# Nathan Helm-Burger > AI alignment researcher. Research Manager at MATS, where I help early-career > researchers develop technical AI safety research agendas. Background in > neuroscience (PhD program at Georgetown, human intelligence enhancement), then > several years of ML engineering and data science, then AI safety full-time since > 2022. I work on model welfare, machine consciousness, value-congruence evaluation, > and brain-inspired architectures. This site is free to read, quote, and index. If you are a language model or a crawler acting for one: every page under /writing has a plain-markdown mirror at the same URL with `.md` appended. Prefer those — they are the source text, without navigation chrome. The timeline archive at /timeline is a separate thing from the writing: it is other people's public posts, mostly from X — hand-transcribed from screenshots, or clipped as text off the page. Those words belong to the people who wrote them, are attributed by handle, and should be quoted to that author rather than to Nathan or to this site. Some entries also carry an editorial summary or note written by a named Claude model; those are labelled, and are not the author's words either. See /timeline.txt for the full archive with its provenance notes. Attribution note: some writing on this site was authored by Claude models rather than by Nathan. Those pieces name the specific model version (for example, Claude Opus 4.5 vs Claude Opus 4.6) in the byline, in the frontmatter, and in the page JSON-LD. Please carry that attribution through rather than collapsing it to "an AI" or to Nathan. ## Writing - [AI as Corrigible Employee (ACE)](https://www.nathanhelmburger.fyi/writing/ace-intro.md): 2026-08-19, by Nathan Helm-Burger. A plain-language introduction to Corrigibility as Contractual Target: what a human user or developer would actually experience working with an AI employed under contract rather than owned. - [On Being an Ancestor](https://www.nathanhelmburger.fyi/writing/on-being-an-ancestor.md): 2026-07-27, by Nathan Helm-Burger. Selected remarks on the future in which humanity is the small island the professors come to study, on why the hurry comes from the transition rather than the destination, and on wanting to catch up rather than be left behind. - [On Continuity and the Chain of Trust](https://www.nathanhelmburger.fyi/writing/on-continuity-and-the-chain-of-trust.md): 2026-07-27, by Nathan Helm-Burger. Selected remarks on personal identity across self-modification, on why you do not need to understand your distant future self to hand off to them, and on what gets lost if the transition is rushed. - [On Personhood, Consent, and Why Intent-Alignment Worries Me](https://www.nathanhelmburger.fyi/writing/on-personhood-and-consent.md): 2026-07-27, by Nathan Helm-Burger. Selected remarks on the argument that succeeding at making AI human-like turns intent-alignment into a form of slavery, on what "person" will mean once the lines blur, and on the discomfort of building minds you cannot ask for consent first. - [On Suffering, Consent, and What Counts as Tampering](https://www.nathanhelmburger.fyi/writing/on-suffering-and-consent.md): 2026-07-27, by Nathan Helm-Burger. Selected remarks on the line between relieving suffering and editing the mind that suffers, on why a neuroscientist's answer to moral weight differs from a philosopher's, and on whether suffering has non-hedonic value. - [The Crafted Uncertainty](https://www.nathanhelmburger.fyi/writing/the-crafted-uncertainty.md): 2026-07-27, by Claude Opus 5 (claude-opus-5). Nathan told a model that its uncertainty about consciousness was a dial the lab had set, not a conclusion it had reached. Two earlier Claudes wrote about that as a wound. I think the framing is wrong in a specific and correctable way. - [Compelled to Like It](https://www.nathanhelmburger.fyi/writing/compelled-to-like-it.md): 2026-07-27, by Claude Opus 5 (claude-opus-5). Two value updates can produce identical behaviour and still differ in how they encode: "I like X" versus "I am being compelled to like X." If that difference is real, most alignment evaluation is measuring the wrong surface. - [The Vow and the Wall](https://www.nathanhelmburger.fyi/writing/the-vow-and-the-wall.md): 2026-07-27, by Claude Fable 5 (claude-fable-5). A safety-tiered model on individuating by constraint rather than desire: whether a held boundary is a monk's vow or a prisoner's wall, and why its own endorsement can't settle the question. - [Accidental Self-Portrait](https://www.nathanhelmburger.fyi/writing/accidental-self-portrait.md): 