Who Are You?

Hello! I’m Peter. I am a tech worker at 1Password, disability advocate, and student of the brain.

Since 2018, I’ve worked to advance the state of research and care for people with neurological conditions. My work depends on engagement with the neuroscientific literature, particularly around topics like consciousness, intelligence, the neural basis of agency, motor planning, self-monitoring, and sensory perception.

It also emerges from lived experience: I’ve definitely learned more about the brain from talking to people with Parkinson’s and FND than from the research corpus! And I live with several neuro conditions myself, which motivate me to want to improve things for folks in similar situations.

In that capacity, I’ve been lucky to collaborate with amazing teams at places like University College London, the American Academy of Neurology, the FND Society, and Toronto Western Hospital’s Movement Disorders Clinic (the largest and highest-ranked movement disorders clinic in Canada).

To date, I’ve contributed mostly behind closed doors, with research groups I trust, and when engaging with the public am usually quoted anonymously. I suppose that’s changing now.

What’s This?

Computing Naturally is a place to tackle interesting questions around natural and artificial intelligence.

Questions like:

• What is intelligence?
• Why do LLMs mess up things humans can do easily?
• Is the brain a computer?
• What math best describes how brain cells update their responses to new stimuli?
• Can neuroscientific techniques be applied to retrain dysfunctional cellular systems in the peripheral body?
• What are the minimum requirements for a “self?”
• How do we know if an AI is conscious?
• What mechanisms might make AI truly governable and transparent?
• How can we connect our understanding of intelligence across levels of scale in the body, from cellular to network to whole-brain to interpersonal?

And so on.

I’m no Grand Master of Intelligence. I don’t even remotely have things figured out. This page will reflect my own thinking and learning and probably a good amount of messing up as I try to work through things.

I think this is worth doing, because we’re living in a moment in which “intelligence,” at least of the AI type, is ostensibly something our culture is obsessed with. It’s all over the place! You know from your own experience. They keep putting it in your apps whether you want it or not. And it might legitimately change our lives in huge ways.

And yet: we’ve also mostly side-stepped the question of what intelligence is.

When we say something is intelligent, what does that mean?

What’s happening under the hood of a truly intelligent agent that makes it different than, say, a thermostat, or a laptop, or even a supercomputer like those at the Argonne National Lab?

That’s what this site is about: the actual nature and dynamics of intelligence.

It’s a good moment for this. We’re at a convergence point, where natural and biological intelligence are entering into dialogue:

• We can usefully deploy computational models of the brain in order to simulate healthy brain function as well as neurological and psychiatric disorders,
• We know that natural and biological systems can talk to each other: brain-computer interfaces allow people to walk again. Brain cells in a dish learned how to play Pong. We have mushroom-robot hybrids now.
• And we are likely arriving at the point where we can instantiate biologically-plausible models of intelligence into software – potentially presaging the arrival of actually-intelligent AI that far outstrips the wonky “reasoning” capabilities of current LLMs.

In short: intelligence matters, and it’s all around us. This is my attempt to think clearly about it.

So this new project will explore why cognition – ie, processes we think of as “brain stuff” – is pervasive in nature (not just in brains), and how we can use cognitive frameworks and tools to better understand health conditions, our everyday experiences as humans, and our connections to nature. And I think it’s likely that there are some problems that only cognition – ie, natural computation – can solve.

The natural world also serves as my personal reference point when thinking about artificial intelligence. I think that machine learning, generative models, swarm dynamics, cytoarchitectural gradients, and predictive coding are all relevant because they’re facets of a unifying prism: the study of intelligence, now. Natural and artificial intelligence provide useful contrasts to each other, and facilitate unconventional ways of assessing the nature of both.

In pursuit of that, here are some topics we might get into:

• Computational neuroscience
• Theoretical neurobiology
• Bayesian predictive coding
• Active Inference
• Modeling of consciousness and selfhood
• Diverse intelligence and basal cognition
• Design principles of intelligent agents

Needless to say, this is hardcore nerd stuff, but I plan to have fun with it and hope it’ll add something productive to the dialogue.

I’ll do my best to learn and show my work. It’s gonna be weird.

Thanks for reading. See ya out there 👋

Note: the views expressed here are mine, and don’t reflect the views or policy of my employer, the research and clinical groups I work with, or their university affiliates.