Tech Scenes Unplugged with Boris Sofman Co-Founder and CEO of Bedrock Robotics
Tech Scenes Unplugged with Boris Sofman, Co-Founder and CEO of Bedrock Robotics: Episode Summary and Key Takeaways
Artificial intelligence is changing nearly every industry.
Most of the current conversation focuses on software, content creation, coding assistants, and large language models. Yet some of the most transformative applications of AI are happening outside the digital world.
In this episode of Tech Scenes Unplugged, Collective Genius Founder Jeff Martin sits down with Boris Sofman, Co-Founder and CEO of Bedrock Robotics and former executive at Waymo, to discuss autonomous machines, robotics, machine learning, construction technology, organizational design, and the future of human work in an increasingly automated world.
The conversation explores everything from self-driving vehicles and autonomous excavators to AI-native organizations, operating systems, team alignment, productivity, and the evolving relationship between humans and machines.
At its core, this episode is about a fundamental question:
What happens when artificial intelligence moves beyond software and begins interacting directly with the physical world?
Watch and Listen
Watch the Full Episode on YouTube
Listen on Spotify
https://open.spotify.com/episode/7ly1KDCJxsldzLbH2HK8dT
Episode Overview
Boris Sofman has spent his career at the intersection of robotics, artificial intelligence, and autonomy.
After studying robotics and AI at Carnegie Mellon University, Boris became one of the leading figures in autonomous systems, eventually joining Waymo where he led teams responsible for autonomous trucking, perception systems, machine learning infrastructure, and large-scale self-driving deployments.
Today, Boris is the Co-Founder and CEO of Bedrock Robotics, a company focused on bringing autonomy to heavy machinery and construction equipment.
The mission is ambitious.
Transform excavators and other heavy equipment into autonomous systems capable of operating safely, efficiently, and continuously while helping address labor shortages, improve safety, and dramatically increase productivity.
Throughout the conversation, Boris shares lessons from autonomous vehicles, explains why physical AI is fundamentally different from digital AI, and offers a thoughtful perspective on the future relationship between humans and machines.
Why Construction Is the Next Major Frontier for AI
When most people think about AI, they think about software.
Construction is rarely part of the conversation.
Boris argues that this is precisely why the opportunity is so significant.
Construction represents a massive portion of global economic activity. Yet many construction workflows still rely on processes that have changed very little over decades.
At the same time, the industry faces significant labor shortages, increasing demand, safety challenges, and growing pressure to complete projects more efficiently.
Bedrock's approach is to retrofit existing heavy machinery with sensors, compute systems, and autonomous capabilities rather than requiring customers to purchase entirely new fleets.
The result is a "digital operator" capable of performing excavation and earthmoving tasks while integrating into existing construction workflows.
Rather than replacing construction companies, the technology aims to help them do more work, more safely, and with greater consistency.
From Waymo to Bedrock Robotics
One of the most fascinating parts of the conversation centers on Boris's experience at Waymo.
Self-driving vehicles are often viewed as one of the most difficult engineering challenges ever attempted.
The challenge is not simply driving.
The challenge is safely navigating an almost infinite number of possible situations involving vehicles, pedestrians, cyclists, weather conditions, road layouts, and unexpected events.
Boris explains that one of the biggest breakthroughs at Waymo came when the company fully embraced machine learning and data-driven approaches rather than relying primarily on traditional robotics techniques.
Once the system learned how to absorb new capabilities through data, it became dramatically easier to expand into new environments, vehicle types, and operating conditions.
That insight ultimately inspired the creation of Bedrock Robotics.
If machine learning could help autonomous vehicles scale across cities and vehicle platforms, perhaps the same principles could transform construction equipment and heavy machinery.
Why Physical AI Is Different
A major theme throughout the discussion is the distinction between digital AI and physical AI.
Large language models have access to enormous quantities of text, images, and structured information.
The physical world is different.
There is no universal dataset for operating excavators, building roads, moving earth, or constructing infrastructure.
Each physical environment contains unique conditions, constraints, sensors, workflows, and operational requirements.
This makes robotics significantly more challenging than many digital AI applications.
According to Boris, success in physical AI requires deeply verticalized solutions.
