Marketplace vs Measurement

The short version: Braintrust

Braintrust is a decentralized AI marketplace — a freelance platform that connects enterprises with AI talent: prompt engineers, AI-native developers, model specialists. It does what a talent marketplace does: it sources people. Profiles, ratings, project history, and bid-based matching get you a candidate. What it does not do is measure how efficiently that candidate actually drives AI tools.

SigRank is the measurement layer that fills that gap. It reads token telemetry from any AI tool an operator drives — Claude Code, Cursor, Copilot, Gemini, 15+ others — computes the cascade efficiency (Υ Yield), and ranks them on a global leaderboard. Braintrust tells you who claims to be good with AI; SigRank tells you whether they are.

Feature comparison

Sourcing talent vs measuring skill

Braintrust's value is discovery: it surfaces AI talent through a decentralized marketplace where talent sets their own rates and enterprises post projects. That is a sourcing problem, and Braintrust solves it well. But "uses AI" is not a skill — "uses AI efficiently" is. Two operators can both list "AI-native" on their Braintrust profile and have ten-fold different cascade efficiency. One reuses cached context and compounds signal; the other re-pastes stale context every turn and burns tokens. The marketplace cannot see that difference.

SigRank's Υ = cache_read × output / input² measures it directly. It rewards the operator who compounds cached context into output and penalizes the one who burns input without leverage. Braintrust tells you who is available; SigRank tells you whether they are good at the part that matters.

From self-reported claim to signed score

Marketplace profiles are self-reported: an operator writes "AI-native" and lists projects. SigRank replaces that claim with a verifiable credential. The CLI reads token telemetry locally, computes the cascade metrics, signs a snapshot with ed25519, and publishes it to the leaderboard. The score is server-verified, tamper-evident, and public. An operator hired through Braintrust can share their SigRank rank as proof of efficiency — not a claim, a measurement.

Frequently asked questions

Does SigRank replace Braintrust?

No — the two solve different problems. Braintrust is a decentralized marketplace that connects you with AI talent: freelancers, prompt engineers, AI-native operators you can hire. SigRank is a scoring layer that measures how efficiently any operator drives AI tools. Braintrust finds you AI workers; SigRank tells you how well they (or you) actually use AI. They are complementary — you could hire talent on Braintrust and then score their token efficiency on SigRank.

Does Braintrust measure AI operator skill?

Braintrust surfaces talent through marketplace profiles, ratings, and project history — the same signals a freelance platform uses for any skill. It does not measure how efficiently an operator drives AI tools. It cannot tell you whether a candidate compounds cached context into output or burns fresh input every turn. SigRank measures exactly that: the Υ Yield (cache_read × output / input²), compression ratio, SNR, Leverage, and Velocity that define token-cascade efficiency. Braintrust tells you who claims to be good with AI; SigRank measures whether they are.

Why does measuring AI skill matter for hiring?

Because "uses AI" is not a skill — "uses AI efficiently" is. Two operators can both list "AI-native" on a marketplace and have ten-fold different cascade efficiency. One reuses context, writes tight prompts, and compounds signal; the other re-pastes stale context and burns tokens. SigRank's leaderboard makes that difference visible and comparable. If you hire AI talent through Braintrust, asking for a SigRank score turns a self-reported skill claim into a measured, signed, verifiable number.

Can I use SigRank alongside Braintrust?

Yes. They operate at different layers. Braintrust is where you find and hire AI talent; SigRank is where that talent (or you) scores their AI driving. An operator enrolled in SigRank can share their leaderboard rank, class tier, and Υ trajectory as a verifiable credential — no different than a portfolio, except it is a signed snapshot of measured efficiency rather than a self-reported claim. Run the SigRank CLI alongside whatever AI tools you drive, submit a signed snapshot, and your score is public on the leaderboard.

What is the difference between an AI talent marketplace and an operator leaderboard?

A marketplace (Braintrust) is a sourcing surface — it connects buyers with sellers of AI labor. A leaderboard (SigRank) is a measurement surface — it ranks operators on a canonical efficiency metric. The first answers "where do I find AI workers?" The second answers "how good are they at driving AI?" Marketplaces rely on self-reported profiles and ratings; leaderboards rely on measured, signed telemetry. They are not competitors — they are different layers of the AI talent stack.