Note: This post is an invite for the retreat, as well as an expression of interest for similar events which will be conducted. We are using the form for both.
TL;DR: Ooty AI Retreat 2.0 (June 23-30, '25, open to all): We're moving beyond e/acc vs doomer debates to practically test AI tools & opportunities. We'll cultivate post-rational flexibility (multiple perspectives, intuition, meditation) through coding, writing sprints, strategy talks, and more. Interested? Fill the form! (deadline : June 17th)
Hey folks,
We're running our Ooty AI Alignment Retreat 2.0 from June 23-30, 2025. The last one happened in June 2024. This is open to non technical people also!
When we say x risks, we don't only mean just threat models and p doom/takeoff discussions. We want to work on identifying and navigating the opportunities that current AI capabilities unlock, striving for nuanced understanding over simplistic e/acc versus doomer dichotomies. We welcome people from all kinds of backgrounds, worldviews, and hope to have a productive discussion. Crucially, we anchor these explorations with quick empirical tests and tight feedback loops with reality and today's models—engaging in a 'near mode' rather than getting lost in purely abstract futures.
When we say post rationality, we are cultivating the ability to hold multiple, sometimes even contradictory, mental models or "systems" at once, and fluidly switching between them depending on the situation. This often involves re-engaging with things previously dismissed as vestigial like intuition, emotions, embodiment, or even spiritual concepts—not as blind belief, but as potentially useful sources of information or ways to operate.
As the memetic environment becomes increasingly adversarial, having tools like metacognitive skills and contemplative technique are important to ensure coordination (alignment) with ourselves across time, integrating the different mind parts, giving us stability to handle these turbulent times better.
To get a sense of what happened last year,
Last year's events
Some of the sessions that we had were,
- The future of Brain Machine Interfaces, discussion on bandwidth constraints, etc
- Talks on Centaurism, Cyborgism and integration between AI and humans
- Sahil gave an introduction to his agenda of Live Theory
- Concrete predictions on the future of AI via manifold, forecasting and calibration
- Long walks around the pristine nature in Ooty
- Microphenomenology - how to surface deep preferences
- Funding landscape in AI safety, fellowships, mapping the opportunities
- Poem writing session with facilitation
- Discussions around emotional intelligence, processing experiences
- Horocruxes for Humanity - how to prep for disaster
- Hackathon - vibe coding individual projects inspired by Live UI
This year's plan
The high level agenda revolves around these questions
- What does alignment mean when using AI tools that are available now -
- Sonnet 4, O4 mini high, deep research, perplexity, Elicit. We pick tractable coding and writing projects and use frontier models to assist us.
- Unpack what it means for the collaboration to feel adversarial, deceptive or incompetent and how can we mitigate it.
- Introducing tools like Threads (and the wider ecosystem of live machinery) that enhance conversation, getting feedback from you all on what the crucial prayers are that you have for integrating with AI?
- Sense-making AI trajectory and how to prepare mentally and physically (similar to Superposition | Improving the discourse on the trajectory of AI), how can we remain agentic as we get pervasive cheap intelligence?
- How can contemplative techniques like meditation help cultivate integrity, embodiment, etc? Specifically how can we use AI while remaining grounded, not letting the tool use you.
Overall we plan to maintain a healthy balance between talking and doing.
A more detailed day by day breakdown can be found below
Tentative Schedule
We have various themes,
- Social events (hiking, board games, movie night, cooking together)
- Writing sprint (peer support to finally get your ideas out into the world, find your voice before we start trusting AI to write for us)
- Vibe coding (building tools that enhance epistemics and empower agency)
- Cyborgism (build a personal app that you can integrate into your life, noticing rituals, culture, and bending)
- Autostructures (High Actuation, demos, ecosystem, design interfaces)
- Strategy (forecasting timelines, threat models from developments in AI and how to prepare( Horcrux for humanity))
- PostRat (Meditation, Circling, Buddhism, MAPLE)
The plan is to do a short taste of all these dimensions and then reallocate our time/attention based on these tests. Which activities are aligned with our skills, interests, and are useful to the group.
Day 1
Kickoff! Settle in, catch up & Introductions. Setting expectations, values and basic rules, going through schedule.
