Robotics

June 1, 2026 by David Such Leave a Comment

A Chip That Thinks using 7 μW

Most AI research is about making the cortex bigger. This one is about making the spinal cord cheaper.

A team at the University of Michigan has built a tiny computing device that controls a balancing propeller using about seven millionths of a watt. For comparison, the LED bulb in your kitchen burns through about ten watts. The Michigan device runs the control task on roughly a millionth of that power.

That is not a typo. It is the finding of a paper published in ACS Nano in March 2026, and it matters because power is the wall that edge AI keeps running into (or falling off?).

Why Power Is the Whole Game

Most of the interesting AI you read about lives in a data centre. It has a wall socket, a cooling system, and an electricity bill measured in millions. Edge AI is what happens when you try to put that intelligence into a hearing aid, a pacemaker, a drone, a soil sensor, or a pair of smart glasses. You are running off a small battery, or whatever energy you can scavenge from sunlight or vibration.

In that world, every microwatt counts, and there is one component that has been quietly eating the budget for decades: the analog-to-digital converter (ADC). Sensors produce continuous signals, but computers think in ones and zeros. Something has to translate between the two, and that translator is the ADC. It is usually the single biggest line item in a battery-powered device’s power budget.

The Michigan team’s trick is to skip the translator entirely.

Your Brain Has a Real Estate Problem, and Dreams Might Be the Solution

What a strange theory about REM sleep tells us about building AI that doesn’t quietly fall apart

Put a blindfold on a sighted adult and stick them in an fMRI scanner. Get them to feel some textured surfaces with their fingers. Then watch what happens in their visual cortex, the part of the brain that is supposed to be doing absolutely nothing in the dark. It lights up. Not after weeks, not after days, but after about forty-five minutes.

That finding, which has been replicated in various forms since the early 2000s, is one of the more unsettling results in modern neuroscience. The visual cortex is not a quiet, dormant region waiting patiently for the lights to come back on. The moment you stop using it for vision, the neighbours start moving in. Touch first, then hearing. Within an hour, real estate that was supposed to be reserved for processing photons is being repurposed for something else entirely.

This raises an obvious and slightly alarming question. If your visual cortex starts getting taken over within an hour of going dark, what happens every single night when you close your eyes for eight hours?

A neuroscientist named David Eagleman thinks he has the answer, and it is one of the more interesting hypotheses about why we dream that has come along in a while. He calls it the defensive activation theory, and once you understand it, you start seeing the same pattern everywhere, including in some surprisingly broken corners of how we currently build AI.

A chip that controls a balancing propeller on seven microwatts

Every battery-powered device you own has a quiet energy hog in it that nobody talks about. It is not the processor, it is not the radio, and it is not the screen. It is the analog-to-digital converter, the small piece of circuitry that translates the messy real world into the clean ones and zeros a computer can think about. For thirty years it has been the line item that decides how long your hearing aid, your pacemaker, or your soil sensor lasts on a battery.

In March 2026, a team at the University of Michigan published a result that quietly removes that converter from the picture for a specific class of problems. Their bismuth selenide memristor runs a closed-loop control task at about seven microwatts, roughly a millionth of what a household LED bulb pulls. The chip does not run code in any conventional sense. The physics does the arithmetic, and the answer drives the motor directly.

In this episode, we walk through what the device actually is, why removing the converter changes the energy budget by orders of magnitude, and which products land first when microwatt-class intelligence becomes buildable. We talk about hearing aids, implants, environmental sensors, and the small drones that have been waiting for this kind of result for a decade. We also talk about what this chip cannot do, because the press releases tend to skip that part. It will not run a language model. It will not recognise your face. It will run the reflexes underneath all of that, and the case for why those reflexes matter more than the cortex gets credit for is the through-line of the episode. #embeddedAI #podcast

Listen to the Podcast…

May 12, 2026 by David Such Leave a Comment

Spot the Robot Can Reason, but it can’t hold a can of soda

Boston Dynamics and Google DeepMind announced last month that Spot is now running Gemini Robotics-ER 1.6, a high-level embodied reasoning model. The headline is that Spot has been taught to reason. The commercial pitch is industrial inspection: wandering around a facility, reading gauges, spotting spills, deciding what to do when something looks wrong. So far, so good.

