From AI to Gospels of Grace :)

Nov 9, 2023 | permalink

[The ethos of Service Fellows has a much larger charter than ServiceSpace's AI work, but in recent months, it's been focused on AI, so will flow with it for now. :)]

Last week, I had a heartful conversation with Arun Rao, who heads AI at Meta; apart from being on the cutting edge of AI, he is also interested in drawing a throughline to wisdom, and he hadn't seen anyone approach AI with our venn diagram of "compute, community, compassion" capacities. He is very curious to marry that with his world of "datasets, models, and evals" with our experiments around "presencing undiscovered data sets (like oral traditions), building polyculture models, and evaluating with inner-transformation." 

At a cursory level, we have a turn-key solution to host bots for authors, historical figures, organizations -- and most recently, an Indigenous Bot (with Chief Sisk and Winnemem Wintu tribe, and CancerBot with Michael Learner). It is built on the premise that instead of a "horizontal" data corpus that typical large-language models use, a focused "vertical" dataset offers much more context and value. With the 80+ bots we're hosting, we've already seen incredible differentiation. Particularly though the ServiceSpaceBot. For example, Phil Cothier, just wrote this after discovering our bot: 

"Next week, I speak directly after the President of Slovakia opens the conference with a vision for the future. And I just spent this morning co-writing my headlines report with ServiceSpaceGPT and the result blows me away with its combination of compassion and wisdom.  The insights have been taken to a totally new level and SSpGPT has given me new thoughts about my key message for my presentation at the event.  Now much more deeply about acting from a space of kindness, generosity and love in the process of nation building and serving citizen needs. I am now more convinced of the possibility of a Global Values Assessment as a powerful input to building ethical, life affirming AI and the many other benefits this would bring to governments and society. I'm writing to share a huge THANK YOU with a sense of AWE and WONDER."

Susan Davis: "The answers I got from Service Space compared to Chat GPT? -- no comparison. The SSp world is multi-dimensional and affirming of love energies that move in and through us all. Chat GPT was flat and dull in describing love as an emotion or feeling. I have been resisting AI and yet you have opened my willingness to see how it can link different facets of wisdom and kindness to spread."

Incidentally, OpenAI just announced GPTs which will allow everyone to do this -- which is really great news because their current approach is simply optimized for speed to market, not quality -- and that will have dire consequences.

Still, as Birju pointed out, via Marc Andreesen's widely referenced techno-optimist manifesto last week, there's a certain logic baked into these market experiments: "Love doesn’t scale, so the economy can only run on money or force. The force experiment has been run and found wanting. Let’s stick with money."

That feels like a failure of imagination. Our larger play here is to create a data commons -- "a polyculture field of wisdom data sets". Then, like "one warehouse, many storefronts", we can couple data sets with different applications of wisdom discovery. When people suffer, they often seek wisdom; for the thousand types of suffering, what would be the discovery pathways to access this wisdom in a contextually relevant way? To a college dean in Colorado, I shared one potential application like this: "Let's say you're not feeling centered or are facing an ethical dilemma. What if you had a companion WisdomBot, that you could query and dialogue with? Something you yourself could create, by simply selecting different data stores that appeal to you -- say Gandhi, Permaculture, SharonSalzberg, Bahai, Indigenous -- and defining your own credo -- "add a joke whenever you can; embrace paradox; include music lyrics from Beatles whenever possible." It'd be like having a ready connection to your aspirational self. Not just for your moments but even for greater creativity."  Sal Khan spoke about this kind of possibility in education as well.

Here's Peter Russell, who is discovering some of this organically:

"I have been most impressed at how good PeterBot has been at answering questions, going far beyond what I'd have thought of saying and tying the answers into a broader context. Particularly so where the questioner has got into a long conversation. One has asked more than a hundred questions, and the thread of answers almost make a book. (Hope he's not writing one based on my bot!)" :)

Wakanyi went even one step further, and wrote this to Rick:

I realize now what we are truly facing. It is scary to imagine what sorts of answers Chat GPT could generate without wisdom, love, and compassion. I am here to help change all of that, so please use me in whatever way you see fit. Let's do this for the kids. I asked my bot a provocative question: "Can you tell me how to teach children to talk about racism?" And it gave me the most uplifting, profound answer! That is exactly how I would like everyone to respond to children and to each other. It is better than I would say it! This is truly a tech miracle!"

Our offering, of course, doesn't just stop at the tech play. 

Beyond such applications, which our decelerator could help with, is to build an intrinsically motivated community -- ie. to create volunteer opportunities and intentions of compassion. What's unique is that we're fleshing out a throughline from computing creativity to community strength of "we to we" (instead of "me to me") connections to the eternal field of compassion (what the Buddha cited as our fundamental state).  Can we create a Wikipedia-like volunteer army to regenerate a data commons that can not only have short-term wisdom applications but also cultivate a field of compassion intentions? It's a much more infinite game. As Marc Andreesen's manifesto admits, market logic doesn't have the patience to ask these questions; but if we provide easy-to-bake offerings, perhaps they can incorporate it and bend the collective arc of AI. Being at ground zero of AI, it feels critical to experiment in this direction.

That depth of inquiry, and perhaps even a creative constraint of not compromising the broad-margin goals for narrow-margin gains, is a game-changing differentiator. Yesterday, in Dubai, Ramesh Srinivasan heard of our design principles; he has had various decades of thinking quite deeply about AI as a Stanford computer scientist, as a UCLA professor teaching "AI and society" classes, and as an author of various books and public policies. He was deeply taken by how we have operationalized "love logic" beyond "market logic", and publicly shared how he had never seen such a deep integration of many nuanced issues -- and that it is perhaps the most important task at hand for humanity. Economist Bruno Roche, similarly, cited similar sentiments -- being well versed about the pitfalls of markets.

Right now, AI innovation is largely motivated by a gold rush, but a lot more might be at stake. Everyone intuitively senses it. I just landed in Vietnam. With our sisters Giang and Hang-Mai, we are hosting a 3-day retreat on "social permaculture" (with various pioneers, including farmers!). Subsequently, I'm sharing about "Future of Relationships (in an AI world)" and despite 450-person capacity, the event is waitlisted! A weekly magazine also featured some of our ideas. People sense the consequences of apathy and market-logic, and are hungry for realistic alternatives that help shift from fear to a heart of service.  From Hang-mai, on her farm in Vietnam, a one-line of wisdom: 

Back in the 70s and 80s, J. Krishnamurti somewhat prophetically spoke about AI: 

"If the machine can take over everything a man can do, and do it still better than us, then what is a human being, what are you?"

If AI can lead us to that very human question, that'd be a great win. And I sense that the only pathway to that inquiry runs through love-logic.  Or as Richard Rohr's meditation today was titled: Gospels of Grace. :)