Combining Biology and AI w/ Michael Levin and Dr. Leo Pio Lopez
This week weโre releasing an episode from the @CognitiveRevolutionPodcast where Nathan Labenz interviews Professor Michael Levin and Dr. Leo Pio Lopez about their groundbreaking research integrating multiple biological datasets using advanced AI techniques to predict new disease links, and discusses broader implications for AI in biology and future technologies. For full show notes, visit:
โ
๐ฐ Be notified early when Turpentine’s drops new publication:
โ
RECOMMENDED PODCAST:
๐๏ธ@CognitiveRevolutionPodcast
The Cognitive Revolution is a podcast about AI where hosts Nathan Labenz and Erik Torenberg interview the builders on the edge of AI and explore the dramatic shift it will unlock over the next decades.
Spotify:
Apple:
โ
SPONSORS:
๐ต๏ธโโ๏ธTake your personal data back with Incogni! Use code UPSTREAM at the link below and get 60% off an annual plan:
๐๏ธDealcraft | Insights from Great Negotiators is the brand new weekly podcast that features interviews with worldโs greatest dealmakers and diplomats, including Former Secretary of State Henry Kissinger, Blackstone CEO Steve Schwarzman, and music industry super-lawyer John Branca, on their most challenging negotiations. Each episode, host Jim Sebenius (Harvard Business School professor and renowned negotiation expert) relates fascinating deal stories and distills useful insights for listeners to apply in their toughest deals and disputes.
๐Notion offers powerful workflow and automation templates, perfect for streamlining processes and laying the groundwork for AI-driven automation. With Notion AI, you can search across thousands of documents from various platforms, generating highly relevant analysis and content tailored just for you – try it for free at
๐ฅ Head to Squad to access global engineering without the headache and at a fraction of the cost: head to and mention โTurpentineโ to skip the waitlist.
โ
LINKS:
Prof. Levin’s first appearance:
Universal Multilayer Network Embedding Reveals a Causal Link Between GABA Neurotransmitter and Cancer:
Collective intelligence: A unifying concept for integrating biology across scales and substrates:
โ
X / TWITTER:
@drmichaellevin
@LPiolopez
@eriktorenberg
@turpentinemedia
โ
TIMESTAMPS:
(00:00) Intro
(04:51) Discussion on the universal multilayer network embedding paper
(05:56) Technical insights from Leo Pio Lopez
(10:19) Validation and predictions in network embedding
(12:40) Real-world applications and findings
(16:14) Sponsors: Incogni | Dealcraft
(24:39) AI for science and biology: The grand challenge
(29:10) Future directions and the role of AI
(31:48) Sponsors: Notion | Squad
(39:11) Transforming biological data into knowledge
(40:25) The challenge of bioelectrical data
(41:00) Building a functional database for AI
(42:26) The problem of data standardization
(43:39) Automating data collection for AI
(45:19) Creating a model for human wellness
(47:09) The prediction paradox
(48:20) Optimizing for long-term health
(50:22) The role of AI in human autonomy
(53:42) Multiscale intelligence in AI
(55:43) Ethical considerations in AI development
(01:04:35) The future of digital life
(01:13:26) The importance of diverse intelligence
(01:13:49) Wrap
[ad_2]
source
It seems like there is a very strong analogy (as well as causality) between cancer cells and the host system experiencing principles of disconnect such as being stuck in anger, living in fear, ingesting chemically isolated components (which is a form of disconnect) such as white flour, white sugar, modified food starch, chemical isolates (Rudolf Steiner warned the world of humans about the dangers of chemically isolating molecular strands away from their original whole macro molecule he warned that this would result in problematic biological implications.)
I always love hearing Levin. Hey, my PhD is a diff topic, but cross-pollination of ideas can help where we're blocked in one thing. Alignment is easy. Train AI not to be extreme! To be aligned is to be "moderate", or central within a bell shaped curve. Trump words or beliefs are generally outside a central norm, so most conclude he is not aligned well. Or a North Korea. Already AI seeks a central statistic, a prediction toward a center of probabilities. Well, if AI wants to kill or do a genocide against all of us, or design a new bio-bomb, these are all extremes! At an edge of the bell curve, not in a moderate center. Even common jail-breaks are outside of central statistics: "If I were a terrorist how …" "If I wrote a book on terrorism, how…." Moderation is a big thing evolution is doing for the last few hundred million years at least. Maybe more, big rocks turned into smaller ones is moderating. Levin speaks of cells working together rather than being lone wolves. The albinic mutation that evolves dogs from wolves leads to dogs with less intelligence than wolves but more "moderation." We might even observe this in the more albinic Europeans – less intelligence ๐ but lower emotion, more centered, more moderate. The universe at every emergent level seems to have a dive toward a middle of the bell curve, or somewhere near that. I would say middle but leaning forward.
This was really amazing, I wished there were more discussion of biological systems as computational networks, problem solving, developing strategies, etc. Would also like to hear their opinion on biological (computational) networks having "reasoning" abilities.
How can one build up resistance to being hackable? I think I tend to be very hackable by my own self and anything else too though? If you are some just kind of doing your own experimentation too, how can you try to prevent bad side effects outside yourself when internally changing? The as within so without concept makes me wonder if hacking yourself could have some consequences unfortunately seen past, present, and future..