Buried in the demo video is a small failure that says more about the architecture than the press release does. Asked to recycle cans in the living room, Spot picks one up sideways. If there is any liquid left in the can, it ends up on the carpet. The problem is using a language model to solve something handled by other more primitive layers in the organic world.

Read the complete article…

May 4, 2026 by David Such Leave a Comment

Why ARC’s $2M Bet Against Priors Is the Wrong Bet for Embedded AI

ARC says intelligence emerges from learning when priors are minimised. Vertebrates disagree and have spent 500 million years proving the opposite.

The ARC Prize is a public competition with $2 million on the table, run by François Chollet (creator of Keras) and Mike Knoop (co-founder of Zapier). The premise is that current AI benchmarks measure what a model already knows, not how well it learns. Their fix is ARC-AGI, a series of reasoning tasks that are easy for humans and, so far, brutal for machines. Frontier models score below 4% on ARC-AGI-2.

The newest version, ARC-AGI-3, swaps static puzzles for interactive game environments where the agent has to figure out the rules by playing, with no instructions and no pre-loaded domain knowledge. Humans score 100%. AI scores under 1%. The gap demonstrates the point. Behind the gap is a philosophy about what intelligence is and where it comes from, and that philosophy is where the Primal Layers framework and ARC differ.

Read Full Article…

April 27, 2026 by David Such Leave a Comment

Why Humans and Robots must Dream

Put a blindfold on a sighted adult and the visual cortex starts being colonised by touch and hearing within forty-five minutes. Not weeks. Not days. Forty-five minutes. This is not a quirk of extreme cases. It is how the cortex works all the time. Every region of the brain is in continuous low-grade negotiation with its neighbours over territory, and the currency of that negotiation is activity. Stop using a subsystem and the neighbours move in, fast. This is the empirical foundation of a hypothesis from neuroscientist David Eagleman called the defensive activation theory: that REM sleep exists specifically to keep the visual cortex active during the eight hours each night when external input is unavailable, defending its territory against takeover by senses that never go offline.

The theory itself is plausible but not yet directly proven. What is proven, and what matters more for engineers, is the underlying principle. A complex system with reconfigurable resources will silently lose capability in any subsystem that is not regularly exercised, even when nothing is actively trying to take that capability away. This is not catastrophic forgetting in the usual machine learning sense, where new training overwrites old parameters. This is something subtler and arguably more dangerous: passive territorial loss in any system that supports continuous adaptation. It shows up wherever capabilities are not being exercised in long-running adaptive AI: rarely-routed experts in mixture-of-experts models, underused sensor pipelines in multi-modal robotics, capabilities that drift out of online reinforcement learning agents over months of deployment. Most current architectures treat their structure as fixed by design. Biology treats its structure as continuously contested.

This episode looks at what defensive activation reveals about a missing primitive in modern AI architecture. Current systems have two fundamental modes, training and inference. Brains have at least three, and the third one, the maintenance mode that operates during REM sleep, has no clean equivalent in the systems we build. We examine what this mode is doing structurally, why generative replay in continual learning is mechanistically closer to dreaming than the field usually acknowledges, and what a telemetry-driven maintenance subsystem might look like for embedded and edge AI. The closing argument is straightforward: if biology has been running this experiment for a few hundred million years and converged on internally-driven activation as the way to maintain a plastic computational substrate, the absence of an equivalent mechanism in our architectures is not a neutral design choice. It is a gap.

Listen to the podcast…

March 27, 2026 by David Such Leave a Comment

Will Robots Evolve into Crabs?

Nature keeps reinventing the crab. At least five times, unrelated crustacean lineages have independently converged on the same compact, flat, modular body plan. Biologists call it carcinisation. Engineers should be paying attention.