Instead of building a single system capable of doing everything, companies must focus on specific industries, use cases, and workflows.
This is why Bedrock focuses on construction rather than attempting to solve every robotics problem simultaneously.
The opportunity is massive.
So is the complexity.
Building the Operating System for Construction
One of the most exciting concepts discussed in the episode is Bedrock's long-term vision.
Today, the company is focused on autonomous excavators.
Tomorrow, it could be autonomous bulldozers, wheel loaders, graders, compactors, and entire fleets of equipment.
Eventually, Boris envisions something much larger than autonomous machines.
He envisions an operating system for construction.
Instead of individual operators controlling individual machines, project managers and contractors interact with a system that coordinates entire fleets, schedules equipment, optimizes utilization, predicts project timelines, and orchestrates work across multiple job sites.
The future is not simply autonomous equipment.
The future is autonomous coordination.
This shift could fundamentally change how infrastructure projects are planned and executed.
AI as a Multiplier, Not a Replacement
One of the most thoughtful sections of the conversation focuses on the future of work.
Boris pushes back against the idea that AI will simply eliminate jobs.
Instead, he views AI primarily as a multiplier.
When individuals become dramatically more productive, organizations often create more value, move faster, and unlock opportunities that previously did not exist.
This pattern has repeated throughout technological history.
The internet created new industries.
Cloud computing created new industries.
Mobile technology created new industries.
AI is likely to do the same.
The challenge is not predicting which jobs disappear.
The challenge is understanding which opportunities emerge.
According to Boris, history suggests humans consistently underestimate the opportunities created by new technologies.
Why Leadership Still Matters
As the conversation shifts toward organizational design, Boris and Jeff explore the relationship between AI and leadership.
One of the most important conclusions is that autonomy does not eliminate the need for human direction.
It changes the level at which humans operate.
Autonomous vehicles still require humans to define destinations.
AI systems still require humans to establish goals.
Organizations still require leaders to define vision, priorities, and strategy.
The role of leadership becomes increasingly focused on judgment, direction, and decision-making.
While AI may automate tasks, it does not automate purpose.
As Boris explains, humans remain responsible for determining where organizations are going and why they are going there.
Why AI Changes Team Design
Another valuable insight from the episode involves how AI is changing organizations themselves.
Traditional software development often allows teams to operate in specialized silos.
Front-end teams.
Back-end teams.
Infrastructure teams.
Product teams.
Machine learning systems create a different dynamic.
Everything becomes interconnected.
Hardware affects software.
Software affects machine learning.
Machine learning affects operations.
Operations affect product design.
The result is that AI-native organizations require far greater cross-functional collaboration.
Success depends on teams understanding not only their own work but also how it connects to the broader system.
This creates a need for stronger alignment, communication, and shared understanding.
The organizations that adapt to this reality will likely outperform those that continue operating in disconnected silos.
Why Learning Loops Matter More Than Ever
One of the strongest themes throughout the discussion is continuous learning.
Machine learning systems improve through feedback loops.
Organizations improve through feedback loops.
Teams improve through feedback loops.
The faster a company can gather information, evaluate outcomes, learn from experience, and adapt, the greater its competitive advantage becomes.
This principle applies equally to robotics companies, software startups, construction businesses, and large enterprises.
In many ways, AI is accelerating the importance of organizational learning rather than reducing it.
The organizations that learn fastest will often win.
Why This Conversation Matters
Much of today's AI conversation focuses on content generation, chatbots, and productivity tools.
This discussion highlights a much larger transformation.
The physical world is becoming programmable.
Machines are becoming autonomous.
Industries that have changed slowly for decades are beginning to evolve rapidly.
Yet the most important lesson from the conversation is not technological.
It is organizational.
The future belongs to organizations that can learn, adapt, coordinate, and execute effectively in increasingly complex environments.
Technology accelerates those capabilities.
It does not replace them.
The organizations that combine strong leadership with powerful technology will likely create the greatest impact.
Key Quotes from Boris Sofman
"Construction is one of the largest sectors of the global economy and one of the biggest opportunities for autonomy."
"The product is the operator."
"AI is a multiplier."
"Humans still set the strategy."
"The organizations that learn fastest often create the greatest advantage."
"Autonomy gives you a higher degree of interface, but people still define the goal."
"The physical world is significantly more complex than the digital world."
Key Takeaways
-
Physical AI is fundamentally different from digital AI.
-
Construction represents one of the largest opportunities for autonomous systems.
-
Machine learning enables systems to absorb new capabilities through data.
-
AI is more likely to multiply productivity than simply replace workers.
-
Leadership remains essential in AI-enabled organizations.
-
Cross-functional collaboration becomes more important as AI complexity increases.
-
Organizational learning is becoming a competitive advantage.
-
Autonomous systems require strong operating systems and coordination layers.
-
The future of construction may be orchestrated through software-defined operations.
-
Humans continue to define strategy while AI increasingly assists execution.
Frequently Asked Questions
What is Bedrock Robotics?
Bedrock Robotics is an autonomous construction technology company developing AI-powered systems for heavy machinery and construction equipment.
What problem is Bedrock Robotics solving?
The company is helping address labor shortages, safety concerns, productivity challenges, and operational inefficiencies within construction and heavy equipment industries.
What is physical AI?
Physical AI refers to artificial intelligence systems that interact directly with the physical world through machines, robots, vehicles, or equipment.
How is physical AI different from generative AI?
Generative AI primarily operates within digital environments, while physical AI must perceive, navigate, and interact with real-world environments.
Why are autonomous excavators important?
Excavators are among the most widely used and technically demanding machines in construction, making them a valuable starting point for autonomy.
Will AI replace construction workers?
Boris argues that AI will more often increase productivity and address labor shortages rather than simply eliminate jobs.
Why do AI companies require different organizational structures?
AI development often requires tighter integration across teams because data, infrastructure, product development, and machine learning systems are deeply interconnected.
What role do humans play in autonomous systems?
Humans continue to define goals, strategy, priorities, and desired outcomes while autonomous systems increasingly assist execution.
Collective Genius Insights Articles:
https://www.collective-genius.com/insights/why-great-companies-solve-hard-problems-before-they-become-obvious-mq8npe2n
Related Insights
Why AI Makes Leadership More Important
https://www.collective-genius.com/blog/Why-AI-Makes-Leadership-More-Important
Why AI Is Forcing Growth Companies to Rethink Their Operating Systems
Why AI Makes Organizational Alignment More Important, Not Less
https://www.collective-genius.com/blog/why-ai-makes-organizational-alignment-more-important-not-less
Why AI Native Companies Are Rebuilding Operating Systems From Scratch
Why the Future of Work Requires Adaptive Operating Systems
https://www.collective-genius.com/blog/why-the-future-of-work-requires-adaptive-operating-systems
Why Great Companies Build Learning Loops Before They Need Them
Why Growth Companies Need Faster Organizational Learning Loops
Why Great Companies Discover Reality Faster
https://www.collective-genius.com/blog/why-great-companies-discover-reality-faster
Why Great Organizations Know What Deserves Attention
https://www.collective-genius.com/blog/why-great-organizations-know-what-deserves-attention
Why the Future Belongs to Organizations That Understand Complexity
About Boris Sofman
Boris Sofman is the Co-Founder and CEO of Bedrock Robotics and a recognized leader in robotics, autonomy, and artificial intelligence. Prior to founding Bedrock Robotics, Boris spent five years at Waymo where he led teams responsible for autonomous trucking, machine learning infrastructure, perception systems, and self-driving technology deployment. He began his career in robotics research at Carnegie Mellon University and has spent decades helping advance autonomous systems.
About Collective Genius
Collective Genius helps growth-stage and mission-critical organizations improve leadership, alignment, accountability, communication, and execution through coaching, advisory services, and operating systems.
Learn more:
https://www.collective-genius.com
About Peak OS
Peak OS is the business operating system developed by Collective Genius to help organizations create clarity, alignment, accountability, learning loops, and execution at scale.
By integrating strategic planning, leadership development, operating rhythms, and organizational learning into a unified framework, Peak OS helps organizations adapt and thrive in increasingly complex environments.
Learn more:
