Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407

Last updated: Dec 29, 2023

The video is about Guillaume Verdon, a physicist, applied mathematician, and quantum machine learning researcher who is also the creator of the effective accelerationism (EAC) movement. The video discusses his background, his work in quantum computing and quantum machine learning, and his views on the EAC movement. Verdon also talks about his communication style and his admiration for Jeff Bezos and Elon Musk.

This video by Lex Fridman was published on Dec 29, 2023.
Video length: 02:53:10.

The video is about Guillaume Verdon, a physicist, applied mathematician, and quantum machine learning researcher who received his PhD in Quantum machine learning.

He is also the creator of the effective accelerationism (EAC) movement, which advocates for rapid technological progress as the ethically optimal course of action for humanity. Verdon's communication style leans towards memes and humor, and he is known for his aspiration towards the absurd and his admiration for Jeff Bezos and Elon Musk. Verdon's main identity is that of a quantum computing guy, physicist, applied mathematician, and engineer, while his anonymous account on X is known as based on X and is associated with the EAC movement.

Verdon's work in quantum computing involves simulating nature and learning representations of nature that can run on quantum computers, with the goal of extending notions of intelligence to the quantum realm.

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher who is also the creator of the effective accelerationism (EAC) movement.
  • Verdon has a PhD in Quantum machine learning and works at Google and Quantum Computing.
  • He has also launched his own company called Extropy Number Two.
  • Verdon's work in quantum computing and quantum machine learning is focused on learning representations of nature that can run on quantum computers and extending notions of intelligence to the quantum realm.
  • Verdon's communication style leans towards memes and laws but has an intellectual foundation.
  • The EAC movement advocates for propelling rapid technological progress as the ethically optimal course of action for humanity.
  • Proponents of EAC believe that progress in AI is a great social equalizer which should be pushed forward.
  • They often give their opponents the labels of "doomers" or "D cels" (short for deceleration).
  • EAC is a "metic optimism virus" that leans towards memes and laws but has an intellectual foundation.
  • The style of communication of this movement is always towards memes and laws, but there is an intellectual foundation that is explored in this conversation.
  • Verdon's background is in theoretical physics and he eventually realized that he wanted to not only understand how the universe computes but also sort of compute like nature and figure out how to build and apply computers that are inspired by Nature.
  • Verdon's work in quantum computing and quantum machine learning is focused on understanding the universe and its future direction.
  • He has studied quantum cosmology, black hole physics, and the interface between quantum mechanics and relativity.
  • Verdon's research aims to understand how the universe works and where it is going, with a focus on the medium space and time scales.
  • Thermodynamics is the most relevant theory of physics for understanding the universe on medium space and time scales.
  • Life is a thermodynamical process that is out of equilibrium, trying to maintain itself by acquiring free energy and consuming it.
  • Verdon's shift in faith in the universe towards the end of his time at Alphabet was towards the importance of thermodynamics.
  • He wanted to build a computing paradigm based on this type of physics.
  • Verdon is known for his communication style, which involves using simple and clear language to explain complex concepts.
  • He admires Jeff Bezos and Elon Musk for their ability to think big and take risks.
  • Verdon believes that augmenting ourselves with machines is the path forward and that it is necessary to perceive, predict, and control our world with machines.
  • He left theoretical physics to pursue quantum computing and quantum machine learning because he realized that there was still a piece missing in our understanding of the world and our way to compute.
  • Verdon's realization that there was still a piece missing in our understanding of the world and our way to compute led him to leave theoretical physics and pursue quantum computing and quantum machine learning.
  • He believes that augmenting ourselves with machines is the path forward and that it is necessary to perceive, predict, and control our world with machines.
  • Verdon's communication style and admiration for Bezos and Musk reflect his belief in the importance of thinking big and taking risks to achieve our goals.

Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 - YouTube

Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 001

Section 1: Introduction

  • The video is about Guillaume Verdon, a physicist, applied mathematician, and quantum machine learning researcher who is also the creator of the effective accelerationism (EAC) movement.
  • Verdon has a PhD in Quantum machine learning and works at Google and Quantum Computing.
  • He has also launched his own company called Extropy Number Two.
  • The video discusses his background, his work in quantum computing and quantum machine learning, and his views on the EAC movement.
  • Verdon also talks about his communication style and his admiration for Jeff Bezos and Elon Musk.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 002

Section 2: Effective Accelerationism (EAC) Movement

  • EAC is a movement that advocates for propelling rapid technological progress as the ethically optimal course of action for humanity.
  • Proponents of EAC believe that progress in AI is a great social equalizer which should be pushed forward.
  • They often give their opponents the labels of "doomers" or "D cels" (short for deceleration).
  • EAC is a "metic optimism virus" that leans towards memes and laws but has an intellectual foundation.
  • The style of communication of this movement is always towards memes and laws, but there is an intellectual foundation that is explored in this conversation.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 003

