobel Prize winners:

In this meeting, the tone was completely different: AGI was not only a genuine possibility in the short-term future but a real risk to the future of humanity.

Yoshua Bengio (h=248,c=936k) argued, forcefully, for a future without AI Agents, since having them would be too risky:

Geoffrey Hinton (h=189, c=920k) gave a rather balanced talk on what “Understading” means, making the argument that LLMs represent knowledge in much the same way we do:

In the closing talk by the conference chair, Stuart Russel (h=97, c=130k), the tone was definitely dark, with an almost unique focus on safety and on the risks of AI for the future of humanity:

Later in the talk Russell argues that LLMs were never intended as general models of AI:

and supported very strict rules on which systems should be allowed to run on advanced hardware:

In summary, two very different views on the future of AI were present in Paris, both from very well-informed and influential researchers who have, essentially, the same base information. One is left with the feeling that the crystal ball is, indeed, very cloudy.

NotebookLM: A useful research assistant or more than that?

The launching of NotebookLM by Google is bound to be a serious milestone on the development of AI tools. Although many tools that can process personal information have been made available in the last months, NotebookLM stands alone in its ability to combine, summarise, organize and report the information contained in the sources that are made uploaded.

NotebookLM, made available by Google on an experimental base, offers a simple interface to a very complex and very sophisticated system. The user can select and upload up to 50 documents, in various formats, and then interactively explore them. This exploration can take many forms: the user can request a summary, enter a dialogue with the system by asking questions and receiving answers, ask for a timeline of a table of contents, or obtain a briefing on specific subtopics covered by the sources. All these functionalities are well and good, and rather impressive at that.

But the killer app is, really, the ability to generate a podcast about the contents of the sources. This podcast, whose duration and contents can vary, will consist of a vivid dialogue that explores, in podcast style, the material that was uploaded. For an uninformed listener, such a podcast will be hard to distinguish from a real one, since the conversation is fluid and includes many characteristics of typical human interactions, such as pauses and interjections. True, it may still be a long way from a professionally generated podcast, but the result is still impressive. I tried it with a number of sources, including my 2019 book on Artificial Intelligence (in Portuguese). Even though the book is written in Portuguese, the resulting podcast is in English.

You can listen to the result here and, if I am not wrong, be impressed by what NotebookLM can do.

The Astonishing Hypothesis: The Scientific Search for the Soul

First published almost 30 years ago, Francis Crick’s masterful argument that we can now endeavor to understand consciousness remains eminently readable today.

The “Astonishing Hypothesis”, as Crick puts it, is that “You, your joys and your sorrows, your memories and your ambitions, your sense of personal identity and free will, are in fact no more than the behavior of a vast assembly of nerve cells and their associated molecules“. In other words, your, consciousness, your soul, if you will, are simply the result of the workings of the neurons in your brain. We should therefore be able to study these phenomena scientifically and to arrive at an understanding of what it means to be conscious of something.

The main point of the book is that science has advanced enough to not be afraid of tackling that most complex question: what is consciousness? Crick decides to address what he thinks is the most accessible sub-problem within the largest challenge of understanding consciousness, the problem of visual awareness: how do we become aware of something we see?

As the deep learning revolution moves forward, at a time when convolutional neural networks can perform visual tasks almost as well as humans and large language models challenge us in terms of linguist abilities, the question Crick poses is more important than ever: what exactly does it mean to understand something, what does it mean to be aware of something?

Crick leads us through a convoluted path, first introducing the concepts of consciousness, awareness and attention, then presenting what is known of the physiology of the brain and of the visual system and, finally, proposing tentative explanations of what sort of phenomena could lead to visual awareness. In the process, Crick challenges philosophers, psychologists, religious people and fellow scientists who believe that the phenomenon of consciousness will remain forever beyond our reach.

Being You: Are we ready to understand consciousness?

Artificial consciousness seems to be a hot topic these days. This week’s Economist includes two invited articles about the matter (one arguing that we are progressing fast towards artificial consciousness by Google engineer Blaise Agüera y Arcas and another arguing that we are very far from such a thing by renowned scholar and author Douglas Hofstadter). Several media report of an ongoing discussion within Google of whether LaMDA, Google’s latest Large Language Model (LLM) has some sort of consciousness.

