AI, consciousness & awakening | Shamil Chandaria
Shamil Chandaria, is an advisor to Google's Deep Mind, and a philosopher, neuroscientist and meditation researcher in consciousness studies at both Oxford, Cambridge, and Imperial College London.
He explains that according to the neuroscience of perception, if you hold your thumb up at arm's length, your thumbnail covers the only part of your visual field where your eyes are receiving sharp, colour-rich data.
The rest of the vivid world you see is generated based on a best guess. A hallucination continuously updated based on a thin trickle of sensory evidence.
And it's not just what you see. Everything you hear, feel, taste... even your emotions and sense of self are all constructed from scant and messy data.
As Shamil explains it, we are living inside our own simulation. A generative model that organises perceptions based on prior expectations.
In meditation these priors can be discovered, dissolved and even redesigned.
In a world of words... could AI wake up?
It now seems that computers are also processing vast amounts of data and developing sophisticated models of the world, even when restricted to language.
But does that mean they could develop consciousness?
Shamil shared a preview of an upcoming paper describing the five criteria being considered as evidence. From behaviour, to algorithms, to the causal structure of IIT, down to living biology and a 'coupling' between organism and environment.
Bernardo is sympathetic to IIT, but suspicious of substrate independent consciousness. In other words, he leans firmly towards biology as the place where alters emerge. For him, the only compelling empirical evidence we have for inner life is... life. The rest of our dashboard lacks clear boundaries, so his challenge to computationalists is a simple question: what exactly is the thing that's meant to be conscious? The microchips? The cooling system? The river that supplies the cooling water?
Having said that, both agree that AI has only just begun and no one can predict where it will end up.
In the meantime, as we discover more about how computers generate models of reality, sometimes based on language alone, we might be learning more about how we do it, too.
The cosmic joke that never happened
For me, the most captivating part of the conversation was when it turned away from machines waking up and back to humans.
Bernardo shared the one realisation that he has experienced personally: everything comes out of nothing, is made of nothing, and never ceased to be nothing.
His response when this happened was not reverence, but laughter.
Shamil recognised the legitimacy of this response, but insists there is more juice in the apple. He took us deep into a phenomenology of awakening and the silence, stillness, boundlessness and timelessness behind all experience. The awareness that is the host to all things without being one.
Whilst Shamil strongly encourages meditation practice as a way to cultivate insights that translate into human flourishing, Bernardo insists he is at peace with dipping his toe in that lake only occasionally. A kind of meta-peace with everyday suffering.
Paradoxically, it's an orientation that Shamil recommends to most people striving for enlightenment. Nevertheless, as an advocate for taking multiple perspectives, he strongly encourages daily meditation, and the beauty the practice unfolds.
(If you want to explore this more, I'm hosting a series of meditation and Q&A evenings with Shamil, and you can join the next one here.)
I know the meeting never happened, but I look forward to hearing your thoughts!
Amir
We also discuss:
- Access vs phenomenal consciousness - and why the first is a misnomer
- Can language model the world?
- How Schopenhauer came so close
- Why the infinite needs finite perspectives to know itself
- How to meditate like an old cow pees
Upcoming meditation events with Shamil
https://dandelion.events/o/adventuresinawareness/a/shamil-chandaria-events
