Doom or Bloom: Mapping AI Worldviews and Community Reactions

Quick Take

Doom or Bloom is a publicly‑available questionnaire that maps respondents onto a two‑dimensional chart ranging from civilizational doom to incremental bloom of AI, and it visualises where well‑known public figures fall on that spectrum. Hacker News users praised the concept, questioned the methodology, and suggested usability improvements.

What the Tool Does

  • Purpose: The site asks a short series of questions about expectations of AI progress, societal impact, and human agency, then places the answer on a chart with a doom axis (existential risk) and a bloom axis (incremental benefit).
  • Public‑Figure Simulations: It shows simulated positions for 45 personalities—from Eliezer Yudkowsky and Geoffrey Hinton to Elon Musk and Barack Obama—allowing users to compare their own view with those of prominent thinkers.
  • Technical Stack: The front‑end is a Next.js app hosted at doom-or-bloom.com; the source code is open‑source on GitHub (github.com/transitive-bullshit/doom-or-bloom).
  • Outcome: After completing the questionnaire, users see a point on the chart and can explore the “simulated” answers that generated the public‑figure placements.

Core Design Choices

  • Binary Framing: The tool reduces the complex AI‑future debate to a single doom‑vs‑bloom continuum, which forces respondents to choose a side.
  • Simulated Personas: The positions for public figures are generated algorithmically rather than quoted directly, a point raised by several commenters.
  • No Authentication Required: The site offers a “log‑in with X” button, but users can proceed without an account, prompting questions about data collection.

Community Reaction on Hacker News

Positive Reception

  • Thought‑Provoking Prompt: Users like Suhinnall27 noted that the questionnaire asks a crucial self‑reflection question that many people should consider.
  • Insight into Diversity: Commenters appreciated seeing how varied the simulated results are, especially when personal results diverged from expectations.

Critiques of Methodology

  • Questionable Persona Generation

    "What are the 'simulated answers' for the public figures based on? They don't seem to be direct quotes, so I assume they're AI‑generated in their voice…" – chroma_zone

    The lack of transparency about how the personas are constructed may misrepresent real opinions.

  • Over‑Simplification

    "Unask the question. Perhaps step outside this foolish dichotomy…" – grey-area

    Some argue the doom‑vs‑bloom framing ignores nuanced outcomes such as the "enshittification" of digital spaces.

  • Placement of Skeptics

    "Gary Marcus in the center? He is a strong AI skeptic, I think he should be lower, away from the 'Civilizational change' direction." – Marha01

    The chart’s positioning of certain figures sparked debate about the underlying scoring algorithm.

Usability Concerns

  • Login Prompt – thevinter asked why a login is needed, suggesting the flow could be streamlined.
  • Navigation Friction – in-silico wanted a way to view public‑figure entries without opening new pages.
  • Score Interpretation – ddxv and purpleflashing reported difficulty interpreting probability scores and felt the questionnaire forced them into extreme categories.
  • Missing Dimensions – ModernMech suggested adding axes for net worth or social‑media reach to capture power dynamics.

Notable User Experiences

  • Extreme Doom Self‑Assessment – cyclopeanutopia reported a 100 % doom probability, indicating the tool can capture strong pessimism.
  • Incremental Bloom Alignment – weinzierl found the incremental bloom quadrant empty and wondered why the tool offered few nuanced options.
  • Unexpected Alignments – kypro landed near Geoffrey Hinton despite personally feeling far more doom‑oriented, highlighting potential mismatches between self‑perception and algorithmic scoring.

What the Discussion Reveals About AI Worldview Mapping

  1. Complexity vs. Simplicity – While a single‑axis model is easy to digest, many participants argue that AI’s societal impact cannot be captured by a binary spectrum.
  2. Transparency Is Crucial – When tools simulate public figures, clear disclosure of data sources and generation methods is essential to avoid misrepresentation.
  3. User Agency Matters – The requirement (or suggestion) to log in, and the inability to explore data without extra clicks, can deter participation and affect the perceived legitimacy of the results.
  4. Interpretation of Scores – Without contextual guidance, probability scores (e.g., “P(doom) ≈ 1 %”) can be confusing, leading users to question the validity of the assessment.
  5. Community Validation – The range of reactions—from enthusiastic endorsement to sharp criticism—demonstrates that any attempt to map AI worldviews will inevitably become a focal point for broader debates about risk, benefit, and power.

Takeaway for Builders of Opinion‑Mapping Tools

  • Provide Source Transparency: Clearly label whether persona positions are derived from public statements, surveys, or AI‑generated approximations.
  • Offer Multi‑Dimensional Views: Consider adding axes for economic concentration, governance, or ethical considerations to capture richer narratives.
  • Simplify Navigation: Enable inline expansion of persona details and optional anonymous participation without login prompts.
  • Explain Scoring: Include a brief legend that interprets probability values and the meaning of each quadrant.
  • Iterate Based on Feedback: Use community comments—like those on Hacker News—to refine question wording, answer options, and visual design.

Doom or Bloom succeeds in sparking conversation about AI futures, but the Hacker News discussion underscores the need for methodological clarity, richer dimensionality, and user‑friendly design when visualising complex societal expectations.

Sources

Related