IoT Unplugged

The episode surfaces a genuinely interesting concept - bio-inspired neuroplastic AI for constrained edge hardware - and includes some real numbers, but it spends several minutes on company naming backstory and repeated platitudes about cloud limitations. The useful idea-per-minute rate is moderate, not dense.

“we typically say from 100 times to 400 times more compute efficient”

“I'm more just trying to find the essence of the circuit, the essence of the feedback, the essence of how the signal processing works”

The neuroplasticity-as-AI-architecture angle is a genuinely uncommon framing in B2B IoT discourse, and the deer-birth analogy offers a fresh first-principles intuition pump. However, the broader argument (cloud AI doesn't fit edge constraints, biology is more efficient) is not entirely new ground.

“If you've ever seen a deer being born, a deer can learn to walk within an hour. What that tells you is that the pattern for walking is already embedded in the deer's brain”

“all biological systems, Our networks are not static. It's not like the weights and connections in our synapses are fixed. Actually, they're plastic. They change and adapt over time”

Matthew Carr has legitimate technical depth - post-doctoral work at JET fusion reactor, a physics background, and is an actual co-founder who has shipped working code - but the company is clearly early-stage and he cannot name a single customer, which limits the 'at-scale practitioner' credibility.

“I then moved to the UK, about 10 years ago now for my post-op working at Jeff's, the world's largest fusion reactor, which is an amazingly sort of complex machine where you have, you know, 10,000 sensors”

“we managed to make an AI controller that could sort of improve the energy efficiency of that machine by almost 10%”

There are real concrete numbers scattered through the episode - 10 kilobytes, 1% of a Raspberry Pi 4 core, 10% energy saving in injection moulding, 100 - 400x efficiency range - but the wide efficiency range, inability to name any customers, and undisclosed pricing undercut the evidentiary weight.

“a network that was around 1% of a single core of a Raspberry Pi 4”

“It's typically about 10 kilobytes. So absolutely tiny in memory footprint”

The host lands one genuine and specific pushback on the ROI question, which surfaces a useful clarification; however, the vast majority of responses are unchallenged and punctuated with 'That sounds amazing' and 'Wow,' and the host never probes the 400x efficiency claim for independent evidence or methodology.

“a saving of between 1% to 8% doesn't sound that much for, I mean, obviously it adds up, but it doesn't sound that significant if it's towards the 1%. What sort of capital outlay would you need to get that 1% saving?”

“That sounds amazing”