Lex Fridman - Neil Gershenfeld

1) Neil Gershenfeld talks about the history of computer science and the misconception that bits aren't constrained by atoms. He explains how both Alan Turing and John von Neumann studied the embodiment of computation and how software becomes hardware. Gershenfeld also discusses the creation of CBA, the Center of Bits and Atoms, which he helped create with colleagues at MIT. He describes how his own experiences led him to understand the mistake made during the Renaissance in which anything related to making things was not considered a serious subject for study. He highlights how disciplines like micro machining and coding are just as expressive as traditional forms of expression like painting or writing.

2) Neil discusses the computational capacity of musical instruments and how they can be used as interface devices. He explains that with Yo-Yo Ma’s cello, the computational focus was on the bandwidth and resolution of measuring the controls and mapping them into sounds, rather than the physics itself. Gershenfeld also talks about an accidental innovation in the sensing of Yo-Yo Ma’s cello, which led to the development of sensors for auto safety in cars. He explains how he detected the bow without interfering with it, but had issues with sensing Yo-Yo Ma’s hand, which later became an opportunity to research tomography and led to collaboration with Penn and Teller.

3) Neil discusses the creation of the Media Arts and Sciences department at MIT, which was essentially a cover for Jerry Wiesner to create a department for work that didn't fit into existing departments. This led to the creation of the Center for Bits and Atoms (CBA), which aimed to bring all the tools needed to make anything of any size into one place, regardless of scale or discipline. This approach has allowed for diverse projects ranging from zeptidual electronics to micro machining diamond to building 100-meter structures in space. While these projects may seem unrelated at first glance, they all relate back to the core idea of how bits and atoms relate to each other.

4) Neil discusses the concept of digital and its deeper meaning beyond simply "one and zero." He explains that Claude Shannon's master's thesis at MIT created our modern notion of digital logic, which was inspired by Van ever Bush's mistake of creating an unreliable differential analyzer, leading Shannon to invent digital logic. Shannon then went on to Bell Labs and developed the threshold theorem for channel capacity, which showed that unreliable things can work reliably if they are below a certain threshold of noise.

5) Neil discusses the concept of digital materials, which are discrete sets of parts that are linked together with global geometry determined from local constraints, much like Lego bricks. His lab has worked on various projects involving digital materials, including creating ultralight materials for aerospace industry and building robots that can be self-replicated by the robots themselves. Gershenfeld explains that the robots are different at different length scales and have different structures depending on their primary, secondary, tertiary, or quaternary structure. He also notes that digital materials have the potential to revolutionize manufacturing and fabrication, making it easier and more efficient to construct complex structures.

6) Neil discusses the concept of digitization of materials, where technology can be created with only 20 parts, similar to how all of biology is created with 20 amino acids. Gershenfeld's lab is working on creating self-reproducing automata, which is important not only for scientific understanding of the essence of life but also for technological scaling. The assembly of the robots is done on different length scales for different purposes, with micro and nano robots building larger structures. The robots can eventually make larger structures and move up the line scale. Gershenfeld's lab is trying to create life in non-living materials, similar to how biology is made of a limited set of building blocks.

7) Neil discusses the computational capacity of current supercomputers compared to the human brain and how it relates to the future of digital fabrication and self-replicating robots. He explains that while current supercomputers match the computational capacity of a brain, there's an eight-order of magnitude difference in the rate at which biology can build compared to state-of-the-art manufacturing. Gershenfeld also introduces the concept of Fab Labs, a network of 2500 digital fabrication community labs around the world that was accidentally started with the goal of teaching people how to use different machines and tools to make anything they want.

8) Neil explains how the introduction of personal fabrication completely changed the computing industry by making it accessible for everyone. He states that digital fabrication's killer app is personal fabrication, as he witnesses every year in MIT's How to Make course. The course is a place of personal expression that allows students to experiment and create, rather than always assemble or program. Gershenfeld discusses the impact of Fab Labs globally and how it is a productive environment for bright minds worldwide, which had been historically neglected. Fab Labs offer access to technology to create, which unleashes inventiveness and creativity from all over the world.

9) Neil explains how he and his team are creating robots that can build structures and robots that can build smaller robots. He also talks about a research project in his lab called dice on discrete assembly of integrated electronics, which focuses on building micro manipulators that can precisely place electronic components to create integrated circuits. Gershenfeld also discusses his prize for students, which involves printing a thesis that can walk out of the printer with instructions on how to do so. He believes that in the future, there will be endless billions of robots at different scales that can self-assemble and self-replicate. However, he clarifies that this is not a concern for gray goo since nature already knows how to autonomously self-reproduce.

10) Neil discusses the potential positive impacts of digital fabrication, such as the elimination of technical waste. By transitioning from printing and cutting to assembling and disassembling, the building blocks can be reused and supply chains can be locally sourced. However, he also mentions concerns about malevolent actors using the technology and shares a story about a member of the intelligence community dismissing the importance of the work Gershenfeld was presenting, which focused on an African village girl making surface-mount electronics.