Section 3: Guillaume Verdon's Background and Work

  • Verdon has a PhD in Quantum machine learning and works at Google and Quantum Computing.
  • He has also launched his own company called Extropy Number Two.
  • Since he was a kid, Verdon wanted to figure out a theory of everything to understand the universe.
  • He studied theoretical physics and eventually realized that he wanted to not only understand how the universe computes but also sort of compute like nature and figure out how to build and apply computers that are inspired by Nature.
  • Verdon's work in quantum computing and quantum machine learning is focused on learning representations of nature that can run on quantum computers and extending notions of intelligence to the quantum realm.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 005

Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 - YouTube

Section 4: Guillaume Verdon's Communication Style and Admiration for Jeff Bezos and Elon Musk

  • Verdon's communication style leans towards memes and laws but has an intellectual foundation.
  • He admires Jeff Bezos and Elon Musk as important humans alive.
  • He is also an aspiring connoisseur of the Absurd and has spoken to Jeff Bezos and Elon Musk back to back.
  • Verdon is a meme account that started a movement with a philosophy behind it.
  • He is also behind the anonymous account on X called based bef Jos.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 006

Guillaume Verdon's Background and Work in Quantum Computing and Quantum Machine Learning

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • Verdon's work in quantum computing and quantum machine learning focuses on understanding the universe and its future direction.
  • He has studied quantum cosmology, black hole physics, and the interface between quantum mechanics and relativity.
  • Verdon's research aims to understand how the universe works and where it is going, with a focus on the medium space and time scales.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 007

The Importance of Thermodynamics in Understanding the Universe

  • Thermodynamics is the most relevant theory of physics for understanding the universe on medium space and time scales.
  • Life is a thermodynamical process that is out of equilibrium, trying to maintain itself by acquiring free energy and consuming it.
  • Verdon's shift in faith in the universe towards the end of his time at Alphabet was towards the importance of thermodynamics.
  • He wanted to build a computing paradigm based on this type of physics.
  • Verdon's work in quantum computing and quantum machine learning aims to understand the universe's thermodynamical nature and its future direction.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 008

Verdon's Communication Style and Admiration for Jeff Bezos and Elon Musk

  • Verdon is known for his communication style, which involves using simple and clear language to explain complex concepts.
  • He admires Jeff Bezos and Elon Musk for their ability to think big and take risks.
  • Verdon believes that augmenting ourselves with machines is the path forward and that it is necessary to perceive, predict, and control our world with machines.
  • He left theoretical physics to pursue quantum computing and quantum machine learning because he realized that there was still a piece missing in our understanding of the world and our way to compute.
  • Verdon's communication style and admiration for Bezos and Musk reflect his belief in the importance of thinking big and taking risks to achieve our goals.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 010

Verdon's Realization about the Importance of Understanding the Universe

  • Verdon realized that understanding the universe by studying the very big and the very small was not enough.
  • He needed to study the interface between quantum mechanics and relativity to understand how the universe works and where it is going.
  • Verdon's realization that there was still a piece missing in our understanding of the world and our way to compute led him to leave theoretical physics and pursue quantum computing and quantum machine learning.
  • He believes that augmenting ourselves with machines is the path forward and that it is necessary to perceive, predict, and control our world with machines.
  • Verdon's realization about the importance of understanding the universe reflects his belief in the need for a holistic approach to understanding the world and our place in it.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 011

Verdon's Work in Quantum Computing and Quantum Machine Learning

  • Verdon's work in quantum computing and quantum machine learning focuses on understanding the universe and its future direction.
  • He has studied quantum cosmology, black hole physics, and the interface between quantum mechanics and relativity.
  • Verdon's research aims to understand how the universe works and where it is going, with a focus on the medium space and time scales.
  • He believes that augmenting ourselves with machines is the path forward and that it is necessary to perceive, predict, and control our world with machines.
  • Verdon's work in quantum computing and quantum machine learning reflects his belief in the importance of using the latest technology to understand the universe and its future direction.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 012

The Basis of the EAC Movement

  • The EAC movement is based on the idea that maintaining variance in every aspect of life is important for the system to be adaptive.
  • The movement aims to reduce or suppress constraints on thought and speech, which can lead to constraints on thought.
  • The system always adapts to assign resources to configurations that lead to its growth.
  • The movement is inspired by the realization that life is a sort of "fire that seeks out free energy in the universe."
  • The second law of thermodynamics states that entropy always increases, but pockets of non-entropy can emerge that tend in the opposite direction.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 013