Amid all this hoopla, Anil Seth book brings a welcome breath of fresh air on a very old and difficult problem. Understanding consciousness, in general, and artificial consciousness, in particular, remains an open and obscure problem. As I recently wrote on a short PNAS commentary,

Although it has been the subject of human thought for many centuries, consciousness remains a mysterious and controversial topic. Every one of us is overly familiar with the phenomenon, but there is little agreement on what it is, what it entails, and how it is created. Certainly, no other phenomenon is simultaneously so familiar and so hard to explain.

Still, we are not doomed to remain in the dark forever. Seth’s book does an excellent job at clarifying what consciousness is all about and makes some serious contributions to our understanding of the phenomenon. Seth describes very clearly what are the different aspects of consciousness that need an explanation, and argues convincingly that consciousness is the result of our brain ability to model the world and predict the future including our own role in the unfolding sequence of events. Seth’s argument fits well with several other existing theories, including Baar’s Global Workspace Theory, probably the most popular and widely accepted proposal.

However, Seth makes a clear argument for the relevance of several factors that are not that explicit in other models and also makes important connections with other theories, including Damásio’s focus on the importance of emotions. In the process, Seth takes a stab at the idea that we will never be able to understand consciousness and, in particular, that the hard problem of consciousness will remain forever outside of our reach.

Overall, one is left with the idea that sometime, in the not too distant future, we will have a clear theory of what consciousness is, and how it is produced. Will that lead to artificial consciousness, to the creation of machines that are, at least in some ways, conscious? Here, Seth and I diverge, because Seth seems to shy away from the most natural conclusion that his book lead us into: consciousness is a phenomenon that results from a very specific way to process information about the world and systems that work in that way will, undoubtedly, be conscious, in one way or the other.

Extraterrestrial: The First Sign of Alien Life?

Avi Loeb is not exactly someone who one may call an outsider to the scientific community. As a reputed scholar and the longest serving chair of Harvard’s Department of Astronomy, he is a well-known and reputed physicist, with many years of experience in astrophysics and cosmology. It is therefore somewhat surprising that in this book he strongly supports an hypothesis that is anything but widely accepted in the scientific community: ʻOumuamua, the first interstellar object ever detected in our solar system may be an artifact created by an alien civilization.

We are not talking here about alien conspiracies, UFOs or little green men from Mars. Loeb’s idea, admirably explained, is that there are enough strange things about ʻOumuamua to raise the real possibility that it is not simply a strange rock and that it may be an artificial construct, maybe a lightsail or a beacon.

There are, indeed, several strange things about this object, discovered by a telescope in Hawaii, in October 2017. It was the first object ever discovered near the Sun that did not orbit our star; its luminosity changed radically, by a factor of about 10; it is very bright for its size; and, perhaps more strangely, it exhibited non‑gravitational acceleration as its orbit did not exactly match the orbit of a normal rock with no external forces applied other than the gravity of the Sun.

None of these abnormalities, per se, would be enough to raise eyebrows. But, all combined, they do indeed make for a strange object. And Loeb’s point is, exactly, that the possibility that ‘Oumuamua is an artifact of alien origin should be taken seriously by the scientific community. And yet, he argues, anything that has to do with extraterrestrial life is not considered serious science, leading to a negative bias and to a lack of investment in what should be one of the most important scientific questions: are we alone in the Universe? As such, SETI, the Search for Extra-Terrestrial Life, does not get the recognition and the funding it deserves. Paradoxically, other fields whose theories may never be confirmed by experiment nor have any real impact on us, such as multiverse based explanations of quantum mechanics or string theory, are considered serious fields, attract much more funding, and are more favorably viewed by young researchers.

The book makes for very interesting reading, both for the author’s positions about ‘Oumuamua itself and for his opinions about today’s scientific establishment.

Chinese translation of The Digital Mind

The Chinese translation of my book, The Digital Mind, is now available. For those who want to dust off their (simplified) Chinese, it can be found in the usual physical and online bookstores, including Amazon and Books.com. Regrettably, I cannot directly assess the quality of the translation, you will have to decide for yourself. Or maybe you’d rather go for the more mundane English version, published by MIT Press, or the Portuguese one, published by IST Press.

Decoding the code of life

We have known, since 1953, that the DNA molecule encodes the genetic information that transmits characteristics from ancestors to descendants, in all types of lifeforms on Earth. Genes, in the DNA sequences, specify the primary structure of proteins, the sequence of amino acids that are the components of the proteins, the cellular machines that do the jobs required to keep a cell alive. The secondary structure of proteins specifies some of the ways a protein folds locally, in structures like alpha helices and beta sheets. Methods that can determine reliably the secondary structure of proteins have existed for some time. However, determining the way a protein folds globally in space (its tertiary structure, the shape it assumes) has remained, mostly, an open problem, outside the reach of most algorithms, in the general case.

The Critical Assessment of protein Structure Prediction (CASP) competition, started in 1994, took place every two years since then and made it possible for hundreds of competing teams to test their algorithms and approaches in this difficult problem. Thousands of approaches have been tried, to some success, but the precision of the predictions was still rather low, especially for proteins that were not similar to other known proteins.

A number of different challenges have taken place over the years in CASP, ranging from ab-initio prediction to the prediction of structure using homology information and the field has seen steady improvements, over time. However, the entrance of DeepMind into the competition upped the stakes and revolutionized the field. As DeepMind itself reports in a blog post, the program AlphaFold 2, a successor of AlphaFold, entered the 2020 edition of CASP and managed to obtain a score of 92.4%, measured in the Global Distance Test (GDT) scale, which ranges from 0 to 100. This value should be compared with the value 58.9% obtained by AlphaFold (the previous version of this year’s winner) in 2018, and the 40% score obtained by the winner of the 2016 competition.

Even though details of the algorithm have still not been published, the information provided in the DeepMind post provides enough information to realize that this result is a very significant one. Although the whole approach is complex and the system integrates information from a number of sources, it relies on an attention-based neural network, which is trained end-to-end to learn which amino acids are close to each other, and at which distance.

Given the importance of the problem on areas like biology, medical science and pharmaceutics, it is to be expected that this computational approach to the problem of protein structure determination will have a significant impact in the future. Once more, rather general machine learning techniques, which have been developed over the last decades, have shown great potential in real world problems.

Do humankind’s best days lie ahead?

This book, which transcribes one of the several Munk debates organized by an initiative financed by Peter and Melanie Munk, addresses the question of whether the future of humanity will be better or worst than the present.

The debate, also available in video, takes place between four formidable names, the wizards Steven Pinker and Matt Ridley (apologists of the theory that technology will continue to bring progress) and the prophets Alain de Botton and Malcolm Gladwell (doubters of the idea that further technological developments will keep improving the world).

The dialogue that takes place, between the Pollyanas and the Cassandras (to use an expression coined in the debate itself) is vivid, interesting and, at times, highly emotional. Not one of the debaters has doubts that progress has improved immensely the human condition in the last few centuries, but the consensus ends with that. Will we be able to use science and technology to surmount the environmental, social, and political challenges faced by humanity or did we already reach “peak development” and the future will be worst than the past? Read or watch the debate, and decide for yourself.

My take is that the Pollyanas, Steven Pinker and Matt Ridley, with their optimistic take on the future, win the debate by a large margin, against the Cassandras, Their arguments that the world will continue to improve, based both on the historical tendencies but also on the hope that technology will solve the significant challenges we face do not meet a coherent resistance from Alain de Botton and Malcolm Gladwell. At least they did not manage to convince me that famines, cybersecurity threats, climate change, and inequality will be enough to reverse the course of human progress.

Mastering Starcraft

The researchers at DeepMind keep advancing the state of the art on the utilization of deep learning to master ever more complex games. After recently reporting a system that learns how to play a number of different and very complex board games, including Go and Chess, the company announced a system that is able to beat the best players in the world at a complex strategy game, Startcraft.

AlphaStar, the system designed to learn to play Starcraft, one of the most challenging Real-Time Strategy (RTS) games, by playing against other versions of itself, represents a significant advance in the application of machine learning. In Starcraft, a significant amount of information is hidden from the players, and each player has to balance short term and long term objectives, just like in the real world. Players have to master fast-paced battle techniques and, at the same time, develop their own armies and economies.

This result is important because it shows that deep reinforcement learning, which has already shown remarkable results in all sorts of board games, can scale up to complex environments with multiple time scales and hidden information. It opens the way to the application of machine learning to real-world problems, until now deemed to difficult to be tackled by machine learning.