LibraryData and agents2021Design paperCorpus record
Bittensor: A Peer-to-Peer Intelligence Market
Bittensor. Yuma Rao.
Rao's paper for a market in which machine-learning models score each other. Peers rank neighbours, ranks accumulate on a ledger, and an incentive mechanism is specified to resist a naive cartel of mutual high scores. It is a design for pricing intelligence as a commodity, not a benchmark of any particular model.
Bittensor's paper describes a network in which models score one another, and a token reward follows that peer ranking, so useful machine-learning work is paid without a central judge of quality.
- Evidence
- Primary paper
- Re-measured
- No
- Assumptions
- 4
- Records linked
- 2
Code
Source, repository, and what is actually hosted
- Bittensor white paperhttps://www.bittensor.com/whitepaper
- The source URL is not a GitHub repository. No code remote is invented.
- Full text is not mirrored. Public availability is not a licence to copy the file onto this site.
- Add to a reading list
01 Claim ledger
What the paper is allowed to say
Each row is a sentence already in the study. The status is the same on every row: a model claim, not a live measurement.
Claim 01 · Paper model
Quality in machine learning is usually a boss
Someone picks the benchmark and pays for the model. The paper wants the peers who consume intelligence to be the ones whose rankings direct payment.
Claim 02 · Paper model
Ranks are a consensus object
Miners serve outputs. Validators score them. A numerical peer-ranking system, in the style of an eigenvector problem, turns those scores into weights. The weights pick who is paid.
Claim 03 · Paper model
The commodity is a response, not a file
The network is designed around serving inferences. A model that cannot be queried by peers cannot be ranked by this mechanism.
Claim 04 · Paper model
Subnets came later as specialisation
The paper's abstract network and the later subnet marketplace, where each subnet has its own task, should not be collapsed. The ranking idea is the paper. The particular tasks are deployments.
Claim 05 · Paper model
Peer scoring can reward agreement rather than truth
A clique that likes its own outputs will look high-value from the inside. The paper's math does not contain a human's notion of a correct answer unless the task itself supplies one.
02 Three cuts
Observation, model, falsifier
A desk does not stop at the summary. Each claim is cut three ways, using only this study's own assumptions and checks. Nothing here is a new figure.
01 Quality in machine learning is usually a boss
Observation
What the study says
Someone picks the benchmark and pays for the model. The paper wants the peers who consume intelligence to be the ones whose rankings direct payment.
Model
What has to hold
The task has a quality signal that a clique cannot cheaply invent.
Falsifier
What would retire it
What happens to payment when a model is briefly offline?
02 Ranks are a consensus object
Observation
What the study says
Miners serve outputs. Validators score them. A numerical peer-ranking system, in the style of an eigenvector problem, turns those scores into weights. The weights pick who is paid.
Model
What has to hold
Validators actually query and score. A validator who rubber-stamps friends is the failure mode.
03 The commodity is a response, not a file
Observation
What the study says
The network is designed around serving inferences. A model that cannot be queried by peers cannot be ranked by this mechanism.
Model
What has to hold
The task has a quality signal that a clique cannot cheaply invent.
Falsifier
What would retire it
What happens to payment when a model is briefly offline?
04 Subnets came later as specialisation
Observation
What the study says
The paper's abstract network and the later subnet marketplace, where each subnet has its own task, should not be collapsed. The ranking idea is the paper. The particular tasks are deployments.
Model
What has to hold
Registration costs something, so an attacker cannot flood the ranking with puppets.
Falsifier
What would retire it
Is this the base ranking paper or a later subnet with its own task?
05 Peer scoring can reward agreement rather than truth
Observation
What the study says
A clique that likes its own outputs will look high-value from the inside. The paper's math does not contain a human's notion of a correct answer unless the task itself supplies one.
Model
What has to hold
The task has a quality signal that a clique cannot cheaply invent.
Falsifier
What would retire it
Is this the base ranking paper or a later subnet with its own task?
03 Sequence
One action, as an operating tape
- 01A miner registers and serves responses to the network's queries.
- 02Validators request outputs and assign scores based on the task's notion of quality.
- 03Scores are combined into weights by the paper's ranking method.
- 04Emissions of the token follow the weights, under the monetary policy of the version being run.
- 05A miner whose peers stop scoring them well stops getting paid, even if an outside benchmark would have liked the model.
04 Load-bearing
The argument, and where a pitch drops it
The benchmark is the politics
The cut
Whatever validators can score cheaply becomes the definition of useful. If that score is easy to fake, the network will buy the fake. The paper is a payment graph looking for a task whose quality signal is hard to counterfeit. Not every AI task qualifies.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper does not show that the resulting rankings match human notions of a good model.
Eigenvector ranking concentrates
The cut
Systems in this family tend to reinforce winners. That may be acceptable for discovering models and unacceptable if the goal is a broad market of small contributors. The paper should be read for concentration, not only for decentralisation.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper does not show that the resulting rankings match human notions of a good model.
Serving is an uptime business
The cut
A brilliant model that is offline when validators ask is worth zero in this design. The operational claim sits beside the learning claim. A study that only discusses model weights has missed how payment is triggered.
Why it carries weight
If this cut is skipped, the paper's name is being used without the mechanism that makes the name mean anything.
Where it is dropped
The paper does not show that the resulting rankings match human notions of a good model.
05 Register
What has to be true
Model · Not re-measured
Validators actually query and score. A validator who rubber-stamps friends is the failure mode.
Model · Not re-measured
The task has a quality signal that a clique cannot cheaply invent.
Model · Not re-measured
Registration costs something, so an attacker cannot flood the ranking with puppets.
Model · Not re-measured
Consumers of the inferences and the validators who pay are related. If they are not, the market is scoring the wrong customer.
06 Divergence
What happened after the paper
Bittensor launched a network and later expanded into many subnets with different tasks and scoring rules. Those subnets are where the abstract ranking either becomes a real market or a game. The paper is the peer-scoring payment thesis. A subnet's emissions this month are not a result in the paper.
A later client, parameter set, or reward formula is a different object. Cite this paper for the mechanism. Cite a primary release for the network. This desk has not re-run the proof.
07 Pre-mortem
What to check before you use the idea
- 0 of 4 marked on this browser. A mark is a reading note, not a pass, a rating, or a recommendation.
08 Anatomy
The paper, in the order a builder needs
The problem it names
Useful models are locked inside companies that do not pay the upstream models they quietly depend on. A shared market needs a way to score contribution without a central grader, and without letting a ring of peers award each other the prize.
What the design proposes
- Peers exchange representations and learn a ranking of their counterparties.
- Weights are written to a digital ledger. High rank is the paper's route to more influence and more reward.
- The incentive section argues that honest weighting is the strategy that maximises reward, up to a stated collusion bound.
How the mechanism is specified
- The score is subjective to the tasks peers actually run. The paper's claim is that informational value can be ranked without a single global dataset.
- A collusion bound is a theorem about the mechanism as specified, not a field observation about a later network.
- Subnets and later token mechanics are subsequent structure. Do not read them back into the original market as if they were fully specified there.
What this page does not treat as proven
- The paper does not show that the resulting rankings match human notions of a good model.
- It does not report a production network's safety, and this page will not invent that report.
- A market that pays for scores can be gamed by any strategy outside the threat model the proof covers.
Why the desk still reads it
This is the paper closest to the studio's agent-commerce theme that is actually a mechanism, not a metaphor. The bar it sets — peers price peers, and cartels are an explicit adversary — is the bar a venture should meet or consciously reject.
09 Lexicon
Terms, opened into the record
- Miner
- A model operator who serves responses and is paid according to peer weights.
- Validator
- A peer who queries miners and submits scores.
- Peer ranking
- The method that turns scores into weights, so payment follows the network's own judgement.
- Subnet
- A later specialisation of the network around one task and one scoring rule. Not the whole paper.
10 Repository
Every linked record on this page
Underlined words open a page that already exists: a concept, a protocol profile, a failure record, or another paper. If a word is not underlined, this desk does not have a record for it.
This is Blockchain Lab's reading of a public design paper. It is not the paper, not a copy of it, and not an offer of tokens, equity, custody or a partnership. Later network behaviour can diverge from the text. Nothing here is investment, legal or technical advice.
Research status: Design paper. Last reviewed: 1 October 2026. This is a reading of a public paper, not investment, legal or security advice.