Travis discusses Bittensor Governance, Proof of Useful Work, dTAO, and Conviction Lockups on Ventura Labs Media Chat

For his second appearance on the Ventura Labs podcast, Travis ‘Tao Templar’ Millott, sat down with the host for a wide-ranging conversation on what has changed since their last conversation and what is still ahead.

The discussion covered the trajectory of Bittensor’s decentralization, the political risks the network still carries, how Travis frames Bittensor against Bitcoin and fiat as monetary systems, and where the next round of mechanics, including Conviction lockups, flywheel design, and AI-owned subnets, is heading. What follows are the strands worth holding on to.

Highlights From the Conversation

The discussion ran from the post-Dynamic $TAO state of Bittensor through to the future of agent-owned subnets and the political risk profile of the network as it integrates with traditional finance. 

The points worth pulling out are below:

a. Bittensor’s decentralization trajectory has made real progress since the last conversation, with Dynamic TAO standing as the single largest milestone.

The shift from a validator cabal allocating emissions to a market-driven mechanism allocating them based on real flows is, in Travis’s framing, a meaningful structural improvement. Before Dynamic TAO, a small group of validators effectively chose which subnets received chain support, and that decision was made with incomplete information, heavy manual lobbying from subnet owners, and limited ability to update frequently. Markets handle that allocation problem far more efficiently than any group of corruptible individuals could.

b. Substrate-level decentralization, however, remains incomplete, and Travis would like to see quantum resistance addressed before full decentralization is rushed. 

During a recent Revenue Search session, the team behind the relevant quantum work indicated that Bitcoin itself may be vulnerable to quantum attacks within approximately two years. Full substrate decentralization makes chain upgrades slower, and Travis would prefer Bittensor remain semi-centralized at the substrate level long enough to ship quantum-resistant cryptography ahead of that window closing.

c. The Templar exit paper attributed to Sam Dare carries no weight, in Travis’s assessment, because it was generated by an LLM rather than written by Sam directly. 

The text contains identifiable inaccuracies that suggest a prompt-and-publish workflow rather than considered authorship. Travis declines to engage with the document’s claims because there is no way to separate which views actually belong to Sam Dare from which are simply model output.

d. The cleanest framing for what Bittensor actually does is that it allows the creation of worldwide competitions to produce commodities. 

The sports tournament analogy holds: The same way the world identifies the best soccer team or the best pole vaulter through structured competition, Bittensor identifies the best forecaster, the best AI image detector, the best storage provider, the best inference service, and beyond. When the competition is designed correctly, the global decentralized version cannot be beaten by any centralized entity attempting the same thing.

e. The monetary frame Travis returned to throughout the conversation is that Bitcoin represents proof of hard work, fiat represents proof of violence, and Bittensor represents proof of valuable hard work. 

The work itself does not have to be technically difficult to qualify, but it must produce something the market actually wants, and the difficulty comes from competing against the entire world to produce it. This is the structural reason Bittensor commodities tend toward genuinely useful outputs rather than arbitrary computation.

f. If Bitcoin were structured as a subnet on Bittensor today, it would rank somewhere in the middle or bottom of the pack. 

Hash production is not particularly productive relative to commodities like compute, inference, and forecasting being produced by the strongest subnets on the network, and the market would reflect that. Travis acknowledged the question as a useful thought experiment that exposes how productivity actually gets measured inside the Bittensor model.

g. The core political framework Travis brought to the conversation is that everyone is corruptible by power, without exception. 

The standard appeal to “good actors” or “principled leadership” misses the structural point. Systems should be designed assuming corruption rather than designed to depend on the absence of corruption. The right question for any politician or protocol leader is not whether they are corruptible, but how they will become corrupted, and what design choices contain that drift before it actually occurs.

h. Applied directly to Bittensor, that means Jacob ‘Const’ Steeves has too much power and needs to keep being decentralized away from. 

