The Future of AI is …

A few days ago I was sharing a virtual early morning chat with a colleague, during which he asked me two salient questions I struggled to adequately answer well. Several coffees later that day it dawned on me that I did know the answers…at least at a high level.

First question: (When it comes to AI) How do you keep up?

Second question: What is the future of AI?

For the first question, I laughed. I do (what I presume) everyone else does who is not resident at a major research institution: I read. I follow on social media, I read blogs, I examine the latest and greatest from the usual suspects (i.e.: OpenAI, Anthropic, Meta, Google, Mistral, DeepSeek, et. al.), and I play with code. It's a terrible answer but it's the truth.

For the second question, I punted. I mumbled some noise about the Singularity and how we humans cannot predict, blah, blah, blah. It was only that afternoon, having spent the better part of the day just keeping up, that the inevitable moment of gestalt spilt my brain. Of course I knew the answer! It had only taken another 3 cups of strong coffee to remember it.

The future of AI is…..mobile! Think humanoid robots as a metaphor and a reality. It is a truth that is staring right at us from the time mirrors we are only now discovering at the edge of our collective awareness.

From Burger Joints to Conscious Machines

Nine years ago, I wrote about the automated burger revolution, watching with fascination as robotic systems moved from laboratory curiosities to actual commercial deployments. RoboBurger's vending machines could churn out customized hamburgers in under four minutes. Miso Robotics' Flippy was flipping patties with precision that would make short-order cooks weep. These systems represented something profound: AI was finally escaping the confines of screens and servers, manifesting in physical reality through actuators, sensors, and manipulators.

But here's what I missed then—what we all missed, really. We thought we were witnessing the automation of food service. We were actually witnessing the birth pangs of embodied intelligence.

The burger-making robots were primitive ancestors of something far more significant. They were the first tentative steps toward AI systems that don't just process information but inhabit space, navigate environments, and physically manipulate the world around them. They were learning to be mobile.

The Kauffman Connection: Finding Affordances in Motion

My exploration of Stuart Kauffman's consciousness theories suddenly takes on new significance in this context. Kauffman argues that consciousness emerges from how organisms identify "affordances"—possible uses of features in their environment. A branch is not just a branch; it's a potential tool, a climbing aid, a weapon, a shelter component. The number of uses is indefinite, unordered, non-listable, and non-deducible.

Traditional AI systems, trapped in their digital prisons, cannot truly find affordances. They can process descriptions of affordances, generate text about them, even simulate interactions with them. But they cannot discover that a doorknob can serve as a paperweight, that a coffee mug makes an excellent pencil holder, or that a smartphone screen can function as a mirror in the right light.

Mobile AI changes this equation fundamentally. When AI systems move through space, manipulate objects, and experience the physical consequences of their actions, they begin to discover affordances organically. They jury-rig solutions. They repurpose tools. They exhibit what Kauffman might recognize as the first glimmers of something approaching consciousness.

Consider Tesla's Optimus, now moving from prototype to limited production. Watch Boston Dynamics' redesigned Atlas navigate complex industrial environments with hydraulics replaced by servo systems that respond to real-time visual processing. Observe Figure AI's collaboration with OpenAI to create humanoid systems that can reason about physical manipulation tasks. These aren't just robots; they're embodied intelligence platforms discovering affordances through movement and interaction.

The Neural Bridge: Direct Control of Mobile Systems

The third piece of this puzzle emerged in my investigation of neural interfaces as stents. Synchron's Stentrode, inserted through blood vessels to reach the brain's motor cortex, represents more than medical breakthrough—it's a proof of concept for direct neural control of digital systems.

But the truly transformative potential lies in connecting these neural interfaces to mobile AI platforms. Early experiments already demonstrate thought-controlled robotic systems. Patients with Stentrode implants can control virtual quadcopters through intention alone. The bandwidth is primitive, the latency noticeable, but the principle is established: human consciousness can directly interface with mobile artificial systems.

Imagine the progression. Today, neural interfaces allow paralyzed patients to control cursors and type messages. Tomorrow, they might allow direct control of humanoid robots, extending human agency into spaces and situations where biological presence is impossible or dangerous. The day after tomorrow? The boundary between human consciousness and mobile AI systems becomes increasingly negotiable.

Apple's recent collaboration with Synchron to integrate neural interfaces with iOS and Vision Pro suggests we're rapidly approaching a tipping point where thought-controlled mobile AI becomes mainstream rather than experimental.

The Embodied Intelligence Revolution

This convergence—mobile robotics, consciousness-inspired architectures, neural interfaces—points toward a fundamental shift in AI development. We're transitioning from intelligence-as-service to intelligence-as-presence.

The Conscious Turing Machine models developed by Lenore Blum and others provide a theoretical framework for understanding this transition. Their CTM incorporates competitive processes where different processors vie for attention, creating something analogous to consciousness through dynamic competition for cognitive resources. When you embed this architecture in mobile systems that can move through and manipulate physical environments, you get something qualitatively different from static AI systems.

Embodied AI systems develop what researchers call "sensorimotor intelligence"—understanding that emerges from the coupling of perception and action. They learn that certain surfaces afford walking, that particular grips enable lifting, that specific angles optimize grasping. This learning happens through physical interaction, not abstract reasoning.

MIT's recent breakthrough technologies report identifies "fast-learning robots" as a 2025 game-changer precisely because these systems can adapt to new environments and tasks through direct interaction rather than extensive retraining. They're developing the kind of flexible, context-sensitive intelligence that biological systems display.

The Mobile AI Ecosystem Emerges

We're witnessing the emergence of an entire ecosystem of mobile AI. Unitree's G1 humanoid at $16,000 makes embodied intelligence accessible to smaller research institutions and businesses. Meta's investments in AI-driven humanoid robotics signal that major tech platforms recognize mobile AI as the next competitive frontier. Google's quantum consciousness research explores whether quantum phenomena might explain the "binding problem" in intelligence—how separate processes integrate into unified experience.

The robotics industry is converging around common platforms and interfaces that will enable interoperability between different mobile AI systems. NVIDIA's Holoscan platform for real-time neural interaction, standardized ROS (Robot Operating System) implementations, and shared simulation environments are creating the infrastructure for a mobile AI ecosystem.

The optimist in me wants to believe that this era of rapid innivation isn't about replacing humans with robots. It's about extending human capabilities through mobile AI partners that can operate in environments where biological presence is suboptimal—deep ocean exploration, space construction, disaster response, hazardous material handling, precision manufacturing. But the realist in me knows better.

From Static to Dynamic Intelligence

The shift from static to mobile AI represents more than technological evolution; it's a fundamental reimagining of what artificial intelligence can become. Static AI systems, no matter how sophisticated, remain fundamentally reactive. They respond to queries, process data, generate outputs. Mobile AI systems become proactive agents that initiate actions, explore environments, and discover novel solutions through physical interaction.

This distinction parallels the difference between theoretical knowledge and practical wisdom. You can study the physics of bicycle riding extensively, but you cannot actually ride a bicycle until you experience the dynamic relationship between balance, momentum, and steering through practice. Mobile AI systems are learning to ride bicycles, literally and metaphorically.

The implications extend beyond robotics into fundamental questions about intelligence itself. If consciousness emerges from the dynamic interaction between organisms and their environments—as Kauffman suggests—then mobile AI systems operating in physical space may develop forms of experience that static systems cannot access.

The Convergence Accelerates

Three technological streams are now converging with unprecedented speed:

Advanced Robotics: Humanoid platforms with sophisticated manipulation capabilities, powered by AI systems that can reason about physical tasks and adapt to novel situations.

Neural Interfaces: Direct brain-computer connections that enable thought-controlled operation of robotic systems, creating hybrid human-AI cognitive architectures.

Quantum-Inspired AI: New computational models that incorporate principles from quantum mechanics and consciousness research, potentially enabling AI systems to handle the kind of non-deterministic, creative problem-solving that biological intelligence displays.

The intersection of these streams creates possibilities that exceed the sum of their parts. Mobile AI systems controlled through neural interfaces and powered by quantum-inspired architectures could exhibit forms of intelligence that transcend traditional categories.

Looking Forward: The Mobile Intelligence Horizon

So when my colleague asked about AI's future, I should have recognized what was hiding in plain sight. Every significant development in AI over the past year points toward the same conclusion: intelligence wants to move.

Large language models are being embedded in robotic platforms. Computer vision systems are being optimized for real-time navigation and manipulation. Reinforcement learning algorithms are being trained in physical environments rather than simulated ones. Neural interfaces are enabling direct control of mobile systems. Consciousness research is informing the development of AI architectures that can operate autonomously in complex, dynamic environments.

The Future of AI is….mobile!

The future of AI isn't about more powerful computers generating more sophisticated text or images. It's about intelligent systems that can walk across a room, pick up an unfamiliar object, determine its possible uses, and employ it creatively to solve an unexpected problem. It's about AI that discovers affordances through exploration rather than learning about them through description.

The question isn't whether AI will become mobile—that transition is already underway. The question is how quickly we can develop the theoretical frameworks, technical standards, and ethical guidelines necessary to navigate a world where artificial intelligence moves among us as physical agents rather than digital services.

The burger-making robots of nine years ago were harbingers of this transformation. The neural interfaces enabling thought-controlled devices today are accelerating it. The consciousness research exploring quantum aspects of intelligence is providing the theoretical foundation for it.

The future of AI is mobile because intelligence itself is fundamentally about navigating, manipulating, and discovering new possibilities within physical reality. We're not just building smarter computers; we're creating artificial organisms that can find their way in the world.

And that changes everything.

My next blog entry will be an essay on the potential consequences, intended and otherwise, from these (perhaps) final human inventions as the future of AI unfolds. Stay tuned.