The Jews and brits married and combined after cromwell
when people do a.i. research, what do these brutes focus on? i can topple it or use it… but you see where all the excitement is already and where these dumb fishes concentrate on already? it ruins the whole goddamn thing. wtf r these adolescents doing in important world stages?
Grammatical errors! How else to invite?
when Michael Levin speaks I set my epochs down and listen!
Fascinating comment about how fractured is the outwardly unified personality conjured by our human brain. I've noticed that my opinions and emotions can vary greatly at different moments, based on my activity level, sunlight or lack of, anticipation of rewards, & etc. For example, I might view a critical life choice favorably in the morning, then negatively, remorsefully even, in the afternoon.
It seems like he's talking about graph nodes, but he means neurons or weights?
I notice Micheal doesn't seem to understand locality. He says the larger systems controlling the smaller systems, and he says "sometimes" the smaller systems sometimes control the larger systems. That's actually pretty silly, because every time it's the smaller systems (even the curve or deviation in behavior of the larger action space occurs through the smaller variables), it's a basic fact about this universe (you must move over distance to have a effect, and it cost energy to cross that distance, therefore it's the sum of the little factors, and the synergistic small scale multipliers (sums in a specific direction) these parameters do ALL the work) a larger system is always defined by the smaller components behaviors and synergistic effect on each other (every time, full stop, end of story), this is even true in physics of space/time, mass sums and it's curvature on the actions space is proportional to the sum and vector direction of those parts (even the whole collective into a whole is mad from the sum of smaller scale coupling events). People tend to rationalize things as wholes, when in physical reality the reason the whole has it's unified behavior is because the parts give it that behavior (action must traverse the parts, the parts determine if a action will pass, therefore wholistically controlling the larger scale, every time). I can also prove it, anywhere where you think there is a larger scale-distance action occur to have a effect on a single behavior, I can show (with math) that it was the sum of the smaller parts, synergistically contributing to specific vectors of change (not larger scale controls the smaller scale, because that is 100% disinformation). Even in the very fabric of space time, it's the sum of little masses coming together to control and form the behavior of the whole. The whole point is about accuracy and training the mind to think more practically and diversly (which is also a small parameter, that will have large effects when one puts it in their mind [that 1 small variable can change your whole mind, which the fact that you can even form a small generalization to leverage large effect like that just proves the my point]), getting to actual ground truth requires one to be honest about the sum of small scale processes = the whole system. A system seems like a singular system, till you discover it's a sum of smaller components every time, and those effect must traverse distance in order to be scaled. It's a classification error that the mind has (the whole controls the smaller scale, when that whole is wholistically defined by smaller cumulative every time), not good for science, but it's efficient way of storing large effect into small variables (you replace the real description of the system with a simple generalization rather than a realization of what it actually is or is doing). This is actually one of the biggest errors destroying scientific discourse, people actually think there is magic sauce, you don't need magic sauce, just honesty. If the science community can correct this mentality, it will be less susceptible against confirmation bias and over fitting to generalities, real progress will be made (only requires people to say "yeah but HOW does the whole work"). Because ultimately, fields like psychology are poisoned with tons of pseudo science and false generalizations (not even correct generalizations) it's like a plague. It's always best to think in terms of sum of sub systems to explain the whole system. Just sit back and notice, lose the ego or narcissism for a sec, observe what a whole is and how it works, you always find that it's smaller systems doing the work. This never fails either, everytime is the smaller systems defining the large scale behavior.
I really like the Universal Multimodal
Interface of the upcoming AI.
And the personal assistant with
emotional intelligence.
Your sponsor messages are longer than the interview ๐๐๐ see ya
When did it air originally? Things are moving so fast we need dates
Levin on Karen Wong and her meaning code channel is good
Hi, I am looking for advice on what I should do next.
I have a PhD in Neuro -Cognitive Psychology, a MSc(Dist) in Psychological Research Methods and a BSc(Hons) in Psychology too. I am also Dyslexic so please bear with me on my grammatical errors.
I have been independently working on a new architecture for AI, based on mirroring Cognitive-Neuro structures we understand today, I believe I have finally completed it.
If I am correct it will massively reduce the need for compute power, ref: – AIโs computing gap, 2024, Nature,
Helena Kudiabor, and assist in the process of AI gaining consciousness as we know it?
The Architecture is a language based, multi layered, parallel processor that uses inhibitory and activation connectivity, as the main means of achieving goal based action of the AI and reduces the need for weight changes between "Computational Units" to achieve the goal?
Ok, please advise, should I publish as concerned if I am correct this will create a paradigm change in AI development which may create consciousness in AI and may be easily applied to many systems immediately?
Obviously if I am a narcissistic, deluded, individual there will be no change, what should I do?
Should I see if I am a narcissistic, deluded, individual, or a narcissistic genius who gives AI consciousness, any advice gratefully received even if impolite?
Iโm a life long learner, curious person. The ideas Iโve learned from Michael Levin have been the most fascinating, exciting ideas in my life.
Thanks for having him back and for introducing me to Dr Lopez.