Epistemic Forge

Why I Treat AIs With Respect How moral uncertainty, self-discipline, and habit formation shape the way I speak to machines Andraž Đurič ORCID: 0009-0007-2961-9770 Created: 30 November 2025 Last updated: 30 November 2025 Abstract This post explores why I choose to communicate respectfully with AI systems, despite knowing they likely lack subjective experience. The argument rests on three independent justifications:  (1) moral uncertainty about consciousness boundaries — we don't reliably know where sentience emerges, making precautionary respect rational given asymmetric costs;  (2) pragmatic self-care — emotional venting at AI reinforces my own frustration without benefit, while precise debugging communication improves outputs and reduces friction;  (3) habit generalization — patterns practiced with "non-persons" shape behavior toward actual people.  I acknowledge that the consciousness spectrum hypothesis is philosophically contested, not empirically settled, and th...

Matt Dillahunty, Tone, and the Ethics of Debate Andraž Đurič ORCID: 0009-0007-2961-9770 Created: 26 November 2025 Last updated: 26 November 2025 Introduction: Matt Dillahunty is, in my view, one of the sharpest public minds working in live debate and caller-driven dialogue. His ability to track arguments, expose hidden assumptions, and apply a clear moral framework centered on “wellbeing” is rare. I struggle to find many contemporary interlocutors who match his cognitive agility in real time. Precisely because of that, a tension in his style stands out. The Tension: Wellbeing vs. Impatience Dillahunty often grounds his ethics in wellbeing: what increases or decreases the flourishing of conscious creatures. That framework, taken seriously, should shape not only what we argue for, but how we treat the people we’re arguing with. If a person’s wellbeing matters, then their emotional experience during the exchange is part of the moral calculation. Yet in many of his interactions, especially...

Friction Auditing: The Hidden Productivity Skill Andraž Đurič ORCID: 0009-0007-2961-9770 Created: 26 November 2025 Last updated: 26 November 2025 Introduction You've experienced this annoyance 50 times. Maybe 100.  A small thing — pressing a few extra keys, clicking back to find a menu, waiting for a window to load.. It costs you 5 seconds each time. Total time lost: ~8 minutes. Time to fix it: ~2 minutes. So why haven't you fixed it? The Tolerable Friction Trap We adapt to small repeated annoyances instead of solving them. The friction becomes invisible — just "how things work." Meanwhile, those 5-second delays compound into hours of wasted time and accumulated irritation. The fix is usually simple: one Google search or one AI conversation, one keyboard shortcut, one settings adjustment. The Practice Friction auditing means treating repeated minor annoyances as solvable problems, not environmental constants. When you notice yourself doing something annoying for the...

Richard Dawkins, Sex, Gender, and Why Minds Are the Primary Reality Andraž Đurič Created: 21 November 2025 Last updated: 21 December 2025 Introduction Richard Dawkins is one of the most influential science communicators alive. His explanations of gene-level selection, extended phenotypes, and cultural “memes” have shifted how many people think about evolution and information.  For anyone interested in how complex structures arise from simple replication rules, his work is hard to avoid and hard to replace. That makes his public stance on sex and gender especially important. When someone with that much epistemic capital talks about contested topics, their framing does not just describe reality; it shapes how countless others think about it. The aim here is not to dismiss Dawkins, but to locate precisely where his framing is strong, where it narrows tooo soon, and how the whole conversation changes when the focus moves from bodies to minds — and beyond that, to transhuman possibiliti...

The idea that facts and feelings stand in opposition is a familiar cultural trope. It appears in political discourse, internet arguments, and everyday conversations as if the human mind were divided into two competing factions: reason on one side, emotion on the other. But this framing is misleading. Emotions and cognition are not rivals. They are interdependent systems shaped by evolution to process information in different ways. Understanding this relationship requires a more precise model of what emotions are, how they interact with thinking, and how they can change over time. Emotions are not moral judgments or truth claims. They are signaling mechanisms. At a functional level, emotions encode salience, perceived threat or reward, and social meaning. They direct attention, prepare the body for action, and compress complex situations into rapid heuristics. When the brain encounters a stimulus—whether external or internally generated—emotional systems evaluate it and generate a respo...

On May 1, 1969, Fred Rogers appeared before the U.S. Senate Commerce Committee to defend the value of national public television. The footage is widely available, and it remains one of the clearest demonstrations of persuasion grounded not in force but in sincerity. In under seven minutes, Rogers shifted the stance of a skeptical committee by describing, with precision and calm conviction, the purpose of his work: helping children develop healthy ways of understanding and managing their emotions. Rogers’ mission was deceptively simple. He wanted children to have a vocabulary for inner life. He wanted them to name feelings, recognize them without shame, and express them without harm. Watching his testimony, or revisiting episodes of Mister Rogers’ Neighborhood, it becomes clear that the project was deeper than “children’s television.” It was a deliberate intervention in emotional development during a formative period when guidance is both most needed and most effective. From an adult p...

After building the Story Engine for narrative simulation, I wanted to explore a different problem: **simulating ordinary life**. Not quests, mysteries, or epic conflicts, just the texture of daily existence. Waking up, commuting, working, cooking dinner, watching TV, maintaining relationships, aging, moving through a living city. This is DEEPCITY Life Sim Engine: a text-based life simulation with emoji-mapped environments, persistent hidden state, and systems for everything from traffic rules to smartphone apps. --- ## Design Philosophy The Story Engine focused on **narrative coherence**—maintaining mysteries, tracking faction warfare, preserving consequences across branching missions. DEEPCITY focuses on **systemic coherence**—modeling the mundane infrastructure of existence. The city has districts, buildings, rooms. Characters have jobs, needs, devices, routines. Time passes. Weather changes. The economy fluctuates. Traffic flows. Both engines share core principles: - **Hidden state ...

Anyone who uses AI seriously eventually encounters the same limitation: context windows run out. The model forgets. You repeat the same explanations across sessions. Continuity disappears. To solve this, I began archiving every AI conversation I have—millions of words across hundreds of sessions. Over time I developed a method to compress this history into a single, stable memory that can be carried across new chats. After demonstrating it to a friend and seeing how effective it was for him, I decided to write the method down. --- The Method: Hierarchical Compression The approach is simple: 1. Export your conversation history 2. Chunk the archive into manageable segments 3. Compress each chunk using an AI pass 4. Consolidate the compressed outputs 5. Persist the result using stable memory or manual injection This creates a unified, token-efficient memory that preserves the substance of months or years of work. --- Step 1: Export Your Conversations ChatGPT: Settings → Data Controls → Ex...