How well do you know the plan?

Thirty tasks: fifteen from the business strategy, fifteen from the product strategy. Graded the way we grade our benchmarks: one known answer per task, no partial credit.

L6Outputs

L5Build

L4Harness

L3Evals and benchmarksyou are here

L2Knowledge graph

L1Infrastructure

Fig. 1 The stack. Today, the execs are the model under test.

Step 2 //: identify the model

Who is under evaluation?

Every finished attempt is logged as a numbered run and shown to the whole exec team.

Keys: A-D to answer, Enter to continue

0/30

Leaderboard best run, ties go to the faster run

Leaderboard //: every run, versioned

Who knows the plan best

Best run decides the rank. Open a row to see every version of that person's runs.