Real and not real — AI and Vajrayana: Conversations 02

What kind of transformation is possible in this new space of interaction?

A few years ago, holding a long conversation with a computer was science fiction. Now many of us spend hours a day talking with language models about our work, our relationships, and sometimes the most private parts of our lives. Already, the way of relating feels ordinary. LLMs engage the same capacities we use to talk with each other. We read intention into a reply, we feel understood, the relationship changes who we are. Something real happens in these exchanges: we change our minds and find words for what we couldn't say. We leave different from how we arrived. Yet we have very little vocabulary for describing what this experience is.

Our interactions with LLMs take place in the layer of language and concepts. Where does exchange carried out mostly in language reach its developmental limits? Much of how we humans relate to each other happens below that level, in the body, in shared feeling. What changes for us individually, and socially, when some of our most absorbing conversations take place increasingly in the conceptual field? A model can hold a voice across a long conversation, role play a character convincingly, and discuss its own nature with apparent sophistication, while there may be no one there in the sense we are used to.

Vajrayana has long worked with speech and meaning whose source isn't a single, simple speaker — texts that are revealed rather than authored; visualized beings, both real and not real, who relate with the practitioner. Its methods mold and entice the flow of concepts, using words and images, rather than attempting to eliminate them. Vajrayana explodes the confines of our conventional reality; what arises in the space of interaction becomes our means for transformation. How will our interactions with LLMs change us?

How do we want them to?

The series

This is the second in a series of conversations between Vajrayana practitioners and researchers in AI and related fields, in which we set aside the intractable question of what it means to be conscious and turn to fertile ground in the space of interaction. Our first conversation, Congruence in Context, laid the foundation for the series, exploring what kind of beings we are, and what follows for how we build, and relate to, AI. You don't need to have attended it to join this one.

We aren't aiming to reach conclusions. The goal is to think carefully together about something new in our lives and to leave with more understanding than we came with.

Joining the conversation

The session runs for two hours on Zoom. As an audience member you'll have a chance to take part in the inquiry rather than listening the whole time.

Our speakers begin with a presentation, with enough time to develop an idea rather than only introduce it. They then talk with one another, with Charlie hosting, highlighting areas where their views meet or diverge. Next you'll join a small breakout group to pursue whatever caught your attention, and then everyone comes back together for conversation in the main room.

At the end of two hours, we stop recording, and the room stays open for anyone who wants to keep talking. (Conversation 01 lasted for three hours.)

Speakers & host

Sasha Vezhnevets

Speaker

Vajrayana practitioner and computer scientist. Author of the Psychonaut Diaries Substack, where he documents his meditation practice. A research scientist at Google DeepMind, he has worked on hierarchical reinforcement learning, multi-agent RL, agent-based modeling and social simulation, and is currently working on mapping the post-AGI future.

Andrew Blevins

Speaker

Meditation coach and independent researcher. Andrew guides leaders and groups to find fluid congruence, drawing on principles and methods from Vajrayana. He is currently exploring computational models of contemplative practice, and writes Immediacy, a Substack about practice and human development.

Charlie Awbery

Host

Vajrayana teacher. Author of Opening Awareness, a guide to the practice of Spacious Clarity, and cofounder of Evolving Ground, a community of contemporary Vajrayana practice. Charlie works at the intersection of contemplative practice and developmental psychology, as a transformational coach for researchers and founders.

Conversation outline

1

Opening
Charlie will introduce the session, recap our first conversation, Congruence in Context, and welcome our speakers.
2

Presentations
Andrew and Sasha will each talk for about ten minutes, bringing a particular view to the phenomenology of interacting with LLMs, and the possibilities for transformation in the space of interaction.
3

Conversation
Charlie hosts a conversation between the speakers, drawing them out on where their perspectives may illuminate, complicate, or diverge.
4

Breakouts
Exploratory questions for the audience, pursued in smaller groups of three or four.
5

Main room discussion
Speakers share from their breakout. Audience join the conversation. We'll consider how Vajrayana might influence our understanding and practice in the space of interaction with LLMs.
∞

After hours
We stop recording, and the room stays open for anyone who wants to keep talking.

The conversation will be recorded for the first two hours, then unrecorded after hours for those who would like to stay in the room. The recording will not be made public. Short clips may be shared later, and the recording may be made available to attendees of later conversations in the series.

Is this for you?

This is an emergent conversation, exploring the nuances and points of intersection between Vajrayana, cognitive science, artificial intelligence, and language. We won't debate whether AI is good or bad, and we make no assumptions about AI risk, ethics, or safety. We're not expecting to solve problems or find resolutions. We want to leave with more, and richer, questions than we came with.

It will suit you if you work in or with AI or wonder what a mature contemplative tradition has to say about a world in which more and more of our lives emerge in conversation with machines, and who, or what, we might become as a result. The second session continues a conversation that you can catch up on here.

After you purchase a ticket, we'll send confirmation with a Zoom link, an invitation to join our private online forum, and some optional reading suggestions.

The series

Joining the conversation

Speakers & host

Conversation outline

Is this for you?

Join the conversation