Rodney Brooks’ Pioneering Role in Robotics and AI:
Rodney Brooks is a trailblazer in the fields of robotics and artificial intelligence (AI). He is one of the founding parents of both fields and has achieved rare distinction in both domains.

Early Contributions to Robotics:
When Brooks entered the field of robotics, there were only three mobile robots of consequence globally. He later founded a company that has produced millions of robots, saving lives and improving household cleanliness.

Early Contributions to AI:
Brooks also co-founded a company that became a leading provider of AI development tools in the 1980s and early 1990s. He balanced his entrepreneurship with a storied academic career, leading a prominent AI center at MIT for many years.

Focus on Historical Context:
The interview delves into the history of robotics and AI, providing context for understanding their present and future. This historical perspective is crucial for comprehending the false starts and great leaps forward in these fields.

Self-Driving Cars:
Brooks discusses self-driving cars, highlighting their intersection with robotics and AI. He offers concrete and verifiable predictions about the timeline for self-driving cars, diverging from overly optimistic forecasts.

Employment and Automation:
Brooks has a unique perspective on the relationship between employment and automation. He is more concerned about a potential labor shortage than job displacement, emphasizing the need for workers to perform jobs that remain after automation advancements.

Super AI Risks:
Brooks’ views on super AI risks differ from those of many prominent tech commentators. He takes a less alarmist stance, arguing that the risks are not as dire as some believe.

Audience Growth and Support:
The interview concludes with a discussion of the podcast’s audience growth and sustainability. Brooks encourages listeners to spread the word about the podcast on social media to increase its reach. He emphasizes the importance of audience support in ensuring the podcast’s continued existence.

Early Life and Fascination with Robotics:
Rodney Brooks grew up in Adelaide, Australia, where access to technology news was limited due to the slow arrival of British magazines by steamship. From a young age, he developed an obsession with robotics and computing, building simple electrical circuits, robots, and computers in his garden shed. By the 1970s, he had advanced his skills, building printed circuit boards and accumulating a surplus of transistors for complex circuit constructions.

University Education and Mathematical Background:
Brooks attended Flinders University in Adelaide, which had a strong mathematics faculty composed of refugees from the Prague Spring in Czechoslovakia. He received a classical Eastern European mathematics education, taking 39 out of 41 classes in mathematics during his four years at the university.

Access to Computing Resources:
Despite the limited computing resources available at the time, Brooks had exceptional access to the university’s mainframe computer, the IBM 1130, for 12 hours every Sunday. This dedicated access allowed him to gain extensive experience and knowledge in computer science, putting him ahead of many contemporaries.

Choice of Stanford University:
When Brooks decided to pursue higher education in the United States, he applied to MIT, Carnegie Mellon, and Stanford. He ultimately chose Stanford due to its proximity to Australia, as determined by looking up the locations of Pittsburgh and Palo Alto in an atlas.

State of Robotics in 1977:
In 1977, when Brooks arrived in the United States, the field of robotics was still in its early stages. The number of mobile robots in the world was minimal, with only a handful of research labs actively working on their development.

The Pioneering Robots of the Time:
There were only three notable robots in existence: Hilaire at LAS in Toulouse, France; a rover at Jet Propulsion Lab in Pasadena; and the CART at Stanford AI Lab.

The CART:
The CART was a unique robot at Stanford AI Lab, not an acronym but simply referred to as “the CART.”

Rodney Brooks’ Work at Stanford AI Lab:
Brooks assisted Hans Moravec with his PhD thesis experiments on the CART. Brooks experimented with a 28-hour day cycle to accommodate his work and social life.

The Limited Capabilities of the CART:
The CART’s movement was slow, taking 15 minutes to move one meter. It scanned the floor to avoid obstacles during its movements.

The State of AI in the 1950s:
AI had been around for over 20 years, with the Dartmouth seminar in the 1950s as a significant event. John McCarthy, the founder of the Stanford AI Lab, was a prominent figure in the field. AI research was conducted in small pockets, mainly at universities, rather than in dedicated departments.

