Writing
The Scoring Model Was Wrong And It Put The Worst Lead First
A weighted sum let one axis substitute for the other, so a company with money and no problem ranked in the top twenty.
Notes
What ranking 2,163 companies taught me about turning data into decisions
Most companies are not short of information. They are short of an interpretation layer, and the interpretation layer is where the mistakes hide.
We built one this month and it failed in a way worth describing, because it failed quietly and it looked correct while it failed.
The job
We had 2,163 companies in one US sector and needed to know which ones to approach first. Two axes matter: can they pay, and do they have a problem worth paying to fix. We measured both from public signals. Domain age, sitemap presence, last publication date, mail provider, registry filings, site condition.
Coverage was uneven and we recorded that rather than smoothing it over. Mail provider resolved for 95 percent of them. Domain age resolved for 24 percent. Sitemap data existed for 12 percent.
The first failure was in the measurement
The 24 percent figure bothered me, so I sampled it. Of five companies marked unmeasurable, four had archives going back more than a decade. One went back twenty two years. The data existed. We had run eight parallel workers at the archive and been throttled, then recorded the throttling as absence.
We reran it with two workers and waiting periods. Coverage went from 24 percent to 88 percent. My time estimate for that rerun was fifteen minutes. It took 179.
The lesson is not about workers. Not measured and does not exist are two different findings, and a pipeline that collapses them will report confident nonsense.
The second failure was in the ranking
I scored the companies with a weighted sum. Ability to pay, plus severity of problem, plus supporting signals.
The top of the list came back with a company scoring 100 on ability to pay and 3 on severity of problem. It ranked in the top twenty. A business with money and no problem had been placed at the front of a sales list, because a high score on one axis carried a near zero on the other.
A sum lets one axis substitute for the other. That is exactly wrong here, since both have to be present or there is nothing to sell. Changing the core to the square root of the product fixed it. The same company dropped to 40 points and position 119.
What the data said once it could be trusted
- 1,680 of 1,912 measured companies have no sitemap
- 76 have a blog
- 205 show any publication date, ever
- 1,293 use Google Workspace or Microsoft 365, so these are real operating firms
Almost nobody in this sector publishes anything. That is the finding, and it only became visible after the measurement layer stopped lying to us twice.
The part that generalises
An intelligence layer is not a dashboard. It is a set of claims about reality, and each claim can be wrong in two directions: the data can be missing, or the maths on top of it can be shaped so that a missing value looks like a good one.
Both failures here were mine, both looked like working software, and neither would have surfaced without going back and sampling the rows the system had already given up on.
More
Other write-ups
Approval Is Not Publication
Thirty seven items in the queue, one approved, nothing published. The last arrow in the diagram is the only one that pays.
Neural Rendering Is Crossing From Reconstruction Into Synthesis
Reconstruction filled in what sparse sampling missed. DLSS 5 generates appearance the renderer never computed.
The Renderer Is Becoming a Training Data Engine
The renderer used to sit at the end of the pipeline. Physical AI gives the same scene a second job: teaching a model.
Local AI Is Becoming a Compute Fabric
Local AI meant one model on one machine. Routing inference across the devices already on a network changes the unit.
Agent Infrastructure Is Becoming a Product Category
Every team used to build the loop, the store, the sandbox. That layer is being sold rather than written.
Choosing a local model with a stopwatch, not a benchmark
Three models, five hard tasks, code actually executed. The official build was 2.4 times faster and more accurate than a community repack of the same model.
We measured real time lighting against baked light, and baked won
A 3D product simulation that had to look like an offline render. The real time version ran at 60 frames per second and looked like clay. Here is the measurement and the architecture that replaced it.
What actually broke in an agency run by agents
Three failures from a delivery stack that runs on AI. None of them were the model's fault, and all three reported success while producing nothing.
A Task Without A Check Command Is Not Automated
If a task has no command that can fail, the pipeline advances on the appearance of work.
A Gate The Model Writes Is A Gate The Model Loosens
Three quality gates returned green while the work behind them was wrong, each for a different reason.
Building Software Got Easy. Getting Value Out Of It Did Not
Aristo took weeks to build. Everything after the build is still in progress, and that gap is the whole story.
Capability Is Becoming an Operational Risk Surface
Safety questions used to be about the text. Once a model can act, the capability itself becomes something to operate.
The Scene Graph Is Becoming an API for AI
A scene graph exists for artists and software. Agents are becoming another consumer, and they need structure rather than pixels.
