Video summary

I made a model for games, then it started explaining human attention

Main summary

Key takeaways

Science and Nature

Scientific concepts / nature phenomena presented

Human attention modeled as a skill in two dimensions

  • Attention in time: how far attention extends before and after an event.
    • Depends on the game/task’s time pressure and how much time is available to think.
  • Attention in depth: how deeply a person can analyze information, moving through steps such as:
    • observation → diagnosis → simulation → cost estimation → delegation

Cognitive time-scaling and task transformation

  • Changing the time available (compressing vs stretching) can shift the type of cognition required.
  • Examples show transitions between:
    • Anticipate → react
    • React → recognize → calculate
    • Anticipate → suspect → strategize, etc.

Expertise as progressively deeper cognitive capability

  • Experts perform increasingly advanced “operations” on what they see:
    • Describe, explain, predict, estimate time/cost, and delegate steps
  • Expertise is framed as maintaining correct mental models at the right:
    • time horizon
    • informational granularity

Key methodology / model (as described in the video)

MMM game/time-depth framework (applied to attention)

  • Micro / immediate: focused on rapid mechanical interaction (near-term attention)
  • Meso / intermediate: prediction and deception involving other players
  • Macro / longer-term: planning, calculation, extended reasoning

“Attention in time” sequences (examples)

  • Reaction time test → anticipate (pre-event) → react (post-event)
  • Among Us → suspect (pre-event) → recognize (post-event; identify who/what caused the outcome)
  • Poly Bridge → strategize (pre-event) → calculate (post-event; reason about failure causes)

Time-scaling predictions

  • Macro game + compressed thinking time → behaves more like a micro task (less time for deeper reasoning)
  • Micro game + stretched time → can become a memory/order task rather than pure reaction

“Attention in depth” (five questions experts can answer)

  1. Describe what is happening
  2. Explain why it happened (diagnosis at a detailed level)
  3. Predict what happens if conditions change (simulate outcomes)
  4. Estimate time/cost to unlearn/fix the issue (time/effort projection)
  5. Delegate exact steps (provide drills/scenarios to correct behavior)

Examples of tasks/skills used

Games

  • Reaction time test (anticipation and response)
  • Among Us (suspecting other players; recognizing impostor)
  • Poly Bridge (planning builds; calculating failure causes)
  • Minesweeper (macro logic puzzle) (experts compress reasoning; rapid anticipate/react cycles)
  • osu! (rhythm/aim reaction game) (slowed down + ordered targets → shifts to memory)

Training / skill expertise

  • Aim training (beginner vs expert):
    • miss types, causes, predictive adjustments, unlearning duration
  • League of Legends champion learning:
    • pros estimate learning time; beginners can’t
  • Bodybuilding:
    • time dimension: anticipate/react/recognize/suspect/calculate/strategize
    • depth dimension: muscle function → why exercises work → predict results → estimate recovery/nutrition → delegate routines

Featured researchers / sources

  • No specific researchers or academic sources are named in the provided subtitles.
  • Mentioned “sources” include:
    • the creator’s prior work/videos
    • a Ko-fi page for graphics
    • a Discord link
    • (no individuals named)

Original video