The Role of John McCarthy in AI:
John McCarthy co-founded the AI labs at MIT and Stanford. He named the field of artificial intelligence and was one of its pioneers. Brooks had the opportunity to learn from McCarthy, a leading figure in AI research.

The Significance of LISP in AI:
LISP, developed by John McCarthy, was the standard language used in AI research at the time. It remained the dominant language in AI for a considerable period.

Lucid’s Origins:
Rodney Brooks, while a professor at Stanford, created a software Lisp system for Sun workstations in 1983. In 1984, Brooks teamed up with others to establish Lucid, a company that would develop a software version of Lisp.

The Advantage of Software Lisp:
Brooks believed that Lisp implemented in software would eventually surpass Lisp implemented in hardware due to the rapid advancements in general-purpose computing.

The Proliferation of Workstations:
Sun Microsystems, founded by Brooks’ friend Andy Bechtolsheim, produced workstations that were more powerful than personal computers. These workstations could perform various tasks, unlike specialized LISP machines.

The Rise of Lucid:
Lucid offered a software Lisp package compatible with various computers, including Sun workstations. The company received funding and began operations, with Brooks contributing his “hacky” compiler as the initial foundation.

Brooks’ Continued Involvement:
Despite leaving Lucid to join MIT as a faculty member, Brooks maintained involvement in the company’s development. He would send code updates to Lucid via 40-megabyte cartridge tapes delivered by Federal Express.

Lucid’s Internal Culture:
One developer at Lucid, tasked with integrating Brooks’ code into the build system, had a less-than-flattering assessment of Brooks’ compiler, referring to it as “Bertha: Brooks extremely ran the twisted hack assortment.”

Lucid and Lisp:
Rodney Brooks and his team created Lucid, a Lisp that ran on 19 different platforms. This made Lisp accessible to a wider audience and contributed to its popularity during that era. Brooks’ involvement with Lucid spanned from its inception to its eventual demise.

Academic Career and AI Lab:
Rodney Brooks had a distinguished academic career at MIT, focusing on building custom mobile robots and teaching Lisp. He eventually became the director of the AI lab, overseeing significant developments in robotics and artificial intelligence.

iRobot and Entrepreneurial Ventures:
In 1990, Brooks co-founded iRobot with Colin Angle and Helen Grainer. The company initially lacked a business model but bootstrapped its operations by selling robots before they were built. iRobot faced numerous failed business models before achieving success with the Roomba vacuum cleaner and military robots.

My Real Baby and Manufacturing:
One of iRobot’s early ventures was My Real Baby, a robotic doll in partnership with Hasbro. Brooks emphasized the importance of low-cost manufacturing, reducing the number of motors in the doll’s face without compromising its appearance.

Publicity and Commercial Success:
In 2002, iRobot gained significant publicity by sending a robot into the Great Pyramid, revealing an empty cavern. The same year, the Roomba vacuum cleaner was released, becoming a household name and a commercial success. iRobot also deployed military robots, known as PacBots, to Afghanistan for reconnaissance purposes, which eventually saw widespread use.

Impact and Legacy:
iRobot’s Roomba became a household name, with millions of units sold worldwide. The company’s military robots played a significant role in detecting improvised explosive devices (IEDs) during the wars in Iraq and Afghanistan. Brooks’ contributions to robotics and AI, both through his academic research and entrepreneurial pursuits, have had a lasting impact on the field.

Founding Rethink Robotics:
Rodney Brooks founded Rethink Robotics in 2008, driven by two key experiences. In the late 90s, Brooks observed labor shortages in China after the Golden Week holiday. As director of CSAIL, he collaborated with Taiwan-based companies facing labor problems in mainland China.

Labor Shortages in China and the US:
Brooks noticed labor shortages in China as early as 2003-2005, despite the common belief of infinite labor in China. He heard from manufacturers that they had difficulty finding enough workers despite advertising and scholarship programs. In the US, Brooks observed an aging manufacturing population and a lack of young people interested in factory work, leading to a looming labor shortage.

