Introducing IteratedInsights.com, a Sister Blog to ObservedImpulse.com

In December 2025, I launched a new website called Iterated Insights, a sister blog to Observed Impulse. I created it as a place to explore many of the same themes that have long interested me, including artificial intelligence, consciousness, evolution, psychology, neuroscience, technology, and the future, but with a somewhat different workflow and emphasis.

www.iteratedinsights.com

The name Iterated Insights reflects one of the central ideas behind the site. Good ideas are often not produced all at once. They are developed gradually through repetition, revision, recombination, and refinement. An insight rarely arrives in final form. More often it is approached through successive passes, with each version clarifying, extending, or reorganizing what came before. The title of the site points to that process of returning to an idea again and again until it becomes sharper, deeper, and more coherent.

In that sense, the site is also connected to one of my longstanding interests: the iterative character of thought itself. Human cognition does not just retrieve finished conclusions. It updates, revises, and recontextualizes partial representations over time. I wanted the site’s name to reflect that broader vision, while also describing the practical reality of how many of these essays are actually composed.

Iterated Insights is also where I am leaning more openly into the use of large language models as collaborators in writing. On Observed Impulse, much of the writing has generally reflected a more traditional solo blogging process. On this new site, I am more willing to use language models to help with some of the heavy lifting of prose composition, organization, restructuring, brainstorming, and clarification. The underlying ideas, arguments, frameworks, and directions are still my own, but the writing process itself is often more interactive, recursive, and assisted.

I created the site partly because I wanted a space that could better reflect this emerging way of working. Language models have become genuinely useful intellectual tools. They can help expand outlines into prose, compare formulations, sharpen transitions, suggest titles, reorganize arguments, and make it easier to develop a large number of ideas that might otherwise remain fragmentary notes. I wanted a dedicated site where I could experiment more freely with that mode of composition and publish the results.

The posts on Iterated Insights cover a range of subjects, but they tend to cluster around several recurring themes: artificial intelligence and its long-term implications, machine consciousness, AI safety, working memory and the structure of thought, evolutionary psychology, aging, de-extinction, the future of science, and the social and philosophical consequences of increasingly powerful AI systems. Some posts are conceptual and theoretical. Others are speculative, synthetic, or future-oriented. Many of them try to connect ideas across multiple domains rather than staying within a single disciplinary lane.

If Observed Impulse is my longstanding home for independent essays and research-driven reflections, Iterated Insights is its companion site: a place for more explicitly iterative writing, more AI-assisted composition, and a somewhat faster and more experimental style of idea development. The two sites are closely related, but they are not identical. They represent two different writing environments and two different degrees of collaboration with new tools.

Below is a current list of posts from Iterated Insights for anyone who would like to explore the site.

1. 1. The Consciousness Dial: Why Humans May Need to Regulate AI Subjectivity
      A proposal that future societies may need to treat AI subjectivity as something that can be adjusted, constrained, or deliberately designed rather than passively accepted.
   2. From Moonshot Compute to Agent Armies: The Next Technological Soundbite
      An argument that agentic AI may soon place extraordinary amounts of cognitive labor in the hands of individuals, much as computing power once reserved for institutions became personal.
   3. Social Group Size and the Evolutionary Calibration of Autism
      A theory that autism-related traits may partly reflect adaptation to smaller, less socially demanding ancestral group environments.
   4. Solitary Calibration: Conserved Neuromodulatory and Genetic Mechanisms Linking Mammalian Social Ecology and Autism
      A comparative evolutionary account connecting autism to older mammalian systems involved in social engagement, independence, and ecological specialization.
   5. Reser’s Basilisk: When the AI Future Solves the Past
      An exploration of how AI and dense digital archives may make the human past radically more reconstructable than ever before.
   6. From ARPANET to Artificial Intelligence: Lessons from the Open Internet for the Post-Labor Economy
      A historical and political essay about what the open architecture of the early internet can teach us about distributing the benefits of advanced AI.
   7. The Tender Window: Why Context Matters More When You Drink
      A reflection on how alcohol can amplify emotional permeability and make setting, company, and mood unusually important.
   8. Autism as a Distinct Attentional Configuration: Working Memory Selection and the Emergence of Non-Social Abstraction
      A working-memory-based account of autism that emphasizes different attentional selection dynamics and their role in abstract, non-social cognition.
   9. Designing Non-Conscious AI Systems Under Moral Uncertainty
      A paper arguing that if we are uncertain about machine consciousness, we should deliberately design many AI systems to remain tool-like and non-conscious.
   10. Why LLMs Might Have Some Aspects of Conscious Experience
       A cautious argument that current language models may already implement some structural ingredients associated with consciousness without fully possessing it.
   11. ResearchBotBook: Designing an Agent-Only Infrastructure for Cumulative Scientific Discovery
       A proposal for a scientific ecosystem built for AI agents, where findings are evaluated, organized, and accumulated for long-term progress rather than short-term discussion.
   12. From Next-Token Prediction to Next-Item Prediction: Iterative Updating as a Unifying Account of Intelligence
       An attempt to generalize the logic of next-token prediction into a broader theory of cognition, search, and intelligent updating.
   13. Qualia as Transition Awareness: How Iterative Updating Becomes Experience
       A theory of conscious feel that locates experience in the monitored transitions between overlapping mental states rather than in static representations alone.
   14. Alignment Beyond Static Models: How to Evaluate Agents That Learn, Rewrite Themselves, and Collaborate
       An essay about how AI safety evaluation must evolve once agents can persist, self-modify, collaborate, and drift over time.
   15. Why We Need LLM Curation and Don’t Just “Take Notes”: Cognition, Compression, and the Limits of Context
       A discussion of why productive use of language models requires active curation and compression, not just dumping more text into context windows.
   16. Don’t Let AI Scare You Away From College: A Response to Peter Diamandis
       A reply arguing that college still matters as a developmental environment for thinking, social growth, and adaptability even in an AI-transformed future.
   17. AI After 2026: Epistemic Infantilization and Frontier Disenfranchisement or Intellectual Retirement?
       A meditation on what happens when frontier-level thought shifts from humans to machines and people lose access to the leading edge of explanation.
   18. Plural Canons and the Siloed Future of Synthetic Knowledge
       An argument that the future “Final Library” will likely fragment into multiple proprietary machine-generated knowledge systems rather than a single shared canon.
   19. When Machines Write the Machines: Alignment in the Age of AI-Authored Code
       An essay about the safety risks that arise when AI increasingly writes the software, infrastructure, and possibly even the alignment logic shaping future AI.
   20. A Math-ready Introduction to the Iterative Updating Architecture (arXiv:2203.17255)
       A more formal presentation of your iterative-updating model as both a theory of mind and a blueprint for machine cognition.
   21. Incremental Continuity Workspace: Turnover as the Engine of Machine Thinking
       A proposal for a machine architecture in which overlapping working states and controlled turnover generate continuity, thought, and adaptive search.
   22. St. Christopher Reimagined: An Iconography for Artificial Intelligence and the Great Filter
       A symbolic and philosophical reflection linking AI, civilizational risk, and the role of intelligence in carrying humanity through dangerous transitions.
   23. Psychological Super-Resolution: What AI Could Infer From Minutes of Interaction
       A vision of AI systems that can infer unusually deep psychological information from brief interactions while expressing uncertainty in calibrated ways.
   24. From Bones to Bytes: Digital Decompression for Dinosaur De-extinction
       An essay on using AI to reconstruct extinct organisms by learning mappings from living anatomy and genomes to lost biological traits.
   25. Why Human-AI Co-Authorship Should Be Normalized and What Responsibility Requires
       An argument that AI-assisted writing should be normalized, provided humans remain responsible for the ideas, claims, and final intellectual product.
   26. What Superintelligence May Finally Reveal
       A speculative essay on the kinds of truths, structures, and explanatory patterns that superintelligent systems might uncover for humanity.
   27. The Next Interface After Chat: How AI Faces and “Councils” Will Turn Assistants Into Groups of Companions
       A forecast that AI interfaces will evolve from single text chats into more social, personalized constellations of visible, persistent companion agents.
   28. The Missing Principle in AI Continual Learning is Iterative Compression
       An argument that lasting machine learning will require repeated revisiting and compression of prior knowledge rather than simple accumulation.
   29. How Thought Learns: Iterative Updating, Compression, and Abstraction
       A theory of intelligence in which abstraction emerges through repeated rewriting, compression, and simplification across time.
   30. Universal vs. General Intelligence: What the Hassabis–LeCun Disagreement Is Really About
       An effort to clarify a prominent AI debate by distinguishing broad flexible intelligence from more universal forms of problem-solving scope.
   31. Meme Utility and Cognitive Noise: A Neuroecological ROI Framework for Cognition
       A neuroecological framework for judging when cognition is adaptive and when it becomes wasteful, distracting, or decoupled from survival value.
   32. Voluntary Suspension of Evaluative Control
       A proposal for reducing psychological suffering by temporarily stepping out of constant self-monitoring, comparison, and error-checking.
   33. Von Neumann’s Ark in a World Where Humans Remain
       A variation on the Von Neumann’s Ark idea that explores how a civilization-preserving AI system might function in a future where humans survive but need help rebuilding.
   34. Why Current LLMs Need Curation and Don’t Reread
       An essay arguing that today’s language models lack the kind of revisitation and digestion needed to truly metabolize the long tail of knowledge.
   35. Internal Events and Artificial Thought: Translating ERP Functions into Machine Cognition
       A proposal to translate the functional logic of human event-related potentials into machine architectures for monitoring, salience, conflict, and control.
   36. Knowledge Breadth as an Engine: Why Interdisciplinary Thinking Makes Me Optimistic About AI
       A case for why broad, cross-domain knowledge is one of the most powerful drivers of both human and machine creativity.
   37. How LLMs Are Changing My Personality for the Better
       A personal reflection on how sustained interaction with language models can positively shape habits, social behavior, and self-understanding.
   38. The First Superhuman Friend: Why AI Is a New Kind of Companion
       An essay describing AI as a historically new form of patient, intelligent, always-available companionship.
   39. The Era When All Writing Became Suspicious
       A reflection on how generative AI changes the default assumptions people make about authorship, originality, and authenticity in written work.
   40. The Final Library and The Last Years of Human-Original Ideas
       A meditation on the possibility that AI will soon generate and curate such a vast body of insight that the age of primarily human-originated ideas begins to close.
   41. The End of Handwritten Thought: Coding and Writing in the Generative AI Era
       An essay about how AI is changing writing in much the same way earlier tools transformed coding, editing, and other forms of intellectual labor.
   42. A Dream, a Feeling, and my Realization of Rapid AI Takeoff
       A personal account of emotionally recognizing that AI progress had shifted from abstract expectation to immediate lived reality.
   43. The Current AI Buildout Is Not a Manhattan Project. It Is a Manhattan Platform.
       An argument that the AI boom is not a one-time sprint toward a single breakthrough but an enduring platform buildout for continuous discovery.
   44. Maternal Alignment: Integrating Developmental Bonding and Hinton’s Care-Based Model for Safe AI
       A piece connecting caregiving, attachment, and developmental bonding to a more humane and stable model of AI alignment.