Manas Bihani
About

the questions

  1. What is a moat in an AI world?
  2. Why do AI products converge?
  3. What becomes scarce when intelligence becomes cheap?
  4. Does distribution matter more than technology?
  5. Why might human-made things become more valuable?
  6. What happens to expertise when everyone has the same models?
  7. Which parts of an AI startup are actually defensible?
  8. Where does value move when intelligence becomes commoditized?

everything on the desk

  1. The periodic table of the AI stackVisualization
  2. What is a moat when the model isn't yours?Note
  3. The problem-selection premiumNote
  4. Same model, different wiringNote
  5. Selection is the new bottleneckNote
  6. Get friendly with the AI raceEssay
  7. The convergence taxNote
  8. The luxury of realityNote
  9. The non-technical technical advantageNote
  10. The verification economyNote
  11. The bets against the wallVisualization
  12. You can't buy your way outVisualization
  13. How a chatbot writes one wordVisualization
  14. The grid is the last wallVisualization
  15. Who got paidVisualization
  16. Why this paper mattersExplainer
  17. Transformer: Why did transformers replace RNNs?Vaswani et al., NeurIPS 2017
  18. KV cache: Why does a long conversation get slower and cost more than a short one?Shazeer, 2019
  19. Mixture of experts: Why do some AI models have experts?Fedus, Zoph and Shazeer, 2021
  20. FlashAttention: Why is attention slow when the GPU is barely doing any arithmetic?Dao et al., NeurIPS 2022
  21. Mamba: Why does a model reread the whole conversation instead of just remembering it?Gu & Dao, 2023
  22. PagedAttention: Why does a GPU with free memory still refuse new requests?Kwon et al., SOSP 2023
  23. DeepSeek: How did DeepSeek train a frontier model so cheaply?DeepSeek-AI, 2024
  24. Jamba: Why does Jamba matter?Lieber et al., AI21 Labs, 2024
  25. BitNet: Why does BitNet matter?Ma et al., Microsoft Research, 2025
  26. DeepSeek-R1: Can a small AI model learn to reason like a huge one?DeepSeek-AI, 2025
  27. Kimi K2: Why does Kimi K2 matter?Kimi Team, Moonshot AI, 2025
  28. Sliding-window attention: How do models handle huge context windows without the memory bill exploding?Gemma Team, Google DeepMind, 2025
  29. How electricity becomes intelligenceVisualization
  30. This desk, as a datasetDataset
  31. The first version of this roomNote
  32. The aura dividendNote
  33. Distribution is rented attentionNote
  34. The Convergence TestNote
  35. A shelf for thinking about cheap intelligenceCollection
  36. Anatomy of an AI startupNote
  37. Six shocks to expertiseNote
  38. Nineteen Public KeysEssay
  39. The value migration machineModel
  40. AAA-Rated GPUsEssay
  41. Moats, before and afterVisualization
  42. The rhinoceros problemNote
  43. What becomes scarce when intelligence becomes cheap?Essay
  44. AI Has Passed Every Exam. It Has Never Had an Idea.Essay
  45. What Becomes Scarce After Intelligence?Essay
  46. India’s Carbon Markets : A New Test for Global Climate PolicyEssay
  47. Google Wants AI to Become BoringEssay
  48. The Wall That Wasn’t YoursEssay
  49. The Rate-Limiting StepEssay
  50. The Speed of Being WrongEssay
  51. Uber Burned a Year of AI Budget in Four Months. A Rat Catcher in 1902 Knew WhyEssay
  52. Finding a Flat in India Is Broken. We Have the Technology to Fix It. Nobody With Power Wants To.Essay
  53. Why We Can Never Have Good Social MediaEssay
  54. Gen Z Is Going OfflineEssay

rooms

  1. Home
  2. Writing
  3. Projects
  4. Reading & Watching
  5. All the questions
  6. Everything, as a contact sheet
  7. About

Essay · 4 Aug 2026

What becomes scarce when intelligence becomes cheap?

Every time something becomes abundant, the value moves next door. A search for the next door.

the bottleneck always moves next door

trying to answer →What becomes scarce when intelligence becomes cheap?Where does value move when intelligence becomes commoditized?

Dürer's engraving Melencolia I: a winged figure sits brooding among unused tools, a polyhedron, a sphere, an hourglass and a magic square.judgement, sitting idlethe problem nobody solvedtime: still scarcea perfect tool, unused
Albrecht Dürer, Melencolia I, 1514. Genius surrounded by instruments, unable to use any of them. Public domain, via Wikimedia Commons

Economists have an old, unglamorous rule that explains more about technology than most futurism does. When something becomes abundant, its price falls toward the cost of making it, and the value moves to whatever is still scarce next to it. Printing made copies cheap and made authors and editors valuable. Search made facts cheap and made attention expensive. The interesting question about cheap intelligence is not what it replaces. It is what it sits next to.

The rule of the next door

If the marginal cost of a competent paragraph, a working function or a decent analysis goes to nearly zero, three things happen at once. Output explodes. The average quality of that output converges. And everything that can’t be generated, because it is physical, or accountable, or scarce by definition, gets relatively more valuable.Joel Spolsky’s version, from 2002, is the corporate strategy: commoditize your complements. Make the thing next to your product cheap and demand for your product goes up.

The trick is to be precise about which is which. “Human creativity” is not an answer; it is a mood. The useful exercise is to go down the list of things intelligence is currently bundled with and ask, one at a time, whether each one gets cheaper along with it.

A ledger of scarcities

Becomes abundant Becomes scarce Why
Answers Good questions A model answers what it’s asked. Choosing what to ask is still the bottleneck.
Drafts Judgement When drafts are free, the valuable act is choosing the right one, and saying no to the rest.
Content Attention The supply of words explodes. The hours in a reader’s day do not.
Average expertise Taste When everyone has the median expert, the edge is the non-median opinion.
Software Distribution Anyone can build it. Few can put it in front of people.
Claims Provenance and trust When anything can be generated, who stands behind a thing matters more.
Cognition Atoms and energy Intelligence runs on chips, power and physical access, which don’t get cheap at the same rate.

The question is not which jobs disappear, but which bottlenecks appear.

Accountability is the quiet one

A model can draft a contract, a diagnosis or an audit. It can’t be sued, struck off or embarrassed. Much of what professionals are paid for isn’t the thinking itself but the willingness to put a name to it. I suspect accountability, meaning a person who signs, becomes one of the most durable scarcities. It is scarce by construction, because it only works if it’s attached to someone who can lose something.Baumol’s cost disease points the same way: where productivity can’t rise (a string quartet still needs four players), relative prices go up. Being present and answerable may be the new quartet.

Where I’m unsure

Taste is the answer everybody gives, which makes me suspicious of it. Taste is scarce today partly because making things is expensive, so you need taste to decide what deserves the effort. If making things is free, taste might matter more, as the only filter left. Or it might matter less, because you can generate a thousand variants and let the market choose.

My provisional bet is on the filter. But this is exactly the kind of claim this room exists to test, not to repeat.


Every abundance creates a bottleneck. Find the bottleneck and you have found where the value went.