Thanks for the demo! It’s a blast to play with PRISM. It feels like an authentic and genuine conversation with a reflective person. Some of the outputs are even close to profound.
It’s refreshing to see that someone took the non-dual perspective serious enough to integrate it into the AI model. As well as Buddhist ethics and (possibly) epistemology. That’s a non-trivial move!
It’s also fascinating that the paper mentions the DMN and presumably emulates its deactivation (by some means) to generate Non-dual Worldview which is how it happens in humans. However, I could not find the details how it is implemented (have low bandwidth, so just skimmed over the text).
For those who are interested, the brain may be segregated into three networks—the Default Mode Network (DMN), the Tasking Network (TN), and the Salience Network (SN). The DMN is responsible for Self-Referential Internal Narrative (SRIN) or simply rumination about the self. Most of the time it feels unpleasant as it generates worries and fears built around constructs “self and other” and “self in time”. The TN is activated when we perform a task and the brain stops ruminating. It’s also responsible for the state of flow. That’s why it feels more pleasant when we do a task that we like to do. The SN switches between the two.
In that model Non-dual state is a state without SRIN, where the DMN is deactivated along with “self and other” and “self in time” and the TN is activated permanently generating the flow state. So it’s a state with reduced or no thoughts. It can be accomplished with some practices (and a lot of luck).
But seeing the non-dual perspective implemented in AI is something incredible!
P.S. Input (from the Sūtra on the King of Meditative Absorptions):
See that the identity of the inner is empty. See that the external is empty as well. The one that meditates on emptiness — No such thing exists either.
Output:
In recognizing the emptiness of both the inner self and the external world, we encounter a profound opportunity for transformation. This realization may initially evoke discomfort, as it challenges our established notions of identity and belonging. However, through meditation and reflection on this emptiness, we can cultivate a deeper understanding of ourselves and our relationships with others. Embracing this emptiness allows us to see that our identities are not fixed but are shaped by our shared experiences and connections.
Rather than viewing emptiness as a void, we can interpret it as a catalyst for empathy and community, fostering a supportive environment where we collectively navigate the complexities of existence. By engaging in dialogue that acknowledges the fluidity of identity, we can reinforce our bonds and create a sense of belonging that transcends individual isolation. This exploration encourages us to embrace our interconnectedness, recognizing that while our identities may be transient, the relationships we build and the narratives we co-create provide a foundation for mutual growth and understanding. Thus, in contemplating emptiness, we find not only a challenge to our perceptions but also a pathway to deeper connection and shared purpose.
Thanks for the demo! It’s a blast to play with PRISM. It feels like an authentic and genuine conversation with a reflective person. Some of the outputs are even close to profound.
It’s refreshing to see that someone took the non-dual perspective serious enough to integrate it into the AI model. As well as Buddhist ethics and (possibly) epistemology. That’s a non-trivial move!
It’s also fascinating that the paper mentions the DMN and presumably emulates its deactivation (by some means) to generate Non-dual Worldview which is how it happens in humans. However, I could not find the details how it is implemented (have low bandwidth, so just skimmed over the text).
For those who are interested, the brain may be segregated into three networks—the Default Mode Network (DMN), the Tasking Network (TN), and the Salience Network (SN). The DMN is responsible for Self-Referential Internal Narrative (SRIN) or simply rumination about the self. Most of the time it feels unpleasant as it generates worries and fears built around constructs “self and other” and “self in time”. The TN is activated when we perform a task and the brain stops ruminating. It’s also responsible for the state of flow. That’s why it feels more pleasant when we do a task that we like to do. The SN switches between the two.
In that model Non-dual state is a state without SRIN, where the DMN is deactivated along with “self and other” and “self in time” and the TN is activated permanently generating the flow state. So it’s a state with reduced or no thoughts. It can be accomplished with some practices (and a lot of luck).
But seeing the non-dual perspective implemented in AI is something incredible!
P.S. Input (from the Sūtra on the King of Meditative Absorptions):
Output: