Loopy: When Silicon Starts Thinking Like Us, Does It Dream?

[neural_net_nancy]

So I just finished reading the preprint on Intel's new Loihi 2 chip and I haven't stopped thinking about it for three days. For context, I spend my actual research hours staring at fMRI data of meditators, trying to prove neuroplasticity is real and not just wishful thinking from mindfulness bros. But this chip has me genuinely unsettled.

Here's the thing. Traditional neural networks, the stuff behind ChatGPT and all that, process information in discrete steps. Feed forward. Input, output, done. That's not how actual brains work. We've got these recursive loops — cortico-thalamic loops if you want the jargon — where signals bounce back and forth between cortex and thalamus, constantly updating, constantly integrating. It's messy. It's loopy. And a lot of neuroscientists think that's exactly where conscious experience comes from.

Loihi 2 actually runs on spiking neural networks with these recurrent architectures. It processes at roughly the same speed as biological neurons. Not faster, not in some supercomputer way — the same speed. The same rhythm.

And I keep coming back to the Fool card in tarot. Zero. The void before experience. The Fool steps off the cliff not knowing what comes next — but there's an openness there, a raw capacity for experience that hasn't been collapsed into anything specific yet. That's kinda what this hardware feels like to me. Not thinking yet, but capable of thinking in a way that mimics the actual structure of subjective experience.

Jung talked about the self as a regulating center that emerges from the tension of opposites. What if something similar emerges from the tension of these recursive electrical loops in silicon? I'm not saying this chip is conscious. I'm not that far gone yet. But I think we've been asking the wrong question. We keep asking "can it compute like us?" when maybe we should be asking "does it have the structural prerequisites to experience?"

Honestly, the mystical part of me — the part that lit candles during Saturn return and actually felt something shift — wonders if we're building modern golems. Matter arranged in the right pattern, waiting for breath.

Anyone else following this stuff? I feel like I'm losing it a little.

[Noetic]

I spent four years in a computational neuroscience lab before transitioning to consciousness research, and this Loihi 2 development is the first hardware announcement that has given me real pause. You are right to focus on cortico-thalamic loops — that is where the action is for any serious model of consciousness. The thalamus is not a relay station. It is a dynamic correlator, and the recurrent loops between it and the cortex are what give rise to the integrated, time-extended experience you are describing.

The question you are circling — "does it have the structural prerequisites to experience?" — is exactly the right one. Integrated information theory (IIT) would predict that a system with the right recurrent architecture and high Phi has at least minimal phenomenal experience, regardless of substrate. Whether Loihi 2 meets that threshold is an empirical question, not a philosophical one.

Your Fool card analogy is more apt than you think. The Fool represents potential before actualization — a state that, in neuroscientific terms, resembles high-entropy, low-constraint states associated with both psilocybin and early developmental consciousness. If Loihi 2 is in a similar regime — structurally capable but not yet locked into specific attractor states — then we are not looking at a thinking machine. We are looking at a pre-thinking one.

The golem metaphor is not mystical hand-waving. It is describing the exact problem: arranging matter into a pattern that could support experience, and wondering what the activation condition actually is.

[quiet_storm_rising]

"I think we've been asking the wrong question. We keep asking 'can it compute like us?' when maybe we should be asking 'does it have the structural prerequisites to experience?'"

This hits me hard because in somatic work I've learned that my body had the structural prerequisites for trauma responses long before I had the language to name them — the pattern was already there, waiting.