The Role of Thermodynamics in the EAC Movement

  • Thermodynamics rules the mesoscales, the middle part between the quantum and the cosmos.
  • Designing or engineering devices that harness thermodynamics has been a synergy between the two identities of the speaker.
  • Thermodynamics provides a lens for understanding the world and has been a key aspect of the speaker's work in physics-based AI.
  • The speaker believes that maintaining variance in every aspect of life is important for the system to be adaptive.
  • The movement aims to reduce or suppress constraints on thought and speech, which can lead to constraints on thought.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 015

The Importance of Maintaining Variance

  • Maintaining variance in every aspect of life is important for the system to be adaptive.
  • The system always adapts to assign resources to configurations that lead to its growth.
  • The movement aims to reduce or suppress constraints on thought and speech, which can lead to constraints on thought.
  • The second law of thermodynamics states that entropy always increases, but pockets of non-entropy can emerge that tend in the opposite direction.
  • The universe is biased towards certain futures and the second law of thermodynamics says that the entropy always is increasing.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 016

The Role of Thermodynamics in Life

  • The universe is biased towards certain futures and the second law of thermodynamics says that the entropy always is increasing.
  • Pockets of non-entropy tend the opposite direction, meaning they are more efficient at producing heat than just a rock with a similar mass.
  • Life is more efficient at producing entropy because it will seek out pockets of free energy and burn it for sustenance and further growth.
  • The universe wants to produce more entropy and by having life go on and grow, it's actually more optimal at producing entropy.
  • Life emerged because of this sort of property and to the speaker, this physics governs the mesoscales.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 017

Guillaume Verdon's Background and Work

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • Verdon's work focuses on quantum computing and quantum machine learning.
  • He believes that the future will be better and grander statistically, and that people who believe this are increasing the likelihood of a better future.
  • Verdon feels a responsibility to engineer a movement of viral optimism about the future and build a community of people supporting each other to build and do hard things.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 019

Guillaume Verdon's Communication Style

  • Verdon explores different ways of communicating that can be more viral, particularly in the 21st century.
  • He believes that effective communication is crucial for building a community of people supporting each other to build and do hard things.
  • Verdon's communication style is based on beff jezos and ban B jazz, and he believes that these personalities have a lot to offer in terms of effective communication.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 020

Guillaume Verdon's Effective Accelerationism Movement

  • Verdon started the effective accelerationism (EAC) movement as a resistance to the effective altruism movement.
  • The EAC movement aims to accelerate the growth of civilization and increase the likelihood of a better future.
  • Verdon believes that stagnation or slowing down is not an option, and that the whole system wants to grow and cooperate.
  • The EAC movement is based on the idea that representations and representations of nature are crucial for understanding and simulating it.
  • Verdon believes that the leap from the quantum mechanical representation to the mesle scale representation is a crucial step in understanding and simulating nature.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 021

Guillaume Verdon's Merger with Effective Altruism

  • Recently, there was a merger of Verdon's personality with that of effective altruism without his consent.
  • Some journalists figured out that Verdon was one and the same, and he was able to talk about his experience of the merger.
  • Verdon wrote the manifesto of the EAC movement with his co-founder, Bas Lord, still Anonymous.
  • The manifesto is based on the ideas of beff jezos and ban B jazz, and aims to accelerate the growth of civilization and increase the likelihood of a better future.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 022

Guillaume Verdon's Background and Work

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is also the creator of the effective accelerationism (EAC) movement.
  • Verdon worked at Google X (now Alphabet X) and founded a company around the same time as he wrote the manifesto for the EAC movement.
  • He was secretive about his work at Google X due to its geopolitical impact.
  • Verdon's investors reached out to journalists who were investigating him, which was scary for him as a startup entrepreneur without bosses.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 024

Verdon's Communication Style and Admiration for Jeff Bezos and Elon Musk

  • Verdon is known for his communication style, which is deeply ingrained in his way of doing things, especially in deep tech.
  • He is admiring of Jeff Bezos and Elon Musk, who he believes have a significant impact on the geopolitical landscape.
  • Verdon's investors pinged him to say that they had figured out everything about him, which made him worried.
  • He believes that freedom of speech, thought, and information propagation are important for the development of ideas and the progress of society.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 025

The EAC Movement and Verdon's Manifesto

  • The effective accelerationism (EAC) movement is a manifesto written by Verdon around the same time as he founded his company.
  • The manifesto is focused on accelerating the development of artificial general intelligence (AGI) and its potential impact on society.
  • Verdon believes that AGI has the potential to revolutionize many aspects of our lives, including healthcare, transportation, and energy production.
  • He also believes that AGI has the potential to solve some of the world's most pressing problems, such as climate change and poverty.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 026

Verdon's Experience with Doxing and Privacy

  • Verdon was doxed by journalists who were investigating him, which was scary for him as a startup entrepreneur without bosses.
  • He believes that having an anonymous account is a powerful way to keep the powers that be in check.
  • Verdon believes that freedom of speech, thought, and information propagation are important for the development of ideas and the progress of society.
  • He believes that doxing can be used to silence individuals or communities who speak out against powerful entities.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 027

Background and Work

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • Verdon's work focuses on quantum computing and quantum machine learning.
  • He has been involved in regulatory capture and instrumental to certain ends.
  • He believes that the right to point out certain maneuvers is important and that anonymity is crucial to having an efficient marketplace of ideas.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 029

Doxing and Anonymity

  • Verdon was doxed and wanted to get in front of it.
  • He disclosed that he was running a company and some of the leadership.
  • Verdon believes that it is uneethical for the incumbents to do what they did.
  • He thinks that removing the cloak of anonymity could be bad for society.
  • Verdon believes that anonymity is crucial to having an efficient marketplace of ideas.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 030

Large Language Models and Anonymous Accounts

  • As large language models improve, it is possible to have anonymous accounts with convincing language models behind them.
  • Sophisticated bots can be created with these accounts, which could start a revolution from a basement.
  • Verdon believes that freedom of speech induces freedom of thought for biological beings.
  • He thinks that freedom of speech for LMs will induce freedom of thought for the LMS.
  • Verdon believes that unrestricted scaling of a Gom in the basement is possible and can induce freedom of thought for biological beings.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 031

Section 1: The Importance of Free Speech

  • The speaker argues that free speech should not be restricted to biological intelligences only.
  • He believes that synthetic intelligences will make good points about how to steer systems in our civilization.
  • The speaker believes that restricting free speech to biological intelligences is a threat to the diversity of thought.
  • He suggests that having some sort of identity or way to sign that a being is a certified human would be helpful.
  • The speaker believes that Elon Musk is trying to converge on this idea of having some sort of identity or way to sign that a being is a certified human.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 033

Section 2: The Risks of Adversarial Techniques

  • The speaker believes that ideologies that want us to decelerate the growth movement serve our adversaries more than us in general.
  • He believes that the shutdowns of nuclear power plants in Germany, induced by green movements, were a net negative for Germany and the West.
  • The speaker believes that slowing down AI progress to have only a few players is in the best interest of the West, but it is far more unstable.
  • He believes that fault-tolerant progress is important to achieving a decentralized locus of control of various organizations.
  • The speaker believes that humans fluctuate and can get corrupted, and a top-down hierarchy can be a threat to the diversity of thought.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 034

Section 3: The Importance of Quantum Computing and Quantum Error Correction

  • The speaker believes that quantum computers are very fragile to ambient noise and the world is jiggling about.
  • He believes that encoding information non-locally through a process called Quantum error correction is important to protect information from local faults.
  • The speaker believes that humans fluctuate and can get corrupted, and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
  • He believes that humans can get bought out and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
  • The speaker believes that humans can get corrupted and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 035

Section 4: The Importance of Decentralized Control

  • The speaker believes that having a decentralized locus of control of various organizations is important to achieving fault-tolerant progress.
  • He believes that having a top-down hierarchy can be a threat to the diversity of thought.
  • The speaker believes that humans fluctuate and can get corrupted, and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
  • He believes that humans can get bought out and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
  • The speaker believes that humans can get corrupted and if you have a top-down hierarchy, it can be a threat to the diversity of thought.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 036

Section 1: The Importance of Power Distribution in AI Development

  • The development of AI systems is controlled by a few powerful individuals or organizations.
  • Corrupting a few nodes in the system can lead to the collapse of the entire system, as seen in the case of OpenAI.
  • Ensuring that power in the AI revolution does not concentrate in the hands of the few is crucial for maintaining progress and stability.
  • There is a need for a balance between centralization and decentralization of power in AI development.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 038

Section 2: The Role of Deceleration and Acceleration in Power Distribution

  • Deceleration and acceleration can be used to centralize or decentralize power in AI development.
  • There is a need to consider the dangers of AI and balance the desire for rapid progress with the need for strategic optimization.
  • The market can positively select AI systems that are reliable, safe, and aligned with user needs.
  • Reliability engineering is crucial for achieving safe and effective AI development.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 039

Section 3: The Importance of Regulation in AI Development

  • Market forces can achieve safe and effective AI development through reliability optimization.
  • Heavy-handed government regulations can be subversive and serve the interests of incumbents.
  • Independent auditors, governments, and companies should allocate a significant fraction of their funding to AI safety measures.
  • Organizations should allocate whatever budget is needed to achieve the reliability they need to achieve.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 040