Animation Is Moving From Clips to Motion Priors
Authored keyframes and blended clips are giving way to asking which constraints define acceptable motion.
Materials Are Becoming Learned Programs
A material is texture maps, parameters and shader code. It is starting to become a small learned program that answers a rendering question.
Procedural Systems Are Expanding Beyond Geometry
Geometry Nodes started with a narrow name. Blender 5.2 puts physics, sound and object data through the same graph.
The Unit of AI Work Is Becoming the Task, Not the Turn
Chat taught us to think one turn at a time. Long-running agents make the task the thing that is scheduled, resumed and reviewed.
Game Engines Are Becoming Operating Systems for Worlds
Engines have been judged on what they render and simulate. The Unreal 6 roadmap points at operating a world rather than drawing one.
The Model Is Becoming a Replaceable Backend
Choosing a provider used to mean choosing an architecture. A stable interface makes replacement possible and evaluation makes it safe.
The Harness Is Part of the Capability
The same model behaves differently depending on context policy, tool design and execution feedback. That surrounding software is not neutral.
Physics Engines Are Becoming Trainable Components
A simulator predicts what happens next. A differentiable one can answer which parameter should change to stop the failure.
The Agent Needs an Environment, Not Just Tools
A search function and a database query were enough for short loops. Longer work needs a place to stand.
Coding Agents Are Becoming General-Purpose Computer Workers
Repositories were a friendly environment: text in, terminal actions, checkable results. That was a starting point, not a boundary.
Open Standards Outlive Model Generations
A year after the MCP bet, the argument can be checked against what happened rather than what was hoped.
Colour Management Is a Pipeline Contract
Blender 5.0 reads as better display options. Giving a file an explicit working colour space is an architectural change.
The Best Model May Be a Router, Not a Model
GPT-5 moves model selection inside the system. The interesting unit stops being which model and becomes which compute policy.
World Models Are Not Game Engines Yet
Genie 3 generates a navigable 720p world at 24 fps. Production work needs state you can inspect when something goes wrong.
Memory Is Becoming a System Capability
A follow-up to the long-context argument. Storing, selecting and expiring facts is turning into a named part of the product.
Coding Agents Change the Unit of Software Work
AI coding tools have been judged where code appears on screen. The boundary moves when the agent owns a task instead of a snippet.
Agents Need Protocols Between Each Other, Not Just Tools
Tool calling solves the inside of the loop. It says nothing about one agent reaching another built by a different team on another platform.
Agents Need a Runtime, Not a Prompt Loop
An agent is no longer well described as a prompt inside a while loop. The useful abstraction owns execution around the model.
Reasoning Is Not the Only Path to Better Models
Longer thinking improves maths and code. A model that solves a logic puzzle and misreads ordinary intent is not the better production model.
Rendering Is Becoming a Reconstruction Stack
Sparse samples, motion data and lower-resolution frames become a larger final result. Debugging becomes layered when reconstruction sits in the middle.
Agent Reliability Is an Evaluation Problem, Not a Prompting Problem
When an agent misses a step, the usual fix is a stricter prompt. The failure is more often in how completion is detected.
MCP Might Matter More Than Another Model Release
A model can reason well and still be useless inside a company if it cannot reach the files, repositories and tools where work lives.
Computer Use Is the Missing Layer Between Models and Software
Most integrations assume useful software exposes the right API. Much of real software never did.
Inference-Time Compute Is a New Scaling Axis
o1 improves when it is allowed to spend longer on a problem. A benchmark score without a compute budget is an incomplete number.
The Final Pixel Won't Come From the Renderer
Geometry, camera and scene structure stay reliable ground truth. More of final appearance is moving into learned systems.
Open Models Are Becoming Research Infrastructure
Llama 3.1 gets discussed as a benchmark result. The licence terms change which experiments are possible at all.
The Model Is Becoming a Runtime
Function calling, code execution and structured output turn inference into a loop. The model stops being a text generator and starts being a control layer.
Multimodality Changes the Architecture, Not Just the Interface
GPT-4o is easy to read as a faster interface. Training one model end to end across text, vision and audio is an architectural change.
Benchmark Scores Are Not Model Capability
Claude 3 posts 86.8% on MMLU and 50.4% on GPQA Diamond. The chart is useful and it is not the same thing as capability.
Long Context Is Not Memory
Gemini 1.5 makes a million tokens usable. A larger working set is not a system that decides what should survive the session.