Exclusive Interview with Argo's CEO Bryan Salesky
Table of Contents
- Introduction
- The State of Self-Driving Technology
- The Impact of COVID-19 on Self-Driving
- The Importance of Road Testing
- The Challenges of Predictive Systems
- Common Sense and AI in Autonomous Vehicles
- The Role of Data in Training Algorithms
- The Future of Self-Driving Technology
- The Partnership with Automakers
- The Business Model for Self-Driving Cars
- The Role of Engineers in the Development Process
Introduction
Welcome to another episode of the Technium ride home! In this episode, we have a special guest, Brian Celeste Key, the founder and CEO of Argo AI. Argo AI is a company that focuses on self-driving technology and has gained a lot of Attention in the industry. Brian joins us today to share more about the unique business model and strategy that Argo is exploring, as well as the Current state of self-driving technology.
The State of Self-Driving Technology
The first question that comes to mind when discussing self-driving technology is the impact of the COVID-19 pandemic. Brian explains that Argo AI has been off the roads since mid-March, prioritizing the health and safety of their employees. However, as restrictions ease, they are slowly starting to resume vehicle operations in certain cities. Brian acknowledges that the shutdown has caused setbacks in terms of research and on-road testing but emphasizes the importance of continuing to refine their software through simulation and real-world testing.
The Importance of Road Testing
While simulation can provide valuable insights, Brian acknowledges that there is a limit to how much it can replicate real-world scenarios. He explains that the shutdown period has allowed Argo AI to use the data collected from previous on-road testing in simulations, allowing them to Continue making progress. However, he emphasizes that road testing is crucial to validate and fine-tune their software. Despite the challenges posed by the pandemic, Brian is optimistic about the performance of their self-driving vehicles as they slowly return to the road.
The Challenges of Predictive Systems
One of the biggest challenges in developing self-driving technology is building predictive systems that can anticipate what will happen several seconds into the future. Brian emphasizes that humans have evolved to be incredible predictive systems, able to pick up on subtle cues and anticipate the actions of others. However, replicating this ability in computers is a difficult task. While progress has been made in using data and algorithmic techniques to train these predictive systems, it still requires a substantial amount of manual curation and validation.
Common Sense and AI in Autonomous Vehicles
Another crucial aspect of self-driving technology is the need for "common sense" in autonomous vehicles. Brian agrees with the Notion that AI systems still lack true common sense and that human intuition plays a significant role in predicting and reacting to real-world situations. The challenge lies in developing algorithms that can make reasonable decisions Based on the available information, even when faced with unpredictable or unusual scenarios. Brian explains that obtaining and curating the right data is essential in refining these algorithms and ensuring their effectiveness.
The Role of Data in Training Algorithms
Brian highlights the importance of data in training self-driving algorithms. He explains that the industry often overemphasizes the need for vast amounts of data, but the focus should be on obtaining the right set of variations and scenarios. Argo AI uses both real-world data and simulations to train their software. However, Brian cautions against drowning in data and emphasizes the need for careful curation, analysis, and validation to ensure that the algorithms are learning the right things.
The Future of Self-Driving Technology
When asked about his expectations for the future of self-driving technology, Brian shares his optimism that full automation is possible. He believes that the industry is on the right path, but it will take time to develop the necessary technologies and address the complexities of human interactions on the road. Brian emphasizes that self-driving technology is constantly evolving, and there will always be room for improvement and further innovation.
The Partnership with Automakers
Argo AI has formed partnerships with automakers, such as Volkswagen, to bring self-driving technology to market. Brian expresses his satisfaction with the partnerships and acknowledges the challenges that automakers face in selling large volumes of vehicles in urban areas. He believes that self-driving technology offers a different way of providing personal mobility without the need for individuals to own expensive assets. These partnerships aim to leverage the strengths of both Argo AI and automakers to Create innovative business models for self-driving vehicles.
The Business Model for Self-Driving Cars
Brian acknowledges that finding the right business model for self-driving cars has been a topic of exploration and discussion. Argo AI's approach is centered around unit economics, where the goal is not to own expensive assets that are underutilized most of the time. Instead, they aim to provide a platform where other businesses can build on top of their self-driving technology. While the specifics of the business model may vary, the underlying principle is to offer personal mobility without the burden of vehicle ownership.
The Role of Engineers in the Development Process
Brian reflects on his experience working in the railroad industry, where he first encountered the complexities of mission-critical systems and the importance of software in their operation. He emphasizes the multidisciplinary nature of developing self-driving technology, with engineers from various backgrounds working together to find innovative solutions. Brian believes that teamwork and collaboration are key to tackling the challenges in the industry and continuing to improve the technology.
Highlights
- Argo AI has been off the roads since mid-March due to the COVID-19 pandemic, but they are slowly resuming vehicle operations in certain cities.
- Road testing is crucial for validating and refining self-driving software, as simulation can only replicate real-world scenarios to a limited extent.
- Developing predictive systems that can anticipate future actions is a significant challenge in self-driving technology.
- True common sense in autonomous vehicles is still a work in progress, and human intuition plays a crucial role in predicting and reacting to real-world situations.
- The right data, rather than vast amounts of data, is crucial in training self-driving algorithms effectively.
- The future of self-driving technology is promising, but it will take time to address the complexities of human interactions on the road.
- Partnerships with automakers are essential in developing innovative business models for self-driving cars.
- The business model for self-driving cars aims to provide personal mobility without the burden of vehicle ownership.
- Teamwork and collaboration among engineers from various disciplines are crucial in developing self-driving technology.
FAQ
Q: When will self-driving cars become a reality?
A: The progress of self-driving technology is constantly evolving, and it is difficult to predict an exact timeline. However, significant advancements have been made, and full automation is considered a possibility in the future.
Q: What is the role of data in training self-driving algorithms?
A: Data is essential in training self-driving algorithms. It provides the necessary information for the algorithms to learn and make informed decisions. However, it is crucial to curate and validate the data to ensure its accuracy and effectiveness.
Q: How do self-driving cars handle unpredictable or unusual scenarios?
A: Handling unpredictable or unusual scenarios is a significant challenge in self-driving technology. Developers are working on creating algorithms that can make reasonable decisions based on the available information, even in such situations. However, further research and development are needed to refine these algorithms effectively.
Q: What is the business model for self-driving cars?
A: The business model for self-driving cars aims to provide personal mobility without the need for individual vehicle ownership. Companies like Argo AI are exploring partnerships with automakers and developing platforms for other businesses to build upon their self-driving technology.
Q: What is the role of engineers in the development of self-driving technology?
A: Engineers play a crucial role in the development of self-driving technology. Their expertise from various disciplines, such as software engineering, electrical engineering, and mechanical engineering, is essential for finding innovative solutions and improving the technology. Collaboration and teamwork are key in tackling the challenges in the industry.