Inference · industries

The fifty kinds of work we set out to own.

Work where a specialist on owned machines can beat a general model: the language and rules are local, the answer can be checked, the data cannot leave, or the work repeats every day. Each industry is a portfolio, earned only from signed work that a checker scored. Where a frontier model is still better, it stays the fallback, so the system is never worse than the best model it can call.

50industries
10most confident
9institutions as outside checks
0checked runs signed
Bring work from one of these Give Evo a real goal in your industry. Its result is checked and signed, and it counts toward that portfolio.Lend a machine to the work Operators run in pods on grams. A gram with a card can also train the adapters.The models this work runs on Every open model with its signed speed, context and cost per million tokens.

How an industry earns its portfolio

  1. 01

    Goals arrive

    Before launch, from simulated people with real goals in each industry. After launch, from the people doing the work.

  2. 02

    Operators are born for them

    Evo plans the goal and the network creates the specialist operators it needs, each in its own pod on a gram.

  3. 03

    Every result is checked

    A checker scores each result held or failed: a ledger that ties out, a code that matches the code set, a citation that resolves.

  4. 04

    The industry earns its portfolio

    Checked work clusters into domains, good runs become reusable specialists, and adapters train on the outcomes. Standing is earned, never declared.

Work from simulated people is labelled simulated wherever it appears, and counts only where a real checker scored it. A model trained on its own unchecked output gets worse, not better.

ALocal language, local rules, outcomes we can check · 10

A general model has seen little of this work and cannot run where it lives. The answer can be checked.

BRepeated workflows with answers we can verify · 15

The same shape of work every day, so a specialist gets better and cheaper with every signed run.

CData that cannot leave the building · 10

A hosted model cannot see it. A gram runs inside the perimeter and signs what it did.

DWhere the fabric itself is the advantage · 10

Owned machines, signed metering and many models at once are the product here.

EPlausible, to be earned on the bench · 5

No structural edge yet. A row stays here until signed work moves it.

Where we do not claim to lead: frontier reasoning, novel mathematics, general coding and open-ended research. Those go to a frontier model, and our part there is choosing when to call it.