We play ice breaker games, social events, and have lightning talks (5 mins) on what each of us has been working on, curious about, interested in for the last 2 months. We can talk about what we hope to get from others and what we can offer the group.
A walk to see the sunset, and helping each other set up laptop coding environments.
Day 2
A) Knowledge sharing - we go around and talk about which tools (AI studio, windsurf, etc) are useful, each person can demonstrate practically what workflows suit them, prompting strategies, case studies.
B) Vibe coding workshop.
Either you come with an existing idea you are working on or you can propose a new idea and seek collaborators for support by pitching your idea to the group
(Babble Phase) mapping all potential project ideas
Day 3
(Prune Phase) ranking ideas based on votes after discussion.
- Will I use this app? is it practically useful? (direct experience)
- Is it possible for me to build in 2 days? (tight loops)
- How does it empower you? (alignment across persons or person-moment)
Finalise team members and corresponding project goals. Create specification document and start vibe-coding! Goal is to have an MVP by end of the day
Day 4
Culture and Interface Design for Vibe Coding
Start the day by sharing the MVP with the rest of the group for testing and explaining the features that work
Effective writing using AI (Not Slop)
Start with brainstorming - “What articles should I have written which I have not done yet”
15 mins feedback and voting from group
Writing Sprint with AI
15 mins feedback and voting from group
Solo writing without AI
15 mins feedback and voting from group
30 mins of reflections and takeaways
Culture and Interface Design for Vibe Coding
Give feedback to teams. What changes did each of us have to make to use the MVPs easily. Discover what rituals allow these apps to work well with your daily life.
Day 5
Session on Buddhism and how it is linked to post-rationalist thinking - 1 hr talk 1 hr Q&A and discussion
Go hiking, maybe trails in Cain Hill
Practice Meditation, Board games
Cooking together
Last 2 hrs writing block before publish deadline
Creative writing - fiction, poem themes only, nothing factual dry or technical!
We’ll explore the idea of the metacrisis - what it is, why it matters and what it takes to stay clear-eyed, grounded, deeply connected and responsive in a time that demands more from our systems, our communities & ourselves.
Workshop on CFAR inspired rationality techniques (Noticing Confusion, Catching Self-Deception, Bayes theorem)
Movie night, suggestions include
Severence
love death + robots
black mirror
Rick and Morty S8
Akira
Pantheon
Psycho-Pass
Ghost in Shell
Undone 2019
Serial Experiments Lain (1998)
Her (2013)
After Yang
Matrix
LOTR
Day 6
A) Table Top Exercise and Sense making discussions and predictions
A tabletop exercise (TTX) is a discussion-based simulation used to prepare for various emergency or crisis scenarios. It involves a group of participants discussing their roles and responses to a simulated event in a non-threatening environment. TTXs are often used to test and refine emergency response plans, identify weaknesses, and enhance communication and coordination
Will make combined predictions in manifold.
Will discuss Gradual Disempowerment risks and possible mitigation.
B) Autostructures - Talks on Live Theory, AI safety camp demos that were built, how extraction, modification, and distribution all work together in collaboration.
Day 7 (buffer day)
Retrospective analysis of time spent, survey, discussion of takeaways from last 5 days, what worked, what did not, and re-prioritise to optimise time
Memo sessions - proposed by group members.
Parting thoughts, gratitude circling, swag, book giveaways
FAQ
- How can I contact the organisers?
DM one of us in lesswrong or you can contact us using WhatsApp - Bhishma, Aditya, Vatsal.
- What about food, logistics, cost?
We will share them with the selected participants in a dedicated WhatsApp group. We expect the cost to be 500 to 1000 per person per day, if finances are a constraint please let us know we are working to secure partial funding from other sources
- I can can only attend few days, is it ok to attend partially?
Absolutely, we have around 7 people who are already attending the full retreat. While full participation is encouraged for the best experience, we're open to discussing partial attendance. Please detail your availability in the form.
- What is the background of the existing participants?
People from EA/rationality space, founders, participants of AI safety camp
- The title mentions x-risks but I dont see any events talking about it?
Since x-risks conversations come up inevitably, we did not want to over emphasise it. And we do have discussions on takeoff speeds and risk models (table top exercise). We wanted to keep the conversations empirically grounded.