Any Human Ever – One life, drawn at random from all who have ever lived

Original Article ↗ Hacker News Discussion ↗

An interactive site that “rolls” a random human life from across history is prompting both awe and skepticism. Many find it a striking way to grasp how dangerous, short and geographically constrained most lives have been compared to the modern West, and to reflect on luck, progress and Rawls-style “veil of ignorance” ethics. Others argue the project leans too heavily on large language models, with dubious statistics, an incorrect sampling of years, and hallucinated historical details that risk turning it into convincing but misleading AI slop rather than a reliable educational tool.

Overall reaction & emotional impact

  • Many find the site “fun” and moving, likening it to a modern Oregon Trail or Dwarf Fortress history generator.
  • It prompts reflection on cosmic luck and how fortunate it feels to live in the modern era, especially for parents imagining historical child mortality.
  • Several users stop after drawing lives where “their” children die young; others say it concretely shows how harsh most human lives were.

Historical perspective & life expectancy

  • Repeated shock at infant/child mortality and low life expectancy at birth; discussion clarifies that survivors often lived well beyond the average.
  • Users note that pre‑modern adults reaching 18 could still see 50s–70s; the site sometimes generates very old ages (70–90+) even for ancient periods.
  • Some question framing around wages, money, and modern-style “marriage” in eras or societies where those concepts didn’t align with current norms.

Randomness, sampling, and statistical issues

  • Many suspect the year sampling is wrong: pre‑1500 (especially ancient Korea) seems overrepresented versus known population curves.
  • Debate over whether the tool is drawing years uniformly or by births; several claim it contradicts its own explanation that births are “more likely near the present.”
  • Users report inconsistent probabilities (e.g., marriage rates vs. high under‑15 mortality) and modern-country stats (infant mortality, fertility, population) that are clearly off.

Accuracy, sourcing, and “AI slop” debate

  • Strong split: some see a clever, educational use of AI; others call it misleading “vibe-coded” or “AI slop” and worry about polluting the information commons.
  • Critics highlight:
    • Misapplied sources or “attributed” citations where the referenced work doesn’t match the claimed statistic.
    • Anachronistic details (e.g., Ottoman Turkish for rural peasants, implausible diets, clothing, travel, or divorces).
    • Logical errors (people marrying after death, being “orphaned at 34,” misaligned plagues and wars).
  • Supporters argue it’s a thought-provoking toy based on rough probabilistic history, not a reference work; detractors respond that the interface and citations imply factual rigor.

Creator’s clarification

  • The creator says it took several weeks, used both manual work and AI, and aims for “mostly accurate” stories from patchy academic data.
  • They note difficulties with granular stats, conditional probabilities, and filling gaps; claim to test stories with AI “history teacher” agents and to revise based on feedback.
  • Acknowledge AI images are especially inaccurate, but many users like having faces.

Use cases, comparisons, and suggestions

  • People suggest it as a prompt for narrative RPGs or journaling games, and compare it to older “Real Lives” simulations seen as more rigorous.
  • Some want options to:
    • Separately see life expectancy excluding childhood deaths.
    • Draw lives truly by global birth distribution vs. by era.
    • Input their own data and get probabilistic life-path stats.
  • Others call for clearer disclaimers about uncertainty and AI-generated content.

Meta: AI, economics, and web culture

  • Debate over whether AI meaningfully changes the “economics” of side projects: easier to build rich, unmonetized experiments vs. risk of flooding the web with convincing nonsense.
  • Broader arguments arise about AI’s future impact on work, education, and “post‑scarcity,” with pushback that political and distributional issues matter more than technology alone.