STATE OF AI

AI is advancing at breakneck speed, reshaping nearly every aspect of society.

HARDWARE:

• Breakthroughs: Nvidia is now the most valuable company globally. At last week’s “Super Bowl of AI,” CEO Jensen Huang declared:

  • “We had AI all wrong last year. Demand is 100x greater than anticipated.”

  • Data centers are evolving into “AI factories” housing our “digital twins.”

  • The paradigm is shifting: instead of retrieving files, computers will now generate “tokens”^(?) to create words, music, art, research, and more.

  • “Time for autonomous vehicles has arrived.”

  • “Physical AI and robotics are moving at such a high pace - everyone should pay attention.”

• Processing speed: in 1939, the fastest computer chip performed 0.6 calculations per DAY. Six months ago: 130 billion per second. Today: 1.35 quintillion per second. Ray Kurzweil has been right about a stunning 30% of his predictions spanning 4 decades – and here’s what he’s saying.

• Electricity: Power is now the primary bottleneck. As Rollie Stanich and others report, exponential breakthroughs in energy are emerging. What will serve as guardrails against AI running wild if this constraint lifts?

SOFTWARE:

• Context Windows: These define how much a model can process at once. ChatGPT began with 3–4 pages. Today, GPT-4 handles 200 pages, Claude handles 500, and Gemini reached 3,000 pages — last year!

• Reasoning: DeepSeek excels in reasoning, is very small (hence mobile-friendly), and is better with non-English data — potentially great for spiritual texts. This has changed things. Since 2017, all leading AI uses “transformer”^(?) based architecture – as that changes, more unexpected leaps could come.

• Algorithmic advances: Moore’s law saw hardware growth doubling every 18 months for decades; Sam Altman writes of a 10x AI multiplier. By 2035, he says that one prompt will be able to marshall the intelligence of all humans put together! 

• Autonomous agents: beyond processing info, agents now make their own decisions. Manus has gained a lot of buzz recently.

• AGI: Artificial General Intelligence, defined as systems capable of nearly all cognitive tasks humans perform, is likely near. NYT’s Kevin Roose notes, “When A.G.I. is announced, there will be debates over definitions and arguments about whether or not it counts as ‘real’ A.G.I., but that these mostly won’t matter, because the broader point — that we are losing our monopoly on human-level intelligence, and transitioning to a world with very powerful A.I. systems in it — will be true.”

SOCIAL:

• Professions: Jobs like software engineers are rapidly dying: AI Writes 95% of the Code for 25% of Startups (Source: Y Combinator)  Many other cognitive jobs will follow that trend.

• Governance: As Ben Buchanan told Ezra Klien: “Governments are not ready for this.” 

• Machines of Loving Grace? Dario, at Anthropic, predicts “treatment of all infectious disease”, “elimination of most cancer”, “doubling of human lifespan”, and more.

META QUESTIONS:

• What happens when we have “a country of geniuses in a datacenter” (ie. when a single datacenter could contain the intelligence of an entire country)? 

• If AGI outpaces labor, what will be the new model of meaning and value beyond employment?

• Is there a middle ground between safety-ism (Anthropic, DeepMind) and acceleration-ism (Marc Andreesen, Musk)? Are governments agile enough to keep up as the frontier shifts??

AWAKIN AI

Awakin AI began by tapping into 25 years of ServiceSpace data. Today, it spans several terabytes, hundreds of bots, and six servers. 

• Peter Russell is co-authoring a book with his AI, built from his past writings.

• Ram Dass Foundation is archiving every recorded teaching. All of Gandhi’s writings and talks are in the Gandhi Bot. Similarly, all scriptures from 1700 world religions are in our Interfaith Bot.

• Phil Clothier is shaping policy with his bot.

• Matthew Brensilver’s college courses are creating a course-specific bot.

• Everyday users are receiving contextually wise responses to personal dilemmas — “What would compassion do?”

Jack Kornfield, spiritual mentor to Sam Altman, called Awakin AI the most inspiring vision of AI he’s heard.

Our broader aim isn’t to be just a chatbot provider — that’s already passé. Nor merely a tech advocate — that too will fade. We seek to bridge:

• Narrow margin goals: Immediate needs of the world.

• Broad margin goals: Building identity-expanding community.

• Infinite margin goals: Regenerative compassion translated into service (in an infinite loop to support narrow/margin goals)

LOW HANGING FRUIT

AI as a Spiritual Companion
Just as a mirror helps us see our physical body, can AI help us as a mirror for the mind? AI can be a great practice ground.

Pivot Towards Everyday Users

• We began by supporting authors and content creators.

• Today, AI is exponentially cheaper than when we started.

• Awakin AI now offers a turnkey solution: select wisdom data, define your bot’s identity (“credo”), and you're set.

Implementation Steps

• Reframe Awakin.ai to focus on personal spiritual bots — or launch a new brand. Tracy and her team at Cisco have volunteered to support this rollout.

• Add features for creators: password-free access, video timestamping, etc.

• Enable voice integration — previously costly, now more affordable. (OpenAI just today entered the fray, challenging industry leader Eleven Labs that we use).

• Support diverse models (e.g. DeepSeek for reasoning or multilingual queries or Gemini for YouTube videos).

• Enhance the chat interface (e.g. clickable follow-up questions).

FIVE-LAYER VISION

As written in this recent post, Weaving Sacred into AI, our vision spans five layers:

• Tech is sustained by a “multiplication of wants” instead of the “fulfillment of needs.” This creates a misalignment between well-being and wisdom. We need to innovate incentive structures that hold our technologies.

• Alongside tech solutions, we need to create a new social operating system in parallel – a new way of coming together. This cannot be done sequentially; it has to be integrated into tech innovations.

• Alongside social innovations, we need to cultivate inner transformation and growth. Again, these wisdom layers have to be done in parallel to all of the above.

More specifically, the vision is at five layers.

1. Data

• Can we build a data-commons? Just like Vinoba Bhave pioneered a “land gift”, what about a data-gift movement?

• We already have 500+ high-quality vertical datasets.

• The edge: integrate oral traditions, indigenous wisdom, consciousness data — a digital version of Sarjah's House of Wisdom? 

2. Models

• Can we build new intelligences by combining diverse datasets?

• Our infrastructure supports many-to-many sharing and experimentation.

• We are focused on “RAG” implementation, but could easily experiment with some “fine-tuned” models.

• Possibility: How could these different vertical intelligences create varied meta intelligences? What intelligence emerges when Aristotle, Gandhi, and Maya Angelou “speak together”? Can we create a many-to-many web of wisdom?

3. Applications

2025 is the year of AI agents — not just responding to prompts, but taking action. Think of it as a combination of multiple prompts that don’t just help you synthesize information but also take action. If you say “I’m hungry” to your Food Agent, it can remember your preferences, find nearby restaurants, and place an order for food delivery.

• Me-Agents: Inner transformation agents.

  • E.g. “I want to learn how to meditate (or forgive, or become kinder).” Meditation Agent pulls from multiple wisdom sources, adapts to your needs, and creates a personalized 21-day challenge.

• We-agents: Collaborative tools. 

  • News Agent to scour the web for “good news” from your value lens, write blurbs, publish it on your website, where the community engages and refines your agent’s value lens.

  • Service Agent that contextually connects givers with needs.

  • Circle Agent streamlines in-person gathering logistics based on your spiritual priorities.

  • A dozen more examples in Birju’s list of 14 agents.

BatGap is a promising pilot community for Me-Agents. Greater Good Science Center can provide practices. ServiceSpace (and 12 other global communities we’ve identified) can prototype We-Agents.

4. Social Fabric

• What are new ways that technology can bring us together? Without this question at the forefront, technology’s bias will be to take us to a cheaper “meta verse”. 

• ServiceSpace is hosting dozens of in-person retreats in 2025 in India, Japan, Vietnam, London, Austria, and the US.

• In aligning with that offline field of wisdom, Awakin AI is considering:

• Volunteer engagement (training bots, creating apps, etc.)

• Nonprofit enablement (helping orgs raise AI competency)

• Thought leadership (quarterly convenings, collaborative writings)

• Public sector partnerships (e.g. with Compassionate City mayors)

5. Collective Emergence:

• When people and intentions converge in the social field, emergence happens.

• Such emergence is non-linear and unpredictable, when held by “laddership”.

• This is the ServiceSpace strength, so the infrastructure is already there. :)

PARTNERS AND CO-CREATORS

We’re launching newsletters and community calls for Awakin AI’s various sub-communities like … (a) users, (b) Bot Hosts, c) Awakin AI Ambassadors, (d) AI + Wisdom practitioners.

Recently, we’ve had more than one conversations with potential collaborators like: 

• Meta Communities – BatGap (spirituality), Spirit Rock (engaged Buddhism), LightEn (library of light), Paws (animal rights)

• Authors/Teachers – Sharon Salzberg, Zulma Ray, Jem Bendell, filmmakers like Maurizio & Zaya Benazzo of SAND.

• Educators – Teach for India, Contentment Foundation Rithm Project, and a lot more.

• AI Groups – People+AI (Rohini/Nandan Nilekani), Evan Sharp’s Feast, and Bo Shao’s Benevolent AI

• Monasteries – DRBU, City of Ten Thousand Buddhas, and others exploring Faith + Tech.

• Researchers – HeartMath, IONS

NEXT STEPS

Tech Layer: We've built the foundation of data and models. Now, we're building Agents. We also sense a need for “AI trainings” for change-makers and nonprofit leaders.

Social Layer: Awakin circles and local events, virtual Pods, and “Moved by Love” retreats are organically spreading their wings. In addition, we have many tools from Smile Cards to heart pins to 21-day diaries to content portals like DailyGood, Awakin Calls, etc. to keep engaging broader arcs of community.

Spiritual Layer: Our “social permaculture” process ensures continuous regeneration of the deeper field.

Stay tuned – the journey is just beginning.

Source: Spontaneous human formation in early 2025