In this episode, we look at what the crab’s repeated emergence tells us about the deep constraints that shape both biological and artificial systems. The crab body succeeds not because it is optimal in the abstract, but because its modularity creates a platform for downstream specialisation. The same logic applies to robotic morphology: compact, laterally stable, segment-based designs consistently outperform human-mimicking forms when the selection pressure is efficiency rather than aesthetics.

We extend the analogy into AI architecture, where the Transformer has undergone its own carcinisation, colonising vision, audio, robotics, and protein folding from its origins in language modelling. That convergence reflects shared hardware and training constraints, not architectural perfection. And just as crab-like forms have been lost at least seven times in nature through decarcinisation, the emergence of hybrid architectures signals that the Transformer monoculture may be a local optimum, not a final destination.

The core argument is that convergence signals constraint, modularity enables both convergence and escape, and the platform matters more than the form. Engineers chasing human mimicry or constant architectural reinvention may be solving the wrong problem. Nature solved it by building modular platforms and letting selection do the rest.

Check out our latest podcast on Embedded AI – https://www.buzzsprout.com/2429696/episodes/18910786

February 1, 2024 by David Such Leave a Comment

The “Hard Problem” of Consciousness

Recently we have been working nearly every day with ChatGPT, DALL-E, Midjourney and Bard, and this has led us to question whether these GenerativeAI models are intelligent.

The first problem you come across when trying to answer this question, is that no one agrees on what intelligence is. From that perspective, the question then becomes unanswerable. However, this is a bit of a cop out. A better yard stick of intelligence may be, “I can’t define intelligence, but I know what it looks like when I see it.” Using that criteria, we think that most people would agree that Artificial Narrow Intelligence (ANI) has been achieved. Artificial Narrow Intelligence (ANI), also known as Weak AI, refers to AI systems designed to perform a specific task or a limited range of tasks. Unlike Artificial General Intelligence (AGI) or Strong AI, ANI lacks the ability to apply its intelligence broadly across a wide range of contexts.

October 31, 2023 by David Such Leave a Comment

Using the ToolkitRC WM150 Watt Meter

A Watt Meter is a useful addition to your tool kit if you are using electric motors, servos, or Electronic Speed Controllers (ESCs). There are a few different types available, but for our purposes, the WM150 from ToolkitRC is an affordable option. Specifically, a Watt Meter allows you to measure the current drawn from a power source, typically a battery, and calculate the power being dissipated. This is important because too much current will cause things to start melting and release the magic smoke from components.

In addition to measuring voltage, current, and power, the WM150 can drive a servo/ESC directly. This is handy when you are not sure whether the servo or the microprocessor is causing a problem and for measuring servo power requirements.

Read our full review of the WM150 Watt Meter on Medium.

October 31, 2023 by David Such Leave a Comment

Arduino Library for the Stewart Flight Simulator Platform

The Stewart Platform, also known as a hexapod, motion base or parallel manipulator, is a mechanical system that consists of a platform connected to a fixed base through six independently actuated legs. This arrangement allows for precise and versatile motion control in all six degrees of freedom (DOF): three translational (surge, sway, heave) and three rotational (roll, pitch, yaw).

In the context of flight simulators, the Stewart Platform is utilized to replicate the dynamic movements and sensations experienced by pilots during flight. It provides a realistic simulation of aircraft motion, enabling pilots and trainees to practice flying maneuvers, emergency procedures, and various flight scenarios in a controlled environment. By synchronizing the movement of the platform with visual and auditory cues, flight simulators enhance the training experience and help pilots develop their skills without the risks associated with actual flight.

The Stewart platform is used in car/flight/VR simulators, machine tool technology, animatronics, crane technology, underwater research, simulation of earthquakes, air-to-sea rescue, mechanical bulls, satellite dish positioning, the Hexapod-Telescope, robotics, and orthopedic surgery.

We will be using it to test our drone flight control hardware, IMUs and associated software. You can read the full Stewart Platform article on Medium.