The Need for Robots in Factories:
Brooks concluded that robots were necessary in factories to address labor shortages in both China and the US. While robots were already present in car factories, they were often expensive and complex to program. Brooks believed there was a need for simpler, more affordable, and easier-to-use robots for a wider range of manufacturing tasks.

The Impact of Robots in Factories:
Brooks highlights the positive impact of robots in factories, such as helping shut down the Fukushima reactor after the disaster and saving the lives of countless soldiers. He also mentions the use of iRobot robots for tasks like cleaning pools, mowing lawns, and cleaning floors.

Background on Traditional Industrial Robots:
Traditional industrial robots were enclosed in cages to protect humans from potential hazards. This separation between humans and robots made it difficult to integrate robots into assembly lines.

Baxter & Sawyer’s Innovative Design:
Baxter and Sawyer, robots created by Rodney Brooks, were designed to address the challenges of human-robot interaction in manufacturing. They employed series elastic actuators, which allowed them to sense and respond to external forces. This enabled the robots to operate safely in close proximity to humans without the need for cages.

Force Perception and Proprioception:
Baxter and Sawyer utilized force perception and proprioception to replicate human movements and tasks. This allowed them to perform tasks such as inserting nuts onto bolts without precise positioning.

Adaptation to Changing Environments:
The robots were able to adapt to variations in the environment and perform tasks without the need for precise alignment. This flexibility made them suitable for a wide range of tasks in manufacturing settings.

Easy Training and Task Teaching:
Baxter and Sawyer were designed to be easily trained by humans. Instead of programming specific trajectories, users could teach the robots tasks by physically moving their arms and demonstrating the desired actions.

Human-Like Behavior and Safety Features:
The robots were equipped with a screen that displayed a face, allowing them to express emotions like surprise. They also looked before reaching, similar to human workers, to avoid unexpected movements. The robots’ ability to sense external forces ensured that they stopped immediately upon contact with humans or objects, enhancing safety.

Impact on the Robotics Industry:
The introduction of uncaged robots, as demonstrated by Baxter and Sawyer at the Automate trade show in 2013, marked a significant shift in the robotics industry. This innovation paved the way for increased collaboration between humans and robots in manufacturing environments.

Safety in Robotics:
In 2013, there was a shift in the perception of robot safety at trade shows. Initially, safety concerns led to restrictions on uncaged robots. However, after a demonstration of safety, one uncaged robot was allowed, and by 2017, there were thousands of uncaged robots on display.

PLCs in Industrial Automation:
Programmable logic controllers (PLCs) were invented in 1968 to control industrial processes. They use abstractions like coils and 16-bit numbers to mimic mechanical relays. PLCs are still widely used today, even in modern factories like Tesla’s.

Friction Points in Robot Deployment:
PLCs can be a friction point for companies new to automation. Rethink Robotics aimed to reduce this friction by using behavior trees on robots to replace the role of traditional automation engineers and PLCs.

Customer-Focused Approach:
Rethink Robotics focused on identifying and reducing friction points in robot deployment for their customers. They identified PLCs as a common source of friction and developed a solution to eliminate the need for them in many cases.

Rodney Brooks’ Concerns About Elder Care:
Rodney Brooks expresses concerns about the lack of assistance for elder care in the near to intermediate future. He emphasizes the importance of independence and dignity for elderly individuals. He highlights the significance of driver assist features in cars, allowing elderly people to drive safely longer. Brooks discusses the need for devices that assist with getting into and out of bed, maintaining independence for elderly individuals. He mentions the lack of one-to-one mapping of caregivers to care recipients and the challenges of managing care for simple tasks.

Rodney Brooks’ Observations on Elder Care Robotics:
There is significant research in Japan focused on developing robots for elder care tasks. Research institutes showcase prototypes of these robots at trade shows, indicating progress in the field. Brooks notes that there are a few places in the US conducting research in this area, including entrepreneurs he is meeting with. He anticipates a flow of venture capital and other investments into elder care robotics in the next 5 to 10 years.

Rodney Brooks’ Predictions for Self-Driving Cars:
Brooks predicts that driverless taxi services with dedicated pickup and drop-off points in restricted areas may emerge by 2022. He believes that arbitrary pick-up and drop-off taxi services in major metropolitan areas are unlikely to be available before 2032. Brooks cautions against mistaking demos for real-world implementations, emphasizing the need for genuine functionality.

Reactions to Rodney Brooks’ Predictions:
Some people in the field may consider Brooks’ predictions to be timid or conservative. Brooks acknowledges receiving pushback from industry professionals regarding his cautious timeline for self-driving cars.

Common Misconceptions About the Development of Self-Driving Cars:
Many people believe that self-driving cars are a recent development, but the technology has actually been in development since the 1980s. This misconception leads to an overestimation of the rate of development and the belief that self-driving cars will be widely available soon.

The Challenges of Real-World Driving Conditions:
Self-driving cars are being developed to operate in ideal conditions, such as freeways and well-marked roads. However, real-world driving conditions are often more complex and unpredictable. This includes situations where the car has to cross double yellow lines, drive on one-way streets, and navigate temporary road closures.

The Ethical Implications of Self-Driving Cars:
Self-driving cars raise a number of ethical questions, such as who is responsible in the event of an accident. Another ethical concern is whether self-driving cars should be allowed to break the law in certain situations, such as when the only way out of a traffic jam is to drive the wrong way down a one-way street. There is also the question of who should be allowed to override the car’s decisions, such as a 14-year-old passenger who wants to go to soccer practice.

Legal Implications and Liability:
As self-driving cars become more common, there are legal implications and liability concerns to consider. Who is liable in the event of an accident involving a self-driving car? Is it the passenger, the car manufacturer, or the software developer? There are also questions about whether self-driving cars should be allowed to break the law, such as stopping in no parking zones or running red lights.

Social Interaction and Pedestrian Safety:
Self-driving cars lack the social interaction that human drivers have with other drivers and pedestrians. In crowded urban environments, drivers and pedestrians rely on eye contact and nonverbal cues to communicate and navigate safely. Without a human driver, self-driving cars may not be able to interpret these cues and could lead to accidents.

Edge Cases and Complex Situations:
Self-driving cars must be able to handle a wide range of edge cases and complex situations. These include distinguishing between pedestrians who are about to cross the road and those who are simply standing on the sidewalk. Self-driving cars must also be able to interpret temporary signs and signals, such as handwritten no parking signs or police officers directing traffic.

General Thoughts on Self-Driving Cars:
Self-driving cars will not solve traffic issues immediately. There will be a long transition period with mixed human and self-driving cars. The cost of deploying software updates is much lower than deploying physical updates.

Virtually Parked Cars:
Self-driving cars could potentially address the challenge of finding parking spots by “virtually parking” – driving around the block while waiting for their owners. This capability has the potential to be abused, such as by parking in front of someone’s driveway or double-parking.

The Trolley Problem:
The trolley problem is a thought experiment that poses a hypothetical situation in which a person must choose between two morally difficult options, often involving saving one person at the expense of another. Rodney Brooks believes that the trolley problem is an interesting intellectual exercise but is unlikely to occur in real-world driving scenarios. In practice, drivers typically react to accidents by applying the brakes as quickly as possible, rather than making calculated decisions about who to save.

Amara’s Law:
Amara’s Law highlights our tendency to overestimate the short-term impact of technology while underestimating its long-term significance.

Overestimation in the Short Term:
People tend to overestimate the capabilities of AI in the near future, expecting it to revolutionize everything within a few years.

Underestimation in the Long Term:
Conversely, people underestimate the long-term potential of AI, assuming its transformative effects will occur within a relatively short timeframe.

Cognitive Distortions:
Cognitive distortions, such as radical breakthroughs and rapid progress in recent years, contribute to the overestimation of AI’s short-term capabilities.

Historical Perspective:
Rodney Brooks’ long experience in AI, dating back to the 1970s, provides a broader perspective, highlighting the gradual and incremental nature of progress in the field.

Starting Point Problem:
The rapid advancements witnessed in recent years create a false impression of accelerated progress, leading to misconceptions about the true pace of AI development.

Background of Deep Learning:
Deep learning is based on neural nets, initially proposed in 1943 by McCulloch and Pitts and further explored in the 1960s. The breakthrough algorithm called backpropagation, developed in the late 1970s to early 1980s, enabled the updating of weights in neural nets based on example results. Backpropagation was initially overestimated and later overshadowed by other machine learning algorithms.

Rise of Deep Learning:
Around 2012, deep learning gained significant attention and success, led by researchers like Yann LeCun and Geoff Hinton. Deep learning outperformed expectations and became a dominant force in various fields, including computer vision, natural language processing, and speech recognition.

Unpredictability of Scientific Progress:
The rapid progress of deep learning was unexpected and cannot be assumed for other areas of research. Scientific and technological advancements often take a long time to materialize, and there is no guarantee that any particular technology will achieve similar success.

Predictions for the Future of Deep Learning:
Rodney Brooks predicts that the hype surrounding deep learning will eventually fade, and investors will realize the need for more than just adding deep learning to existing technologies to achieve profits. The emergence of the next big thing in AI beyond deep learning is anticipated between 2023 and 2027, with many potential contenders.

Cognitive Distortion and Superintelligence Risk:
The idea that advanced technology can appear like magic, as suggested by Arthur C. Clarke, may lead to cognitive distortions in perceiving AI’s potential and the risks associated with superintelligence. The lack of prototypes or examples of superintelligence makes it difficult to fully understand its capabilities and potential risks.

Newton’s Marvel at an iPhone:
If Newton were introduced to an iPhone, he would be amazed by its capabilities, such as its screen, camera, and internet access. The iPhone’s ability to project moving images and sounds would astound Newton, who was familiar with prisms but not modern technology.

Limited Understanding of an iPhone’s Power Source:
Newton would likely be puzzled by the iPhone’s limited battery life, as he has no concept of devices requiring external power sources.

Omnipotent Perception of AI:
When presented with a list of abilities, Newton would struggle to discern which the iPhone could perform and which it couldn’t due to its advanced nature. This highlights the difficulty in understanding the boundaries of super AI due to its unprecedented capabilities.

Perception of AI as Functionally Omniscient and Omnipotent:
The advanced nature of super AI makes it appear almost magical and beyond comprehension. This leads to the misconception that super AI possesses limitless knowledge and capabilities.

Potential Side Journeys in AI’s Path to Omniscience:
The journey towards super AI achieving omniscience and omnipotence may involve numerous detours and challenges. Performance versus competence is considered a key factor in this journey, emphasizing the need for AI to not only perform tasks but also understand the underlying concepts.

Chess Playing Programs:
Chess playing programs, despite their superior performance, lack the ability to teach or explain chess strategies to humans. They can only interact through winning or losing games.

Image Labeling Programs:
Programs that label images may produce detailed descriptions, but they lack an understanding of the concepts and context behind the images. They cannot answer questions about weather, objects, or the significance of objects.

Exponentialism in AI:
The belief that AI technology follows an exponential growth pattern is common, but it is not always accurate. Market saturation and technological limitations can hinder exponential growth.

S-Curves vs. Exponential Curves:
Moore’s Law and other technological advancements often follow S-curves rather than exponential curves. At certain points, further improvements become difficult or impossible.

GPS Accuracy:
GPS technology has improved significantly, but there is a limit to the accuracy needed for practical applications like driving a car.

Predicting Technological Advancements:
Predicting the future of technology can be challenging. Early projections for iPods and online music services underestimated market saturation and the impact of streaming services.

Existential Risks of AI:
The potential existential risks posed by advanced AI systems are a concern for many experts, including Stephen Hawking, Bill Gates, and Elon Musk.

Super AI Risks:
Rodney Brooks, an expert in AI, argues that the risks of super AI are overblown and that we should focus on more immediate issues. He compares the concerns about super AI to the fears people had about hot air balloons in the 18th century. Brooks believes that we cannot predict the specific risks of super AI because the technology is still too far in the future.

Regulation of AI:
Brooks questions the effectiveness of regulating AI without a clear understanding of the risks. He argues that regulations should be specific and targeted at changing specific behaviors, rather than being general and ineffective.

Alignment Problem:
The alignment problem refers to the danger of a super AI having goals that are not aligned with human values. Brooks argues that the alignment problem is not unique to AI but exists in many other technologies and platforms. He suggests that the alignment problem is a more immediate concern than the risks of super AI.

Timing of Super AI Development:
Some experts believe that an intelligence explosion, leading to the development of super AI, could occur within the next 50 to 200 years. Brooks acknowledges that arbitrarily amazing things could be possible in this timeframe but argues that we should wait for more concrete signs before worrying about super AI.

AI and Biological Intelligence:
Rodney Brooks believes we are more likely to see renegade intelligence emerging from biological material rather than pure AI. Biological intelligence, such as brain-machine interfaces, edited animals, or organoids, could pose earlier existential risks than pure AI. Synthetic biological intelligence, using CRISPR and design, is a potential concern that we don’t fully understand.

Early Warnings of Dangerous AI:
Self-awareness, intentionality, and a sense of time are signs of advanced intelligence that we don’t see in current AI systems. Without ongoing existence, planning abilities, and understanding, AI cannot pose a significant threat.

Regulation and Alignment:
We need to regulate AI and synthetic biological intelligence as we do other technologies, except for guns. Addressing alignment issues in existing technologies is more important than focusing solely on future AI risks.

Rodney Brooks’ Perspective:
Rodney Brooks’ deep understanding of technology and robotics provides comfort in his lack of concern about super-AI risks. However, the vehement disagreement among brilliant minds on this issue warrants caution and further consideration.

Factory Worker Shortages and Robots:
Rodney Brooks reported factory worker shortages in China as early as the 1990s. China’s aging population will further strain its economy, requiring help from automation. Rodney’s robots are designed to enhance human productivity rather than replace them, creating a symbiotic relationship.

ATM Machines and Bank Tellers:
The rise of ATM machines initially threatened bank teller jobs, but tellers shifted to higher-value tasks and became more valuable. A similar trend could occur with factory workers collaborating with robots, increasing their value and job opportunities.

Rodney Brooks’ Exponentialism:
Rodney Brooks’ concept of exponentialism is explored, focusing on the tendency to assume that steady improvements in certain areas will continue exponentially, which is often not the case.

Punctuated Equilibrium in Various Fields:
Moore’s law does not apply to all fields, as demonstrated by the history of air travel and robotics. Periods of rapid development are often followed by long periods of relative equilibrium. In robotics, the emergence of uncaged robots after decades of caged robots illustrates this phenomenon.

Neural Networks and Backpropagation:
Rodney Brooks highlights the seemingly sudden emergence of neural networks and backpropagation as an example of how a long-term development can appear sudden to outsiders.

George Church’s Work in Genomics:
George Church’s contributions to synthetic biology and genomics are discussed. The rapid advancements in DNA sequencing and synthesis are noted, following a steep cost improvement curve. Despite this rapid progress, punctuated equilibrium is still observed in this field, with long-term developments appearing sudden to outsiders.

Rodney Brooks’ Patreon Content:
Additional insights into Rodney Brooks’ thinking can be found in his Patreon content. The speaker recommends the bonus content for patrons, which includes discussions on Brooks’ blog posts and a wider range of topics.

Spreading the Word about the Podcast:
The speaker encourages listeners to help spread the word about the podcast by retweeting, sharing on Facebook, or personally recommending it to friends. The future of the podcast after June 30th is mentioned as being dependent on listener support.