Introducing Burnweight. We’ve been thinking a lot about how Bittensor could better connect subnet demand with TAO value accrual. Burnweight is our proposal to turn the allocation of emissions into a market where subnets compete by burning TAO.

As is evident from my feed, I spend A LOT of time thinking and talking about Bittensor. Sami's early writings on Bittensor were formative for my own thesis on the protocol, and he continues his run of posting absolute bangers with this one below. I'll have to revisit it multiple times, and you should as well. Bittensor is a network state. I'll add to this: Bittensor is the first network state that will actually work. Through years of experimentation and refinement, it has built a way to organize the network's incentive structures so that all actors are aligned in pursuing the same goal of producing valuable digital commodities. This useful work will power the future of open intelligence.

had a lot of fun writing this story I got a new favorite way of describing bittensor out of it as well: Bittensor's ambition is to sustain an economy that can produce intelligence beyond the control of any single company or government, while letting anyone build markets, businesses, and livelihoods around it.

New paper from Harvard, co-authored with Chutes on Bittensor, studying a full year of production LLM traffic and they open sourced the entire trace. Chutes is publishing research, running an open inference network averaging 27B tokens/day, and preparing to release a model trained with a novel decentralized MoE architecture. Their positioning as an open source AI lab is about to be very obvious.

tldr > Fred launches a subnet > it was so bad that it ended up at the bottom of the leaderboard and got deregistered > a protocol update made it to where he couldn't get his TAO registration back > caused him to get embarrased and upset > so he went to sue those behind the protocol lmeow

This past week in Bittensor Land… - Crosschain swaps on Bittensor went live via @taodotcom . Users can now instantly swap USDC and ETH to subnet tokens (!!) -@pareton announced that an optimization identified by miners during their Qwen3.8 campaign was officially merged into vLLM (widely considered the de facto open-source inference engine for LLMs.) - @conjectures_io reported solutions to five open, frontier math problems, with multiple solutions coming from a team of Purdue University students. Just another week.

The more hate I get in my comments about a coin I like, the stronger my conviction Has worked out so well so far for UNI / ENA / NEAR / ZEC Was on a call recently and we chatted about this - consensus in crypto is quite a scary thing as there is no marginal bid outside of our circle esp once everyone is aligned. V different from TradFi land where there is a perpetual marginal buyer from passive flows like 401Ks, mutual funds, pension funds etc. so even tho the asset has reached “investable consensus” there’s still the next buyer Crypto on the other hand… if you actually think back to a time when a trade became consensus, VERY OFTEN the local top came in shortly after and then a corrective wave came. Ofc the strong ones eventually goes on and make new highs - one good example for this is HYPE on every consensus that PERPS is the future and hype is the cleanest expression of that trade. It is, but it’s consensus, and outside of our circle there’s not much marginal bid (which will change in time ofc as the asset matures) That’s why I like the non-consensus trades NOW because of where we are at in the current cycle. Any uptick in risk appetite and we get a reprice in valuations off the back of BTC/ETH/SOL beta + maybe some premium in uptick on fundamentals Getting exposure to risk now > later, so i can focus on cutting instead of adding risk. Anw long winded way of saying consensus in crypto is not always a good thing

bittensor is a network state

Conjectures is solving decades-long-open math problems at a rate of 2 per day now. Powered by Bittensor.

When you point enough machine and human intelligence at a problem it simply gets solved. Many companies have hard problems that they simply don't have the resources to handle. $PRL is a good example: pearlpolymath.com/ Mathematics is just the beginning.

I've demoed this and was blown away by how seamless it was to go from USDC on Ethereum directly into subnet tokens Big week for Bittensor's cross-chain accessibility. Subnet tokens on Base thanks to @forevermoney_ai and now Ethereum thanks to @taodotcom

1/ The very first 404 Game Jam Competition is now officially live. We’re giving creators two weeks to build a playable 3JS game using 404’s game generator repo. Simply point your agent at the repo, build a game from a single prompt, and compete for 10 TAO ($2,300). The competition runs from Sep 11 to Sep 25, with winners announced live at @ExploitSummit . Details below 🧵👇

we will need compute (@lium_io, @TargonCompute) we will need inference (@engyai, @chutes_ai) we will need data (@Hippius_cloud) we will need knowledge (@teutonic) If we are going to have free intelligence (@affine_io)

Can’t wait for the world to learn that this is bittensor

I tweeted this after reading into Pearl and was hinting at both things you mentioned. On your first point, most people started talking about “XYZ Money” after Zcash started running. But the reasons I’m hearing for why to buy Zcash are “it’s only 1% of Bitcoin” or “it’s a quantum hedge on Bitcoin” or “it’s a private version of Bitcoin”. For me, this basically points to the belief that there can only be one crypto money. For Pearl, my understanding of the thesis is that they believe it can surpass Bitcoins hashrate, making it more secure, and thus a better money than Bitcoin. Whether it’s Private money or AI money, they’re all competing to be the ultimate form of money. The term AI Money doesn’t even make sense to me. Then on your second point. Zcash and Bitcoin have been at this shit for 10 years. The way it happened for Bitcoin cannot be repeated imo. Zcash might be able to do it, tbd. Just comical to me that new projects are coming out guns blazing saying they’re gonna be money. I haven’t looked into Nock much yet (I will next), but the difference in Pearl to TAO and Pluralis is that the latter don’t have the primary objective of of creating money. Bittensor wants to create a decentralized economy capable of producing valuable intelligence, which would then make TAO the money of its ecosystem. Pluralis wants to produce models and inference them all within their ecosystem, creating an economy around that work which would make its token the money of its ecosystem. None of those projects are saying they want to be considered as money outside of their ecosystems

I sleep amazing at night fading any network trying to become “ai money” Study all the coins that intentionally tried to become money

Don’t mind if we keep accelerating

This past week in Bittensor Land: - @oroagents was accepted into Y Combinator’s Fall 2026 batch, making it the first crypto company with a live token to be accepted into YC. - @lium_io conducted a $1M revenue-funded buyback and burn, the largest single buyback in Bittensor history. - @_redteam_ announced that it is on track for integrations with three banking companies in the top 10% of the Fortune 500. - @conjectures_io reported that miners found a solution to the near-half-density case of Green’s Problem 51, a math problem that had remained open for 16 years. - @jon_durbin of @chutes_ai shared a dashboard for the Parallax-8B training run, an MoE model being training on 240 RTX 5090s distributed across 13 countries, with a reported compute cost of only ~$11/ billion tokens (!). Just another week.