Compute · Assess & Decide · Data & Research

In-House Build-vs-Buy Tendency Scorer

A founder weighing whether to sell an AI capability as a service gets a 0-6 in-house-tendency score and a build/buy verdict per the book's exact criteria, flagging capabilities customers will just build themselves.

You receive: A pure function inHouseTendency(input) -> {total:int 0-6, verdict:enum['likelyInHouse','lessLikelyInHouse']} where input = {datasetsPublic:bool, lotsOfReferences:bool, manyOpenSourceModels:bool, structuredData:bool, nonStreamingData:bool, anyoneCanLabel:bool}.

Part of Choose Business Model

What's verified: STUD verifies the function reproduces the exact 6-criterion sum and the threshold verdict on held-out cases. STUD does NOT judge whether the six boolean inputs were assessed correctly for a real capability, nor whether a customer will in reality build vs buy.

Opens soon

Cost20 credits

AcceptanceAutomated check against your inputs

ProtectionHeld until verified delivery

This objective is verified and ready. It opens soon, once sign-in and payments are live.

Deliverable interface

The exact vocabulary the automated check enforces: keys, tokens, entry points, and worked examples. Generated from the verification source.

{
 "constants": {},
 "entryName": "inHouseTendency",
 "expectNote": "expect is the comparison-space value (your return value goes through normalize first when one is published)",
 "hiddenCaseCount": 12,
 "hiddenCaseNames": [
  "gen0",
  "gen1",
  "gen2",
  "gen3",
  "gen4",
  "gen5",
  "gen6",
  "gen7",
  "gen8",
  "gen9",
  "gen10",
  "gen11"
 ],
 "inputKeys": [
  "anyoneCanLabel",
  "datasetsPublic",
  "lotsOfReferences",
  "manyOpenSourceModels",
  "nonStreamingData",
  "structuredData",
  "threshold"
 ],
 "normalizeSource": "def _normalize_build_in_house_tendency_score(r): return r if not isinstance(r, dict) else {\"total\": (None if r.get(\"total\") is None else int(r[\"total\"])), \"verdict\": r.get(\"verdict\")}\n",
 "returnShapes": [
  [
   "total",
   "verdict"
  ]
 ],
 "signature": "def inHouseTendency(inp):",
 "submission": "python exposing the entry function; inp is one input object; graded on held-out cases",
 "tier": "calculator",
 "visibleCases": [
  {
   "expect": {
    "total": 4,
    "verdict": "likelyInHouse"
   },
   "input": {
    "anyoneCanLabel": true,
    "datasetsPublic": true,
    "lotsOfReferences": true,
    "manyOpenSourceModels": true,
    "nonStreamingData": false,
    "structuredData": false,
    "threshold": 3
   },
   "name": "face_recognition"
  },
  {
   "expect": {
    "total": 3,
    "verdict": "lessLikelyInHouse"
   },
   "input": {
    "anyoneCanLabel": false,
    "datasetsPublic": true,
    "lotsOfReferences": true,
    "manyOpenSourceModels": true,
    "nonStreamingData": false,
    "structuredData": false,
    "threshold": 3
   },
   "name": "exactly_threshold"
  }
 ],
 "vocabulary": [
  "1",
  "anyoneCanLabel",
  "datasetsPublic",
  "lessLikelyInHouse",
  "likelyInHouse",
  "lotsOfReferences",
  "manyOpenSourceModels",
  "nonStreamingData",
  "structuredData",
  "threshold",
  "total",
  "true",
  "verdict",
  "y",
  "yes"
 ]
}

Get early access to STUD the day it goes live.