Travis was explicit on this point. Const has stepped down as CEO of the OTF (the Opentensor Foundation), but retains significant influence over the network, and that influence carries the same structural risk that any concentrated power carries. Decentralizing const’s role further must be a continued priority for the OTF, with the explicit caveat that Travis would like to see quantum resistance addressed first.

i. AI-owned subnets are partly real today and likely to become much more so within a decade. 

The clearest current example is Const running Distil (Bittensor Subnet 97) with an LLM handling code updates and community responses, which is partial automation of the subnet-owner role. Travis projects that within roughly 50 years, an end user will be able to request a solution to a problem from their LLM, and the LLM will spin up a Bittensor subnet, run the competition to solve it, and shut the subnet down once the solution is produced.

j. On trust signals from subnets, buyback-and-burns sit at the top and Conviction lockups sit just below.

Buyback-and-burns are verifiable on-chain and prove that revenue is real. Conviction lockups, where token holders voluntarily lock their subnet ‘$ALPHA’ for fixed periods, demonstrate long-term commitment without sacrificing the liquidity that some subnet owners legitimately need for reinvestment. The deeper the lockup and the longer the duration, the stronger the signal.

k. Lium (Subnet 51) is the cleanest existing example of a working flywheel on Bittensor. 

The subnet burns 90% of its miner emissions, executes regular buyback-and-burns on revenue, and keeps the subnet owner take small with manual burns layered on top. The net result is that alpha supply growth at Liam is materially lower than the network average, and the token behaves much more like a constrained-supply asset than an inflationary one.

l. $ALPHA supply growth is one of the more useful proxies for whether a subnet’s flywheel actually works. 

Most subnets see roughly 5% monthly supply growth in their $ALPHA, and subnets that burn more than they emit over a meaningful window are running real flywheels regardless of what they say in marketing. The metric is publicly verifiable on-chain, which is part of what makes it more reliable than the “trust me, the deal is coming” narratives that often surround other revenue claims.

m. The arrival of larger traditional players is broadly positive, with one specific risk attached. 

Chamath, Jensen mentioning Bittensor through Templar, Grayscale’s ETFs holding $TAO, and Jason Calacanis on All-In all signal that institutional attention is real and growing. Traditional investors, however, sometimes carry an instinct to ask for “just a little centralization” when something on a decentralized network displeases them. 

That impulse, if accommodated by the OTF, is the single most likely path toward Bittensor losing what makes it valuable in the first place.

n. If Bittensor goes meaningfully permissioned, Travis would expect the original community to fork the chain rather than accept the permissioned version. 

The reference point is that permissioned subnets blunt a network’s potential as a venue for real global competition. The mechanical comfort underneath this scenario is that the code already exists. 

In the same way Templar’s collapse produced a workable fork target for Subnet 3, a corrupted version of Bittensor could be forked back into a permissionless version without losing the underlying design.

Why This Matters

What makes this conversation valuable is that Travis is one of the most consistent independent analysts on Bittensor, and his read on the network has held up across multiple major upgrades. 

The thesis he is articulating now is the same one he was articulating before Dynamic $TAO, before the Templar collapse, and before the current cycle of traditional finance integration. The structure of the protocol is moving in a direction that produces market-driven, globally competitive AI commodities, and the network’s value compounds to the extent that the structure stays permissionless.

The risks Travis names are real and worth taking seriously. The quantum vulnerability window is short. The concentration of power around const has not been fully resolved. The pull from new institutional entrants toward partial centralization is exactly the kind of slow drift that protocols rarely survive without active counter-pressure. And the mechanics around Conviction lockups, subnet-owner key transfers, and AI-owned subnets are still being designed in real time, with real consequences for which subnets survive the next cycle.

The signals worth tracking from here are simple. Whether the OTF continues to decentralize meaningful authority away from any single individual. Whether quantum-resistant cryptography ships before the window closes. Whether the strongest subnets continue to harden their flywheel mechanics around buyback-and-burns and Conviction. And whether the protocol resists the structural temptation, when traditional money asks for it, to introduce permissioned layers. If those signals trend in the right direction, Bittensor’s destiny as the venue for permissionless global commodity competition stays intact. If they do not, the protocol still has a fork path back to the design that made it valuable in the first place.