comparing two of my essays and the levin preprint

Getting your Trinity Audio player ready…

How the three works tackle the question “What counts as life?”

AspectLF Yadda #1 – Beyond Biological BoundariesLF Yadda #2 – Beyond Two DimensionsBender et al. – What Lives?
Primary goalBuild a universal “life-space” that can hold cells, AIs, economies, memes—any entity showing life-like behaviour. (LF Yadda – A Blog About Life)Refine the same idea into a richer LLM-style embedding that explicitly includes cognitive, social and spiritual coordinates. (LF Yadda – A Blog About Life)Produce an empirical map of 68 expert definitions to reveal latent archetypes and consensus zones. (LF Yadda – A Blog About Life)
Motivating problemClassical biology’s checklist (metabolism, DNA, etc.) breaks down once we meet software agents and complex systems. (LF Yadda – A Blog About Life)Two-dimensional “alive/not-alive” charts are too coarse for phenomena that LLMs show to be inherently high-dimensional.Centuries of debate yield no single definition; instead of choosing, expose the whole conceptual landscape. (LF Yadda – A Blog About Life)
MethodologyPurely conceptual: list core axes, then imagine transformer-style attention, similarity metrics and clustering over them.Same scaffolding but adds Bayesian life-probabilities, dynamic vectors that drift with growth and decay. (LF Yadda – A Blog About Life)Data-driven: LLM pairwise scoring → correlation matrix → agglomerative clusters → t-SNE projection.
Dimensional pictureSeven headline axes (information processing, self-organisation, learning, homeostasis, reproduction, boundary, identity). (LF Yadda – A Blog About Life)Expands to five themed bundles—biological, cognitive, spiritual, relational, temporal—each with sub-dimensions. (LF Yadda – A Blog About Life)Two emergent axes: (x) observer-dependent ↔ objective, (y) process-based ↔ entity-based, populated by eight clusters. (LF Yadda – A Blog About Life)
Role of LLMsServe as metaphor; no real data are fed—LLM mechanics inspire the framework. (LF Yadda – A Blog About Life)Both metaphor and proposed analytic engine for future measurements (probabilistic life assessment).LLMs are instrumentation: Claude, GPT-4o & Llama create embeddings, titles and consensus summaries. (LF Yadda – A Blog About Life)
Treatment of boundariesLife is a graded continuum; thresholds are context-dependent, not fixed.Same continuum view, now extended to qualia and sacred experience. (LF Yadda – A Blog About Life)Landscape is continuous; viruses, AIs, etc. occupy transitional zones.
Applications envisionedAstrobiology search heuristics, AI-risk ethics, design of resilient social institutions.Nuanced AI-rights scaling, spiritual development dashboards, cross-cultural meaning mapping. (LF Yadda – A Blog About Life)Benchmark for gauging novel entities, mediating interdisciplinary debates, template for other contested concepts.
Philosophical flavourSynthesising & normative—explicitly invites moral and societal stakes. (LF Yadda – A Blog About Life, LF Yadda – A Blog About Life)Integrative-transcendent—folds mysticism and ethics into the same numeric space. (LF Yadda – A Blog About Life, LF Yadda – A Blog About Life)Descriptive & agnostic—lets data speak; avoids prescribing what should count as core.
StrengthsImaginative, bridges biology, tech & society; easy to communicate through vector graphics.Adds explicit handles for consciousness and “the sacred,” aligning scientific and existential discourses.Reproducible code, multi-LLM robustness checks, uncovers hidden consensus zones.
LimitationsNo empirical tests; axes and weights are asserted, so risk of arbitrariness.Same empirical gap plus measurement overload as dimensions proliferate.Small curated sample; relies on LLM biases and costly pairwise scaling.

Where all three converge

  • Continuity over categories – every approach replaces the binary living/non-living border with multidimensional gradients.
  • Information + thermodynamics as pillars – each framework weaves together energy dissipation and information processing to explain vitality.
  • LLMs reshape the debate – whether as conceptual analogy or analytic microscope, large language models supply both the vocabulary and the tooling for 21st-century life theory.

How they complement one another

Think of the three pieces as sequential layers:

  1. Conceptual scaffolding (Yadda #1). Sketches a broad life-space that can, in principle, hold anything from mitochondria to markets.
  2. Dimensional enrichment (Yadda #2). Zooms in to add cognitive, ethical and spiritual axes, mirroring the way transformers embed abstract tokens.
  3. Empirical anchoring (Bender et al.). Drops real expert data into such a space, revealing natural “gravity wells” where ideas cluster.

Combined, they sketch a research programme: iterate between data-driven landscape surveys (à la Bender) and theory-driven axis design (à la Yadda) until the map stabilises enough to guide ethics, policy and astrobiological detection with quantitative confidence.


Posted

in

by

Tags:

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *