Map is not the territory

07 Apr 2026

Shreyas Prakash headshot

Shreyas Prakash

One of my favorite ideas from Shane Parrish’s Mental Models book which I read recently is this:

The map is never be the territory.

harold fisk mississippi river

Harold Fisk’s Meander Maps of the Mississippi River (1944)

In other words, the “description” of the thing, is not the thing in itself.

It’s so applicable, that it doesn’t exist to me as this abstract idea anymore.

So, the Argentinian writer Luis Borges, in his brilliant allegorical style, summarized the mental model nicely in a one-paragraph short story, On Exactitude in Science:

… In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography.

—Suarez Miranda, Viajes de varones prudentes, Libro IV,Cap. XLV, Lerida, 1658

This is a fictional tale of a map that is so precise of its’ “territory” that it is “useless”, after all, one could just follow the territory instead of using the map. Why use a map then?

It’s a brilliant thought experiment which highlights the fact that maps are lossy, and you lose a lot of details in the process, and this is not a bug, but a feature of it being a map.

Even the best maps are reductions, and they’re never perfect. They differ in their exactitude..

And this is my commentary on the notion that AI is taking away all the “human jobs”, it would at the end of the day, AI can only be an approximation, and cannot exactly replace the human.

It could probably be more than what a human can do, or sometimes, less than what humans are meant to do. Love, care and kindness etc, are not something they can exact, and can only mimic.

My thinking now is that, as long as these approximations exist, the human jobs are not going away any time soon. Talking about approximations, the London underground map is a terrific example, in that, the approximation served more useful than the actual map, in this essay by Ellla Watkins-Dulaney for Asimov Press, she writes:

Consider the map of the London Underground. Until 1933, the map plotted stations at geographically accurate locations in London. But this made central London, where most stations clustered, an unreadable tangle, while the outer suburbs, devoid of relevant data, took up most of the space. The draughtsman Harry Beck solved this inefficiency by abandoning geographic accuracy and instead redrawing the network as a circuit diagram of colored lines and evenly spaced stations.

A pocket map of London’s Underground system in 1908.

Harry Beck’s 1933 map of the London Underground.

I do have a similar argument to suggest folks to [read raw transcripts.](https://shreyasprakash.com/read-raw-transcripts/

This is quite a contrary opinion still, and we’re more used to referring to summaries. Maybe that’s a quick nugget and has faster absorption, but we do mistake comprehension for understanding sometimes. They are not the same. If I do have the luxury of time, I would personally prefer reading through the raw transcripts to get more higher dimensional granularity to what’s going on..

If we extend the analogy, the raw transcript is almost like the “territory” we’re speaking about. And the insights generated, are more or less like the “map”. And the map can never be equivalent to the territory.

Apart from this quirk to read raw transcripts, I also have a similar asymmetric thumb rule when it comes to reading long form blog posts. Some of them are really high alpha, high signal, so when a writer deeply resonates with me, I go the last mile and read the whole essay end to end.

I want to read it in its lossless format, without it being stripped of any details.. as “map is not the territory”, I do prefer waltzing through the territory in some occasions, and not use the map instead..

Subscribe to get future posts via email (or grab the RSS feed). 2-3 ideas every month across design and tech

Read more

  1. Home is where the feeling of home residesreflections
  2. Notes on doing most good, doing some good, and living goodphilanthropy
  3. Notes on Londontravel
  4. 18 key ideas from reading Henrik Karlsson this monthwriting
  5. Stop arguing, and start drawing circles togethermathematics
  6. The whole world is just the snake eating its own tailmental-models
  7. Life lessons and hot takes from my 30slifestyle
  8. Building a skill for coherent science illustrationsscience
  9. My agentic engineering workflow (step by step)agentic-coding
  10. Hammock driven developmentagentic-coding
  11. Peculiar ways number three fits into our funny little brainsmental-models
  12. AI sandwich as a defacto principle for anything agentic engineering relatedagentic-coding
  13. Authority in the guise of evidencecritical-rationalism
  14. Map is not the territoryphilosophy
  15. Self hypnosis as a manifestation ritualmeditation
  16. Hegelian dialectic for structured reasoning with AI agentsphilosophy
  17. How I prepare for tough negotiations nowadaysnegotiation
  18. When should we steelthread somethingproduct-development
  19. Breadboarding, shaping, slicing, and steelthreading solutions with AI agentsproduct
  20. Healthy conflict in teams have a tipping pointteam-building
  21. How I deslopify AI writingwriting
  22. How I started building softwares with AI agents being non technicalagentic-coding
  23. Read raw transcriptswriting
  24. Legible and illegible tasks in organisationsproduct
  25. L2 Fat marker sketchesdesign
  26. Writing as moats for humanswriting
  27. Beauty of second degree probesdecision-making
  28. Boundary objects as the new prototypesprototyping
  29. One way door decisionsproduct
  30. Finished softwares should existproduct
  31. How I periodically rank my rough draftsobsidian
  32. Flipping questions on its headinterviewing
  33. Vibe writing maximswriting
  34. How I blog with Obsidian, Cloudflare, AstroJS, Githubwriting
  35. How I build greenfield apps with AI-assisted codingagentic-coding
  36. We have been scammed by the Gaussian distribution clubmathematics
  37. Classify incentive problems into stag hunts, and prisoners dilemmasgame-theory
  38. I was wrong about optimal stoppingmathematics
  39. Thinking like a shipmental-models
  40. Hyperpersonalised N=1 learningeducation
  41. New mediums for humans to complement superintelligenceagentic-coding
  42. Maxims for AI assisted codingagentic-coding
  43. Virtual bookshelvesaesthetics
  44. It's computational and AI everythingagentic-coding
  45. Public gardens, secret routesdigital-garden
  46. Git way of learning to codeagentic-coding
  47. Style Transfer in AI writingagentic-coding
  48. Understanding codebases without using codeagentic-coding
  49. Vibe coding with Cursoragentic-coding
  50. Virtuoso Guide for Personal Memory Systemsmemory
  51. Writing in Future Pastwriting
  52. Publish Originally, Syndicate Elsewhereblogging
  53. Poetic License of Designdesign
  54. Idea in the shower, testing before breakfastsoftware
  55. Technology and regulation have a dance of ice and firetechnology
  56. How I ship "stuff"software
  57. Writing is thinkingwriting
  58. Song of Shapes, Words and Pathscreativity
  59. How do we absorb ideas better?knowledge
  60. Read writers who operatewriting
  61. Brew your ideas lazilyideas
  62. Trees, Branches, Twigs and Leaves — Mental Models for Writingwriting
  63. Compound Interest of Private Noteswriting
  64. Conceptual Compression for LLMsagentic-coding
  65. Meta-analysis for contradictory research findingsdigital-health
  66. Proof of workproduct
  67. Gauging previous work of new joinees to the teamleadership
  68. Task management for product managersproduct
  69. Beauty of Zettelswriting
  70. Stitching React and Rails togetheragentic-coding
  71. Exploring "smart connections" for note takingwriting
  72. Deploying Home Cooked Apps with Railssoftware
  73. Repetitive Copypromptingwriting
  74. Questions to ask every decadejournalling
  75. Balancing work, time and focusproductivity
  76. Hyperlinks are like cashew nutswriting
  77. Brand treatments, Design Systems, Vibesdesign
  78. How to spot human writing on the internetwriting
  79. Can a thought be an algorithm?product
  80. Opportunity Harvestingcareers
  81. Everything is a prioritisation problemproduct
  82. How I do product roastsproduct
  83. The Modern Startup Stacksoftware
  84. In-person vision transmissionproduct
  85. How might we help children invent for social good?social-design
  86. The meeting before the meetingmeetings
  87. Design that's so bad it's actually gooddesign
  88. Lessons learnt interview prepping for product rolesinterviewing
  89. Obsessing over personal websitessoftware
  90. English is the hot new programming languagesoftware
  91. Better way to think about conflictsrisk-management
  92. The role of taste in building productsdesign
  93. Dear enterprises, we're tired of your subscriptionssoftware
  94. Products need not be user centereddesign
  95. World's most ancient public health problemsoftware
  96. Pluginisation of Modern Softwaredesign
  97. Let's make every work 'strategic'consulting
  98. Making Nielsen's heuristics more digestibledesign
  99. Startups are a fertile ground for risk takingentrepreneurship
  100. Insights are not just a salad of factsdesign
  101. Minimum Lovable Productproduct
  102. Methods are lifejackets not straight jacketsmethodology
  103. How to arrive at on-brand colours?design
  104. Minto principle for writing memoswriting
  105. Importance of Whytask-management
  106. Quality Ideas Trump Executionsoftware
  107. Why I prefer indie softwareslifestyle
  108. Use code only if no code failscode
  109. Self Marketing
  110. Personal Observation Techniquesdesign
  111. Design is a confusing worddesign
  112. A Primer to Service Design Blueprintsdesign
  113. Rapid Journey Prototypingdesign
  114. Visualise detailed file structures on CLIcli
  115. Do's and Don'ts of User Researchdesign
  116. Design Manifestodesign
  117. Complex project management for productproducts
  118. How might we enable patients and caregivers to overcome preventable health conditions?digital-health
  119. Pedagogy of the Uncharted — What for, and Where to?education
  120. Future of Equity with Ludovick Petersinterviewing
  121. Future of Ageing with Mehdi Yacoubiinterviewing
  122. Future of Tacit knowledge with Celeste Volpiinterviewing
  123. Future of Mental Health with Kavya Raointerviewing
  124. Future of unschooling with Che Vanniinterviewing
  125. Future of Rural Innovation with Thabiso Blak Mashabainterviewing
  126. Future of work with Laetitia Vitaudinterviewing
  127. How might we prevent acquired infections in hospitals?digital-health
  128. The why to endure any howentrepreneurship
  129. Design education amidst social tribulationsdesign
  130. How might we assist deafblind runners to navigate?social-design