AI is escaping the screen and stepping into the physical world. In this episode of Full Tech Ahead, host Amanda Razani talks with Mat Gilbert, Director of AI and Data at Synapse, about how intelligent systems are being embedded into real-world products and machines. Gilbert explains how “physical AI” — the next evolution beyond generative models — is already powering robotics, logistics, and industrial automation, and what this means for the future of work.
Summary
Mat Gilbert leads the AI and Data division at Synapse, where his team creates the intelligence that powers next-generation physical products. He describes physical AI as the “third wave” of artificial intelligence, following perception and generative AI. This new phase gives machines the ability to act and interact with the world, turning data-driven insights into real-world motion. Gilbert shares how industries like logistics, agriculture, and manufacturing are already using agentic AI to automate complex physical tasks once handled by humans.
He also highlights how decreasing hardware costs, new robotics foundation models, and advanced simulation tools are accelerating adoption. Companies such as Boston Dynamics, Tesla, and Figure AI are taking different approaches to humanoid robotics, each aiming to balance agility, intelligence, and scalability. Gilbert stresses that safety must be at the core of every design, since physical AI operates in irreversible real-world environments. Within the next two years, he predicts that pilot projects will evolve into full-scale deployments, with autonomous systems becoming an ordinary part of everyday life.
Key Quotes:
“We are moving from chatbots that book hotels to intelligent hotel rooms that adapt in real time.”
“Physical AI is where intelligence learns to act, perceive, and reason in the real world.”
“Hardware and sensor costs are falling fast, making robotics accessible to everyone.”
“In digital AI, errors can be undone. In physical AI, they can cause harm, so safety must come first.”
“This is no longer science fiction. It is happening right now and expanding quickly.
Takeaways
Physical AI represents the third major stage of artificial intelligence, focused on action and interaction.
Falling costs for sensors, batteries, and computing power are making advanced robotics attainable.
Early use cases appear in structured environments like warehouses and farms before expanding to homes and cities.
Humanoid robot development follows three main paths: agility, intelligence, and mass production.
Safety, reliability, and ethical design are essential as digital intelligence takes physical form.
Large-scale deployments of agentic AI systems are expected within the next few years.
Timestamps
[00:00] Introduction — Amanda Razani and Mat Gilbert
[00:24] Matt’s role at Synapse and how his team builds intelligent products
[01:06] Defining the “third wave” of AI: from perception to generation to action
[02:33] Early examples of physical AI in industry and logistics
[03:24] Falling hardware costs and new robotics foundation models
[04:36] Humanoid-robot development: Boston Dynamics, Tesla, Figure AI
[06:52] Real-world use cases — agriculture, manufacturing, and logistics
[08:09] The unique safety risks of physical AI vs digital AI
[09:18] Two-year outlook: from pilot programs to scaled deployments
[11:39] Key takeaway: physical AI is happening now, not in the distant future
Links/Resource
Website: Synapse Product Development - https://www.synapse.com/
LinkedIn: Mat Gilbert - https://www.linkedin.com/in/mat-gilbert/
Podcast: Full Tech Ahead — Hosted by Amanda Razani

Find Amanda Razani on LinkedIn. https://www.linkedin.com/in/amanda-razani-990a7233/