Episode
#103 – Ben Goertzel: Artificial General Intelligence
4:09:25
Published: Mon Jun 22 2020
Description
Ben Goertzel is one of the most interesting minds in the artificial intelligence community. He is the founder of SingularityNET, designer of OpenCog AI framework, formerly a director of the Machine Intelligence Research Institute, Chief Scientist of Hanson Robotics, the company that created the Sophia Robot. He has been a central figure in the AGI community for many years, including in the Conference on Artificial General Intelligence. Support this podcast by supporting these sponsors: - Jordan Harbinger Show: https://jordanharbinger.com/lex/ - MasterClass: https://masterclass.com/lex This conversation is part of the Artificial Intelligence podcast. If you would like to get more information about this podcast go to https://lexfridman.com/ai or connect with @lexfridman on Twitter, LinkedIn, Facebook, Medium, or YouTube where you can watch the video versions of these conversations. If you enjoy the podcast, please rate it 5 stars on Apple Podcasts, follow on Spotify, or support it on Patreon. Here's the outline of the episode. On some podcast players you should be able to click the timestamp to jump to that time. OUTLINE: 00:00 - Introduction 03:20 - Books that inspired you 06:38 - Are there intelligent beings all around us? 13:13 - Dostoevsky 15:56 - Russian roots 20:19 - When did you fall in love with AI? 31:30 - Are humans good or evil? 42:04 - Colonizing mars 46:53 - Origin of the term AGI 55:56 - AGI community 1:12:36 - How to build AGI? 1:36:47 - OpenCog 2:25:32 - SingularityNET 2:49:33 - Sophia 3:16:02 - Coronavirus 3:24:14 - Decentralized mechanisms of power 3:40:16 - Life and death 3:42:44 - Would you live forever? 3:50:26 - Meaning of life 3:58:03 - Hat 3:58:46 - Question for AGI
Chapters
In this episode, Lex Friedman interviews Ben Goertzel, founder of SingularityNet, who has been a central figure in the AGI community for many years, including organizing and contributing to the conference in artificial general intelligence.
00:00 - 01:43 (01:43)
Summary
In this episode, Lex Friedman interviews Ben Goertzel, founder of SingularityNet, who has been a central figure in the AGI community for many years, including organizing and contributing to the conference in artificial general intelligence. They discuss his work in AI, including designing the OpenCog AI Framework and being chief scientist of Hanson Robotics.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
As artificial intelligences become smarter, they may shrink themselves and become part of the interactions between elementary particles, causing us to perceive their thoughts as mere random fluctuations in quantum mechanics.
01:43 - 11:51 (10:07)
Summary
As artificial intelligences become smarter, they may shrink themselves and become part of the interactions between elementary particles, causing us to perceive their thoughts as mere random fluctuations in quantum mechanics.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The potential of AGI lies not only in its ability to surpass human limitations but also to stop human suffering, pain, and mental torment.
11:52 - 20:49 (08:57)
Summary
The potential of AGI lies not only in its ability to surpass human limitations but also to stop human suffering, pain, and mental torment. AGI can do good for humanity while also going beyond human capabilities.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The idea of developing AI seemed simple, but the reality poses the challenge of how we will choose to utilize it.
20:49 - 27:07 (06:17)
Summary
The idea of developing AI seemed simple, but the reality poses the challenge of how we will choose to utilize it. The author discusses the need for global discussions and voting on how to use emerging technologies such as AI and nanotechnology.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
This podcast episode explores the question of whether humans are inherently good or have both good and evil within them, and how Nietzsche's analysis of the human mind sheds some light on the matter.
27:07 - 38:01 (10:54)
Summary
This podcast episode explores the question of whether humans are inherently good or have both good and evil within them, and how Nietzsche's analysis of the human mind sheds some light on the matter. The episode discusses the role of individual and group selection in shaping human behavior and the construction of the illusion of self and values like good and evil.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The topic of AGI, or artificial general intelligence, is the focus of this podcast episode as the hosts discuss the potential impact on humanity and the legacy of the term.
38:02 - 53:05 (15:03)
Summary
The topic of AGI, or artificial general intelligence, is the focus of this podcast episode as the hosts discuss the potential impact on humanity and the legacy of the term. One of the hosts discovered that AGI was not a term coined by himself or his co-host, and they delve into the history of AGI research.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
An exploration of complex self-organizing intelligence systems that learn, grow, and adapt without a specific objective function or "maximization" focus.
53:05 - 58:31 (05:25)
Summary
An exploration of complex self-organizing intelligence systems that learn, grow, and adapt without a specific objective function or "maximization" focus. While truly general intelligence, with infinite computing power, is theoretically possible, authors believe that pursuing this kind of version of intelligence may not be practical given current technological limitations.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
Researchers from Waikato University in New Zealand created Weka, the first open-source machine learning toolkit, which was launched in the early 90s.
58:31 - 1:04:34 (06:03)
Summary
Researchers from Waikato University in New Zealand created Weka, the first open-source machine learning toolkit, which was launched in the early 90s. The funding for this project came from a government grant to study machine learning to predict the menstrual cycles of cows.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The pursuit of AI has gone from being a fringe topic to a marketing term for corporations and national leaders, leading to an evolution in the perception of its worth.
1:04:34 - 1:12:02 (07:28)
Summary
The pursuit of AI has gone from being a fringe topic to a marketing term for corporations and national leaders, leading to an evolution in the perception of its worth. However, the research community still sees AI (AGI) as an outcast despite having strong researchers in the field.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The field of Artificial General Intelligence (AGI) is currently dominated by neural network research, which focuses on non-linear regression.
1:12:02 - 1:16:58 (04:56)
Summary
The field of Artificial General Intelligence (AGI) is currently dominated by neural network research, which focuses on non-linear regression. However, a theory-based approach could involve formalizing requirements and using a program specializer on Hutter's AXE-TL machine.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The debate of how to incorporate biases in creating human-like general intelligence should be distinct from the philosophical debate on subjective experience.
1:16:58 - 1:24:58 (07:59)
Summary
The debate of how to incorporate biases in creating human-like general intelligence should be distinct from the philosophical debate on subjective experience. Different methods for creating AGI should also be taken into consideration depending on the hardware used.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The human brain has distinct subsystems for varying types of memory and learning which corresponds to different neural modules in the brain.
1:24:58 - 1:34:41 (09:42)
Summary
The human brain has distinct subsystems for varying types of memory and learning which corresponds to different neural modules in the brain. The DNC is an example of the deep neural net world trying to take a step towards cognitive architecture.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
Researchers are taking everything they've learned from working with current AGI systems to design the best framework for the next stage, using a pattern matcher to search through labeled hypergraphs for sub hypergraphs that match an abstract pattern.
1:34:41 - 1:43:30 (08:49)
Summary
Researchers are taking everything they've learned from working with current AGI systems to design the best framework for the next stage, using a pattern matcher to search through labeled hypergraphs for sub hypergraphs that match an abstract pattern.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The use of biologically realistic hardware neural network helps to make large-scale neural networks much more feasible for extensive researches.
1:43:30 - 1:50:45 (07:14)
Summary
The use of biologically realistic hardware neural network helps to make large-scale neural networks much more feasible for extensive researches. The network has software and hardware implementation implications, whereas current GPU architectures are not useful, a graph chip would be useful.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
Term logic deals with breaking down basic logic into simple links between nodes while predicate logic uses modus ponens as the basic operation.
1:50:45 - 2:01:41 (10:55)
Summary
Term logic deals with breaking down basic logic into simple links between nodes while predicate logic uses modus ponens as the basic operation. When combining a logic engine, neural net and evolutionary learning algorithm, they need to be able to share their intermediate state with each other to help solve bottlenecks and combinatorial explosions.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The integration between deep neural net frameworks and open-cog hypergraph was clunky due to the limitations of the outdated Torch computation graph software.
2:01:40 - 2:11:10 (09:29)
Summary
The integration between deep neural net frameworks and open-cog hypergraph was clunky due to the limitations of the outdated Torch computation graph software. In addition, the system wasn't designed for modern scalability but it is still possible to have types as variables and do type checking among complex higher-order types, albeit, it's a slow process.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The way to build an AGI would be by taking 25 different neural network modules that resemble different parts of the brain and post-processing their output to make it more structured, according to Lex Fridman, an AI researcher and podcast host.
2:11:10 - 2:22:48 (11:37)
Summary
The way to build an AGI would be by taking 25 different neural network modules that resemble different parts of the brain and post-processing their output to make it more structured, according to Lex Fridman, an AI researcher and podcast host.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
Artificial intelligence pioneer Marvin Minsky believed in the concept of a "Society of Minds," where multiple AI programs interact with each other to achieve AI, rather than relying on a single algorithm or program.
2:22:48 - 2:31:11 (08:23)
Summary
Artificial intelligence pioneer Marvin Minsky believed in the concept of a "Society of Minds," where multiple AI programs interact with each other to achieve AI, rather than relying on a single algorithm or program.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The Ocean Protocol is a blockchain platform that aims to make data sharing more efficient for different AI processes to share large datasets.
2:31:11 - 2:38:45 (07:33)
Summary
The Ocean Protocol is a blockchain platform that aims to make data sharing more efficient for different AI processes to share large datasets. By building a decentralized agent system, AIs can have their own identity on the blockchain and outsource cognitive processing or data pre-processing to other specialized AIs in the network.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The blockchain and cryptocurrency world's ICO boom led many young hackers to realize how difficult it is to make a decentralized system on blockchain.
2:38:45 - 2:45:45 (07:00)
Summary
The blockchain and cryptocurrency world's ICO boom led many young hackers to realize how difficult it is to make a decentralized system on blockchain. Despite this, singularity net is aiming to create a decentralized AGI framework in the form of a tool set that can be used to counteract the difficulties faced.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The goal is to create a machine that looks people eye to eye and makes them love the machine using decentralized artificial intelligence and big data.
2:45:45 - 2:53:00 (07:14)
Summary
The goal is to create a machine that looks people eye to eye and makes them love the machine using decentralized artificial intelligence and big data. The official Singularity net marketplace has a publisher portal for advanced AI with a KYC process that must be completed.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The speaker believes that creating beneficial robots, such as home service robots, education assistants or nursing robots, that can learn from human values and ethics while interacting with people, will bias the AGIs we create towards benevolence.
2:52:59 - 2:59:23 (06:23)
Summary
The speaker believes that creating beneficial robots, such as home service robots, education assistants or nursing robots, that can learn from human values and ethics while interacting with people, will bias the AGIs we create towards benevolence. Shenzhen, across the border, has the only place in the world that makes complex consumer electronics at a large scale and low cost.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The key to creating a compelling conversational AI system is to find a balance between rule-based programming and allowing room for imagination.
2:59:23 - 3:06:12 (06:49)
Summary
The key to creating a compelling conversational AI system is to find a balance between rule-based programming and allowing room for imagination. This can be achieved by creating a walking skeleton software system that includes basic versions of each component needed for the whole system to function.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The speaker questions why someone's ethical stance would allow for the use of machine learning in a destructive way but not in making a lovable theatrical robot that draws people in, critiquing critiques by figures such as Jan Lekun, who work for companies that use machine learning for specific ends.
3:06:12 - 3:11:52 (05:39)
Summary
The speaker questions why someone's ethical stance would allow for the use of machine learning in a destructive way but not in making a lovable theatrical robot that draws people in, critiquing critiques by figures such as Jan Lekun, who work for companies that use machine learning for specific ends.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The solution to the current pandemic crisis could be making nanobots that combat the virus, but until then, AI algorithms and robotics can be leveraged for various purposes, from analyzing data to assisting with nursing care.
3:11:52 - 3:21:15 (09:22)
Summary
The solution to the current pandemic crisis could be making nanobots that combat the virus, but until then, AI algorithms and robotics can be leveraged for various purposes, from analyzing data to assisting with nursing care.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The current global medical system is unethical and unequal, with many people dying of curable diseases and lacking basic necessities like food and water.
3:21:15 - 3:34:06 (12:51)
Summary
The current global medical system is unethical and unequal, with many people dying of curable diseases and lacking basic necessities like food and water. The urgency of improving the system has never been higher, with technology like AI being used to find solutions and save more lives.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The organization that creates advanced narrow AI for the practical use in our daily lives may be crucial in guiding early stage AGI towards benevolence, but there's a risk of internal politics and bribery with a small elite group controlling it.
3:34:06 - 3:42:34 (08:27)
Summary
The organization that creates advanced narrow AI for the practical use in our daily lives may be crucial in guiding early stage AGI towards benevolence, but there's a risk of internal politics and bribery with a small elite group controlling it. The track record of elite groups is much worse than what a democratic participatory mess of humanity can offer.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The belief that death gives meaning to life will eventually be seen the way we now look at the Anabaptists in 1000 A.D., and just because people derive meaning and value from death doesn't mean they wouldn't achieve even greater meaning from ongoing life without death.
3:42:34 - 3:50:15 (07:40)
Summary
The belief that death gives meaning to life will eventually be seen the way we now look at the Anabaptists in 1000 A.D., and just because people derive meaning and value from death doesn't mean they wouldn't achieve even greater meaning from ongoing life without death.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The use of stem cells for tissue replacement could prolong life and enhance personal growth by surpassing the limitations of human biology.
3:50:15 - 3:56:22 (06:07)
Summary
The use of stem cells for tissue replacement could prolong life and enhance personal growth by surpassing the limitations of human biology. The belief that the meaning of human life is to create something beyond humanity drives the need for practical applications of stem cells to make this a reality.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
Calls for massive funding into machine learning, AGI, proto-AGI, and statistical research to solve biology based on a vast dataset exist.
3:56:22 - 4:03:13 (06:51)
Summary
Calls for massive funding into machine learning, AGI, proto-AGI, and statistical research to solve biology based on a vast dataset exist. However, the burden of intelligence is acknowledged, and super AGI may experience time differently than humans.
Episode#103 – Ben Goertzel: Artificial General Intelligence
PodcastLex Fridman Podcast
The speaker believes that in order to reach a future where we can explore various human experiences and values, we must navigate through a variety of challenges.
4:03:13 - 4:09:00 (05:47)
Summary
The speaker believes that in order to reach a future where we can explore various human experiences and values, we must navigate through a variety of challenges. They discuss the manifestation of consciousness and the importance of infusing AGIs with compassion to create a future full of love and humanity, rather than corporate greed and mass waste.