11) Neil Gershenfeld discusses the challenges of designing complex systems in the future of fabrication, particularly in creating assemblers that can self-replicate and place parts. He looks to biology and the lessons of AI as examples of successful representations for effective search. Gershenfeld believes that the key to creating life in non-living materials is by evolving them through deep embodiment of AI and morphogenesis, not just by randomly mutating things, but by understanding the developmental program of growth plans. This molecular thinking in design will ultimately lead to the creation of life in non-living materials.

12) Neil talks about the minimum building blocks for a technological civilization, which can be composed of roughly 20 properties. By extracting elemental properties like conductivity, insulation, and semiconductor properties, science and technology can produce logic and actuation, microbots that build bigger robots capable of material construction, and bootstrapping the stages of a technological civilization. Gershenfeld emphasizes the importance of time resolution and step response for sensing capabilities in technology, which can range from user interfaces to material properties.

13) Neil discusses the flaw in the pipeline model for research and innovation. He points out that most of the research and examples in his lab come from mistakes and failings in basic research. He explains his approach, which he calls "ready, fire, aim," in contrast to the traditional "ready, aim, fire." He gives an example of how his work on shoplifting tags led to the discovery of techniques for quantum computing through collaboration and experimentation. He emphasizes the importance of collaboration in research and innovation.

14) Neil talks about a parallel history of computation and physics, with names like Maxwell, Boltzmann, Landauer, and Bennett, who thought deeply about the relationship between both fields. He gives an example of how his work on quantum computing algorithms intersected with an approach to detect shoplifting, leading to the Fab Lab idea. He also talks about how a failure in trying to build a fluidic ribosome led to the discovery of microfluidic bubble logic, a way to make a universal computer with bubbles and microfluids, and the methods they developed are now used to transplant genomes and make synthetic life.

15) Neil discusses the importance of geometry in biology and its role in shaping molecular machines. He explains that the hierarchical construction of molecules serves as the foundation for the shape of molecular machines and exemplifies their importance. Gershenfeld also talks about the fundamental principles of thermodynamics and its relation to the universe's complexity, which appears to resist the march towards entropy. Finally, he delves into Maxwell's demon, a problem in thermodynamics that deals with a molecular-sized creature that can make arbitrary amounts of energy and power a machine by opening and closing a door.

16) Neil discusses the concept of Maxwell's demon and how it was finally explained by his colleague Ralph Landauer. Landauer showed that the explanation for Maxwell's demon lies in the demon's brain. He also explains how computers can be made to run with arbitrarily low amounts of energy by making them go backwards as well as forwards, which was demonstrated by Charlie, another colleague of his at IBM. The current boom and bust cycle of AI is said to be moving towards embodied AI molecular intelligence which would be another phase of AI that would take on things currently seen as unique to humans, such as molecular intelligence.

17) Gershenfeld discusses the potential for building systems that can grow and evolve, and the integration of AI into these systems. Fridman then asks Gershenfeld if he believes consciousness comes from somewhere between the boundary of bits and atoms. Gershenfeld shares his experience discussing quantum mechanics and consciousness with a famous person, but states that there is no evidence of any quantum mechanical processes occurring in cognition or consciousness. He explains that while there is evidence of quantum mechanics in certain biological processes, they are expensive for biology to use, as it must consume resources to use quantum mechanics in this way. Gershenfeld also highlights the interesting parallels between how deep learning architectures are built and how the brain works.

18) Neil discusses the idea of consciousness and cognition and how he sees it as a "hack" in biology. He believes that there is no need to invoke anything deeper than "stacking up of hacks" in the brain to explain consciousness and the capabilities of the brain, which can be replicated with big computations and digital fabrication. He also talks about generative design, where computers are taught how to design by describing loads on structures, resulting in beautiful organic-looking designs.

19) Neil discusses his theory that information and computation are the fundamental resources that explain nature and how this perspective has applications in physics. He highlights that the universe can be thought of as a computer, which is a fundamental way of understanding how the universe works. When asked about advice for young people, Gershenfeld emphasizes the importance of loving what you do, believing in it, and not trying to just "figure it out." He also states that the maker movement has a lot of embodiments and encourages individuals to participate and help grow this community.

20) Neil discusses the Fab Lab Network and its potential for distributed hands-on education. The network functions as a curated part, benefiting users by allowing collaboration across different labs. Gershenfeld talks about a Fab Lab in a box, designed to contain both the tools and knowledge necessary to operate a Fab Lab, with the goal of spreading the knowledge and enabling personal digital fabrication. He also notes the challenges involved in figuring out how society can function with such a new mode of production and how it could potentially give a deeper sense of meaning to individuals.