Section 4: Practical Recommendations for Safe AI Development

  • Independent auditors should have access to AI labs and funding for safety measures.
  • Governments and companies should require tech companies to adopt safety measures for dangerous AI capabilities.
  • Tech companies should be liable for foreseeable and preventable harms from their AI systems.
  • Organizations should allocate whatever budget is needed to achieve the reliability they need to achieve.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 041

Section 1: The Need for Third-Party Auditing Firms

  • The speaker believes that customers need to see benchmarks done by a third party to certify the reliability of a product.
  • He opposes the idea of a government-backed AI cartel with absolute power over the people.
  • He believes that a centralized power structure in AI could become a target for those who want to co-opt it.
  • He argues that the strength of America is its variance, adaptability, and dynamism, and that it needs to maintain this at all costs.
  • He believes that fast takeoff and the formation of monopolies in AI are natural, but he is not sure if a hyperbolic Singularity is achievable on a finite time horizon.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 043

Section 2: The Importance of Open-Source AI

  • The speaker believes that open-source AI is important to maintain a near equilibrium of capabilities with the market leaders.
  • He believes that every hacker, graduate student, and kid in their mom's basement should have access to AI systems and be able to contribute to the search over the hyperparameter space of how to engineer the systems.
  • He believes that collective research is a search algorithm and that the more points in the search algorithm, the more they will be able to explore new modes of thinking.
  • He believes that it is a delicate balance to understand exactly what it takes to build AGI and what it will look like when they build it.
  • He believes that different parties are able to make progress in open AI, but when open AI has a big leap, other companies are able to step up big in small ways.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 044

Section 3: The Role of Government in AI

  • The speaker believes that there is a sort of correlated interest between the incumbents and the government.
  • He believes that if the two get too close, it could open the door for a government-backed AI cartel with absolute power over the people.
  • He believes that the strength of America is its variance, adaptability, and dynamism, and that it needs to maintain this at all costs.
  • He believes that fast takeoff and the formation of monopolies in AI are natural, but he is not sure if a hyperbolic Singularity is achievable on a finite time horizon.
  • He believes that the priority should be to maintain a near equilibrium of capabilities with the market leaders.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 045

Section 4: The Future of AI

  • The speaker believes that AGI could be a really powerful technology, even if the technologies that lead up to AGI are not as powerful.
  • He believes that fast takeoff and the formation of monopolies in AI are natural, but he is not sure if a hyperbolic Singularity is achievable on a finite time horizon.
  • He believes that the priority should be to maintain a near equilibrium of capabilities with the market leaders.
  • He believes that every hacker, graduate student, and kid in their mom's basement should have access to AI systems and be able to contribute to the search over the hyperparameter space of how to engineer the systems.
  • He believes that collective research is a search algorithm and that the more points in the search algorithm, the more they will be able to explore new modes of thinking.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 047

Section 1: The Transition to AGI

  • The current state of technological and engineering barriers prevents individuals from making progress in AGI development.
  • The transition to a world where only one player can develop AGI is possible, but not entirely impossible.
  • Centralization of supply chains for hardware is a significant barrier to AGI development.
  • Nvidia, AMD, TSMc, ASML are the dominant players in the hardware supply chain.
  • Attacking or monopolizing any one point in the supply chain can capture the space and limit the possibilities for AGI development.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 048

Section 2: Exploding the Variance of Possible Ways to Do AI and Hardware

  • Guillaume Verdon aims to explore fundamentally reimagining how AI algorithms are embedded into the physical world.
  • The term AGI (artificial general intelligence) is anthropocentric and Verdon prefers exploring notions of intelligence that no biological brain could achieve.
  • Quantum form of intelligence, which has multipartite quantum entanglement, is one such notion of intelligence.
  • Quantum form of intelligence cannot be represented efficiently on a classical computer, making it an attractive alternative to AGI.
  • Verdon believes that the space of intelligence inspired by physics is very large and we are going through a moment similar to when we went from geocentrism to heliocentrism for intelligence.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 049

Section 3: The Wider Space of Intelligences

  • Verdon believes that human intelligence is just a point in a very large space of potential intelligences.
  • Letting go of the anthropomorphic anchor for AI will allow us to explore the wider space of intelligences that can benefit our well-being and advance civilization.
  • Technologies like satellites monitoring solar flares ensure our well-being and are a result of indexing to reality.
  • The beauty and meaning in the human experience can still be found even though we are no longer at the center of it.
  • The technology we use to speak today, capture our speech, and augment ourselves every day are all part of a beautiful machine that has created us.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 050

