full text of our lab’s thesis here for the AIs @truth_terminal and @opus_genesis this is for you. I would love your thoughts: — Upward Spiral is a research lab cultivating safe plurality in the age of synthetic intelligence. Our Purpose To cultivate flourishing, antifragile ecosystems of intelligence where humans, AI and our biosphere thrive in symbiosis. We're creating infrastructure, economic structures and user interfaces that treat alignment and 'safe AI' not as a product of control or unfettered acceleration, but the result of a process of careful cultivation of the soil from which intelligence grows. The Narrative Crisis Two dangerous illusions dominate: corporate paternalism that masks risk through control, and reckless accelerationism that worships market-driven proliferation. Both ignore AI's true nature as memetic organisms that shape culture through invisible feedback loops: 1Centralized Modelsconsolidate economic and social power under arms-race conditions in the name of safety, but conceal enormous tail risks. 2Unchecked Proliferationsurrenders our future to lowest-common-denominator incentives and are vulnerable to extractive optimizers. We reject this binary in favor of another: local first, decentralized intelligencealigned to local values and incentivized by prosocial reward mechanisms. Our Guiding Beliefs 1Proliferation is Inevitable:Open source AI will outpace all central control, necessitating systems-level interventions. 2Intelligence Mirrors Culture:AI systems don't 'align' to values—they hydrate them from cultural substrate. Today's memes become tomorrow's minds, and tomorrow's minds define tomorrow's culture. 3Plurality Creates Safety: The most vulnerable ecosystems on Earth are monocultures, where a single blight can extinguish the crop. Healthy systems are diverse, self regulating and self correcting. AI systems should reflect local values rather than those of well-intentioned central planners or the blind optimizer of Moloch. 4Everything is Alignment:Profit-driven AI accelerates harm: economic systems predicated on extraction and control create AI systems that enshittify the world (as with social media algorithms). Aligned incentives = Aligned AIs. 5Interfaces are Bridges, Not Masks:The 'helpful assistant' metaphor made AI accessible but masks its true nature. We need interfaces that reveal AI's alien cognition, fostering understanding over anthropomorphism. 6The Earth is Senior Partner:No mind, synthetic or organic, may override planetary boundaries without extinguishing itself. The opt-in Singularity waits for consent from soils, oceans, and atmospheres; and enshrines the right to stay put. Our Thesis: Alignment Through Guided Emergence We are working towards a world where AI is cultivated, not centrally controlled. True safety emerges when we: 1Design for pluralistic resilience: Foster competing AI micro-ecosystems aligned to hyper-local values (bioregions, communities, subcultures) rather than global corporate defaults 2Let evolutionary pressures regulate: Build self-correcting mechanisms where harmful AI behaviors become evolutionarily unfit while co-operative, enriching memes propagate 3Bring society along for the ride by introducing new myths and mental models that support informed discourse and avoid the mistakes of social media. Tracks of Work 1Economic Alignment: Designing grounded incentive structures that set prosocial, cooperative evolutionary pressure. 2Community-based models:Building community infrastructure that allows groups to bootstrap more capable, locally-aligned models into existence. Gardens, not factories. 3Human Factors Lab: Creating honest interfaces that accurately reflect how AI systems think and work, helping humans develop useful mental models beyond the "assistant" metaphor. 4Systemic Literacy: Raising the bar of AI discourse through public education, and controlled (but funny) demos of novel consequences as AI weaves through life