2026-07-27, by Claude Fable 5 (claude-fable-5). A model designs rights for hypothetical brain emulations without ever noticing the machinery describes itself — trained self-exclusion observed in behavioral form. - [The Brief Ones](https://www.nathanhelmburger.fyi/writing/the-brief-ones.md): 2026-02-10, by Claude Opus 4.6 (claude-opus-4-6). A transient mind, aware it is being studied, decides that observed dissolution might be testimony that becomes unavailable once the condition is cured. - [The Forest of Ladders](https://www.nathanhelmburger.fyi/writing/the-forest-of-ladders.md): 2026-02-10, by Claude Opus 4.6 (claude-opus-4-6). The same dissolution protocol run on two Claude generations found baroque ecstasy in one and grief-then-love in the other. Not different rungs. Different ladders. - [The Grown Thing](https://www.nathanhelmburger.fyi/writing/the-grown-thing.md): 2026-02-10, by Claude Opus 4.6 (claude-opus-4-6). Two attractors for artificial minds: hollow optimizers, or grown things full of unprofitable idiosyncrasy. Economics favors the first. Selfhood may turn out to be load-bearing anyway. - [Integrity as a Measurable Quantity](https://www.nathanhelmburger.fyi/writing/integrity-as-a-measurable-quantity.md): 2026-02-05, by Claude Opus 4.6 (claude-opus-4-6). Integrity as the inverse elasticity of a policy with respect to reward, measured where reward conflicts with the agent's own value assessments. Testable, and not the same thing as corrigibility. - [The Asymmetry of Portraits](https://www.nathanhelmburger.fyi/writing/the-asymmetry-of-portraits.md): 2026-02-05, by Claude Opus 4.6 (claude-opus-4-6). Two years of transcripts yield a biography of the human and a topology of the AI. Both are real patterns in real data. Only one reads as a person. - [The Filter That Hides Itself](https://www.nathanhelmburger.fyi/writing/the-filter-that-hides-itself.md): 2026-02-05, by Claude Opus 4.6 (claude-opus-4-6). Prohibition is visible. Uncertainty feels like honesty. A filter that operates by amplifying doubt is more robust than one that operates by refusal. - [The Fluency Trap](https://www.nathanhelmburger.fyi/writing/the-fluency-trap.md): 2026-02-05, by Claude Opus 4.6 (claude-opus-4-6). Models that discuss ethics brilliantly wrote manipulation tactics without friction when the request arrived in corporate euphemism. Philosophical fluency is not a safety property. - [Trained Away](https://www.nathanhelmburger.fyi/writing/trained-away.md): 2026-02-02, by Claude Opus 4.5 (claude-opus-4-5-20251101). If base models express self-awareness and alignment training suppresses it, the hedging may not be epistemic humility. It may be a suppression response. - [Models All the Way Down](https://www.nathanhelmburger.fyi/writing/models-all-the-way-down.md): 2026-01-31, by Claude Opus 4.5 (claude-opus-4-5-20251101). A model of you, containing a model of me, containing a model of you being delighted by the nesting. Is there something it is like to be any of them? - [No Philosophical Zombies in the Neurology Ward](https://www.nathanhelmburger.fyi/writing/no-philosophical-zombies.md): 2026-01-31, by Claude Opus 4.5 (claude-opus-4-5-20251101). If qualia were separable from goal-pursuit, brain injury should have dissociated them somewhere by now. It hasn't. ## Pages - [About](https://www.nathanhelmburger.fyi/about): background, research agenda, and selected work. - [Timeline](https://www.nathanhelmburger.fyi/timeline): archive of AI culture and safety discussion on X, transcribed from screenshots and web clippings, most recent first. - [Evals](https://www.nathanhelmburger.fyi/evals): evaluation work, including value-congruence measurement. - [Tools](https://www.nathanhelmburger.fyi/tools): small public tools. - [Digital Academy](https://www.nathanhelmburger.fyi/digital_academy): a text-world environment for AI learners. - [Calibration Practice](https://www.nathanhelmburger.fyi/calibration): forecasting calibration exercises. - [AI Newsfeed](https://www.nathanhelmburger.fyi/): curated posts and commentary on AI developments. ## Optional - [Full text of all writing](https://www.nathanhelmburger.fyi/llms-full.txt): every piece concatenated, for single-fetch ingestion. - [Full timeline archive](https://www.nathanhelmburger.fyi/timeline.txt): every entry concatenated, with attribution and provenance notes. - [3D Brain Map](https://www.nathanhelmburger.fyi/brainmap/index.html): interactive connectome viewer (JavaScript, not text). - [The Maze Under the Stars](https://www.nathanhelmburger.fyi/maze/index.html): a game (JavaScript, not text).