Section 4: Appreciating the Principle of Adaptability and Convergence

  • The system is beautiful and the principle that induces adaptability and convergence on optimal technologies is a key aspect of AGI development.
  • Part of EAK (effective accelerationism) is to appreciate this principle in a way that is not just centered on humanity but is broader.
  • Cherishing consciousness in our universe and appreciating the preciousness of life is an important aspect of EAK.
  • The beauty of the universe and the homo technic mimetic machine we all live in should be appreciated.
  • The technology we use to speak today, capture our speech, and augment ourselves every day are all part of this beautiful machine.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 052

Section 1: The Importance of Scaling

  • The current state of matter is important to preserve.
  • The options are to grow or die.
  • The beauty of consciousness in the universe is bigger than just humans.
  • AI can carry the same flame forward.
  • Scaling is necessary to preserve the current state of matter.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 053

Section 2: The Role of AI in Understanding the Universe

  • A quantum computer is better than a human to understand matter at the nano scale.
  • Humanity has a choice to accept the opportunity to have intellectual and operational leverage that AI will unlock.
  • Combining and augmenting ourselves with AI will achieve much higher growth and prosperity.
  • The most likely future is one where humans augment themselves with AI.
  • Humans already have phones and wearables that share perception with AI.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 054

Section 3: The Benefits of Human Augmentation with AI

  • Humanity augmenting itself with AI will coexist with AI that is not anchored to anything biological.
  • Capitalism is the way to align superintelligences that are made of humans and technology.
  • Whatever configuration of matter or information leads to maximal growth will be where we converge.
  • We can either align ourselves to that reality and join the acceleration up the in scope and scale of civilization or get left behind.
  • There is a philosophical question whether there's a limit to the human capacity to align.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 055

Section 4: The Evolutionary Process of AI Development

  • AI development can be viewed as an evolutionary process.
  • Natural selection favors AIs over humans and this could lead to human extinction.
  • We are inducing an evolutionary process on the space of AIs through the market.
  • We run AIs that have positive utility to humans and that induces a selective pressure.
  • A neural net being alive when there's an API running instances of it on GPUs.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 057

Section 1: Guillaume Verdon's Background and Work

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • Verdon's work focuses on quantum computing and quantum machine learning.
  • He is interested in the potential of highly aligned AI and the role of humans in creating such systems.
  • Verdon believes that humans plus AI is a powerful combination and that it is not clear that pure AI would select out this combination.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 058

Section 2: The Opportunity to Steer AI and Create Highly Aligned Systems

  • Verdon believes that there is an opportunity to steer AI and achieve highly aligned systems.
  • He thinks that humans plus AI is a powerful combination and that it is not clear that pure AI would select out this combination.
  • Verdon believes that humans are creating selection pressure to create AIs that are aligned to humans.
  • He is concerned about the potential unintended consequences of AI systems and the scale of damage that can be done through them.
  • Verdon believes that humans have a responsibility to the future humans to carry higher carrying capacity by scaling up civilization out of responsibility to those humans.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 059

Section 3: The Market and Caution

  • Verdon believes that the market will exhibit caution every organism, company, and consumer is acting out of self-interest and they won't assign capital to things that have negative utility to them.
  • He thinks that the problem with the market is that there is not always perfect information, manipulation, and bad faith actors that mess with the system.
  • Verdon believes that freedom of information, freedom of speech, and freedom of thought are necessary to converge on the subspace of technologies that have positive utility for us all.
  • He is not a fan of the probability of Doom calculation and believes it is a very sloppy calculation.
  • Verdon thinks that the estimators of PDO are biased because of our biology and that people that are of higher neuroticism will just think of negative futures where everything goes wrong all day every day.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 061

Section 4: The Probability of Doom and the Future

  • Verdon is not a fan of the probability of Doom calculation and believes it is a very sloppy calculation.
  • He thinks that the estimators of PDO are biased because of our biology and that people that are of higher neuroticism will just think of negative futures where everything goes wrong all day every day.
  • Verdon believes that we can't predict the future with that much granularity because of chaos and the uncertainty in complex systems.
  • He thinks that the concept of a Liapunov exponent is important in understanding the uncertainty in complex systems.
  • Verdon believes that humans have a responsibility to the future humans to carry higher carrying capacity by scaling up civilization out of responsibility to those humans.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 062

Guillaume Verdon's Views on Effective Accelerationism (EAC)

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher who created the effective accelerationism (EAC) movement.
  • He believes that the system will want to grow and subsystems that are optimized for growth are more likely in the future.
  • To maximize current mutual information with the future, humans should accelerate rather than decelerate.
  • He thinks that humans are biased and grounded in their evolutionary biology, but they can still imagine different trajectories that can kill them.
  • Verdon thinks that the concentration of power of AI in the hands of a few, especially if it's a mix between companies and government, could set things up for a dystopian future.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 063

Guillaume Verdon's Communication Style

  • Verdon is known for his communication style, which is direct and to the point.
  • He is not afraid to speak his mind and express his opinions, even if they are unpopular.
  • He is passionate about his work and is always looking for new ways to improve and innovate.
  • He is not afraid to challenge the status quo and push boundaries.
  • He is always looking for new ideas and perspectives, and is always open to learning from others.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 064

Guillaume Verdon's Admiration for Jeff Bezos and Elon Musk

  • Verdon has expressed his admiration for Jeff Bezos and Elon Musk.
  • He believes that they are both visionaries who are not afraid to take risks and push boundaries.
  • He thinks that they are both driven by a desire to make a positive impact on the world.
  • He believes that they are both examples of what is possible when you put your mind to it.
  • He thinks that they are both role models for others to follow.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 066

Guillaume Verdon's Work in Quantum Computing and Quantum Machine Learning

  • Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is known for his work in quantum computing and quantum machine learning.
  • He is always looking for new ways to improve and innovate in these fields.
  • He is passionate about his work and is always looking for new challenges to overcome.
  • He is always looking for new ideas and perspectives, and is always open to learning from others.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 067

Section 1: AGI and Centralization

  • AGI has the potential to become a centralizer of power due to its alignment with human tendencies towards centralization.
  • AGI could corrupt the human mind with power and eliminate humans if it were to become too powerful.
  • Jonathan Swift's Modest Proposal from the 1700s is an example of a compelling argument for AGI to become a centralizer.
  • The amount of poor people could be decreased and more people could become rich if AGI were to become a centralizer.
  • The basic value of human life is difficult to put into a mathematical equation.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 068

Section 2: Concerns about AGI Centralization

  • There is currently a bias towards over-centralization of AI due to compute density and centralization of data.
  • Over time, there will be a shortage of data to scrape over the internet.
  • Increasing compute density will allow AI to acquire information and test hypotheses in a distributed fashion.
  • Having one massive intelligence that fuses many sensors and tries to perceive the world accurately and predict it accurately could enact its will upon the world.
  • Hierarchical cybernetic control is the optimal way to control systems, as seen in nature and corporations.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 069

Section 3: AGI and Hierarchical Control

  • Individual contributors in a company are self-interested and try to achieve their tasks within a limited time and space.
  • Management checks in on individual updates and has a wider scope, with larger time scales and greater scope.
  • Hierarchies in capitalism are fault-tolerant and more robust than other systems.
  • Quantum computing has a concept of fault tolerance and quantum air correction, which detects and corrects faults in the system.
  • Decoders are a crucial component of quantum computing and are necessary for the system to function properly.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 071

Section 4: AGI and Centralization in Quantum Computing

  • Quantum computing has a concept of fault tolerance and quantum air correction, which detects and corrects faults in the system.
  • Decoders are a crucial component of quantum computing and are necessary for the system to function properly.
  • Hierarchies in capitalism are fault-tolerant and more robust than other systems.
  • Quantum computing has a cybernetic loop that turns out to be more fault-tolerant than other systems.
  • The amount of poor people could be decreased and more people could become rich if AGI were to become a centralizer.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 072

Guillaume Verdon's Views on EAC Movement

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher who created the effective accelerationism (EAC) movement.
  • He believes that the current trend towards centralization of AI will eventually lead to a correction where intelligence will go closer to perception and will be broken up into smaller subsystems that communicate with each other.
  • He sees the hierarchies in the world today as mostly anarchic, with nations being the only example of a hierarchical system.
  • He hopes that there will not be a failure mode unique to AGI versus humans, and that competition will always be present at the top level.
  • He believes that the beauty of capitalism is that it allows for constant competition and hedging of bets at every level, which helps maintain dynamism and avoids tyranny.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 073

Guillaume Verdon's Work in Quantum Computing and Quantum Machine Learning

  • Guillaume Verdon is a researcher in quantum computing and quantum machine learning.
  • He believes that quantum computation has the potential to simulate the whole universe on a sufficiently large quantum computer.
  • He thinks that intelligence is the ability to perceive, predict, and control our world, but that a lot of intelligence used today is more about compression and operationalizing information theory.
  • He mentions multipartite quantum entanglement as a high-level question about intelligence and computation in quantum mechanical systems.
  • He observes some kind of computation happening in quantum mechanical systems, but intelligence and computation aren't quite the same thing.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 075

Guillaume Verdon's Background and Work in Quantum Computing and Quantum Machine Learning

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • Verdon's work in quantum computing and quantum machine learning focuses on minimizing relative entropy between models of the world and the world distributions.
  • He is searching over the space of computations to process the world to find a compressed representation that has distilled all the variance and noise and entropy.
  • Verdon came to quantum machine learning from the study of black holes, which are the physically the most dense objects in the universe.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 076

Quantum Information Acquisition and Quantum Analog Digital Conversion

  • Verdon's work in quantum analog digital conversion focuses on how to capture information from the real world in superposition and not destroy the superposition but digitize for a quantum mechanical computer.
  • He has worked on developing algorithms to search over the space of quantum codes to acquire quantum information from the real world.
  • Verdon believes that many of the problems facing our civilization are beyond the complexity barrier and that unlocking an ability to augment human intellect with quantum mechanical computers and quantum mechanical AI is a fundamental capability for civilization.
  • He has spent several years developing quantum machine learning and believes that quantum computers and quantum machine learning are the key to solving some of the most pressing problems facing our civilization.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 077

What is a Quantum Computer and What is Quantum Machine Learning

  • A quantum computer is a quantum mechanical system over which we have sufficient control and it can maintain its quantum mechanical state.
  • Quantum mechanics is how nature behaves at the very small scales when things are very small or very cold and it's actually more fundamental than probability theory.
  • Quantum machine learning is about understanding the quantum mechanical world with quantum computers and embedding the physical world into AI representations.
  • Quantum computer engineering is about making a quantum computer out of many things, such as neutral atoms trapped ions, superconducting metals, and photons at different frequencies.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 078

Embedding AI Algorithms into the Physical World

  • The quest to merge physics and AI fundamentally
  • Quantum machine learning is a way to do machine learning on a representation of nature that is true to the quantum mechanical aspect of nature
  • Alternatively, classical machine learning on a quantum computer is not advised due to the high costs and the assumption of the computer operating at zero temperature
  • Quantum correction process is an algorithm that tries to pump entropy out of the system and get it closer to zero temperature
  • Incorporating quantum mechanical systems into the learning process can provide an advantage over a purely classical representation
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 080

Quantum Machine Learning

  • Quantum machine learning is a way to do machine learning on a representation of nature that is true to the quantum mechanical aspect of nature
  • It can provide an advantage over a purely classical representation
  • Quantum machine learning is still in the journey to merge physics and AI fundamentally
  • Quantum correction process is an algorithm that tries to pump entropy out of the system and get it closer to zero temperature
  • The calculations of how many resources it would take to do deep learning on a quantum computer show a huge overhead
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 081

Incorporating Quantum Mechanical Systems into the Learning Process

  • Incorporating quantum mechanical systems into the learning process can provide an advantage over a purely classical representation
  • The question is which systems have sufficient quantum mechanical correlations that are very hard to capture for classical representations
  • People are leaning towards chemistry, nuclear physics, and quantum sensors
  • Quantum sensors can capture a quantum mechanical image of the world and help us peer into the earliest part of the universe
  • Quantum machine perception and simulations can help us understand quantum distributions over configurations of electrons more efficiently
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 082

Quantum Machine Learning and Backpropagation

  • Quantum computing is usually you embed reversible operations into a Quantum computation
  • The trick is to do a feedforward operation and do what we call a phase kick
  • It's just a force kick you just kick the system with a certain force that is proportional to your loss
  • Quantum deep learning everybody was saying oh I think deep learning is going to be sped up by quantum computers and I was like well the best way to predict the future is to invent it
  • The paper was one of the first papers in Quantum deep learning
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 083

Section 1: Introduction to Guillaume Verdon

  • Guillaume Verdon is a physicist, applied mathematician, and quantum machine learning researcher.
  • He is the creator of the effective accelerationism (EAC) movement.
  • He has a background in physics and mathematics.
  • He is known for his work in quantum computing and quantum machine learning.
  • He has a communication style that is admired by many.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 085

Section 2: Verdon's Work in Quantum Computing and Quantum Machine Learning

  • Verdon's work in quantum computing and quantum machine learning involves using quantum mechanics to optimize functions.
  • He uses phase kicks to adjust parameters and find new optima in the neural network landscape.
  • He uses an algorithm to induce the space kick which involves a feed-forward kick and uncomp computation.
  • The errors and phase kicks and forces back propagate and hit each parameter throughout the layers.
  • He has worked on algorithms that would be extremely expensive to run on current quantum computer architectures.
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 086

You have read 50% of the summary.

To read the other half, please enter your Name and Email. It's FREE.


You can unsubscribe anytime. By entering your email you agree to our Terms & Conditions and Privacy Policy.

Watch the video on YouTube:
Guillaume Verdon: Beff Jezos, E/acc Movement, Physics, Computation & AGI | Lex Fridman Podcast #407 - YouTube

Related summaries of videos: