# This Week in Startups E2326: "Bittensor creator Const on Affine, dTAO, 'mining reasoning,' and more"

- Title: Bittensor creator Const on Affine, dTAO, "mining reasoning," and more | E2326
- Channel: This Week in Startups (hosts Jason Calacanis and Lon Harris)
- Date: 17 August 2026
- Length: 1:24:54 (podcast audio)
- URL: https://www.youtube.com/watch?v=quq00obUVQ4
- Jacob's start offset: none given; Jacob joins at 2:40, first real answer at 3:50
- Method: YouTube blocked from this VM. The show's podcast RSS feed (Libsyn) carries the same episode; the MP3 was downloaded from the feed and transcribed locally with faster-whisper `medium`. Timestamps are from the podcast audio and are accurate to a few seconds. The YouTube cut may differ by a few seconds around ad breaks.
- Format: research transcript. Hosts paraphrased; Jacob's key lines quoted verbatim. Ad reads skipped.
- Speakers: JASON = Jason Calacanis; LON = Lon Harris; JACOB = Jacob Steeves.

## Outline with timestamps

0:00 Cold-open montage of lines from the episode, then ads. JASON says he is "TAO-pilled," recounts his Bitcoin history, says he bet on Bittensor because "these subnets actually solving problems in the real world." JASON predicts he will "make tens of millions."

2:40 JACOB joins. On the name: "Just Jake, it's fine." Const comes "from Constantine, the emperor."

3:50 JASON: why did you create Bittensor? JACOB starts with what Jason skipped, Ethereum: it abstracted Bitcoin's contract layer into smart contracts (MakerDAO as the example). Then the second Bitcoin innovation, proof-of-work mining: "aggregate all of these contributors from across the globe together to solve this one very difficult computational problem." The consequence: "this market for a digital commodity... the first digital commodity, something that you could mine."

6:12 JACOB defines the term: "It's computationally defined commodity. Oil is not computationally defined. It's physically defined. But a Bitcoin hash is actually computationally defined." Bitcoin is "probably the most efficient market we've ever seen in history. Anybody anywhere can buy and sell it. Anybody can contribute to it from any place on earth by plugging something into the wall." Price is variable; "the proof of work power of Bitcoin is just purely exponential. It just never goes down."

8:28 JASON asks for plain-English definitions of peer-to-peer and permissionless. JACOB: peer-to-peer is "I can send directly to you without an intermediary," the first sentence of the whitepaper. Permissionless "was a quality that he needed to get to that." Then the passage on blindness: "Think of it like the ultimate form of non bias... a permissionless market is purely blind... It's purely a meritocracy. We don't care if you're in a communist country. We don't care if you're in a democracy. We don't care if you're 12 years old or 72 years old." The street-lamp story: he knew Bitcoin was truly permissionless "when people started hijacking like New York City public lights and run a cable into their apartment to make Bitcoin."

11:36 JACOB: "we just measure the output. And when you just measure the output, it means you can optimize the output." Permissionless systems are hard "because anybody anywhere can try to cheat"; before Bitcoin "everything was manually done by humans and humans have biases."

13:08 The founding sentence: "a really powerful computer should be applied to the most important computational problem of the 21st century, which is artificial intelligence." "How do we connect the most powerful computer in the world to the most important computational problem in the world? And that's the founding raison d'être."

13:58 Mark Jeffrey in the chat: Bitcoin is a new way to run a company. JACOB: the invention needed was an "abstract consensus mechanism." "A hash is just binary. It's just true or false." Measuring AI is "very high dimensional," like "the informational significance of a 1,024 dimensional vector."

15:31 JASON on Bitcoin's energy critique and the shift to a compute-constrained world. JACOB: the original intention was training; the primitive turned out to apply to inference and compute too. Analogy: DAOs failed first, then MakerDAO. "At first, we tried to do information mining, which was cool, but not enough to push AGI." Higher-order commodities "require somebody going on and solving a problem": hardware, then software, then "the innovation and the algorithm and the intuition and the creativity."

20:18 JACOB's one-liner for the platform: "Ethereum is a blockchain where you can create lots of smart contracts. Bittensor is a blockchain where you can produce a different type of mining network. Think of it like a contract but it mines a particular commodity." Includes a joke in the list of commodities: "stealing API keys and selling them." Tokenomics: 21 million cap, half a TAO per 12-second block after the first halving, minted to subnets "as basically additional liquidity."

22:40 JASON asks for the top subnets. JACOB insists on the primitive first: "Bittensor is now a meta. It's actually an abstraction on top of itself." A subnet is "an open network that you can join. And by join, you burn a little bit of a token to prove that you're willing to play the game." Worked example: run a model, answer Jason's Claude Code requests ("the capital of Texas is Austin"), validators check you did it faithfully and fast. The visual: a curve of 256 participants; "think of it like a league, like the Premier League." "each of the networks on Bittensor has this relegation system." Names Lium, Engy, Affine as top examples of compute, inference, model competitions. Test of a good subnet: it uses "this permissionless hyper-competitive market to produce this digital commodity... faster and better than anyone else in the world because we're using the power of Bitcoin. And if that's not their premise, it doesn't make much sense to build on Bittensor."

27:15 Alpha tokens: "a secondary cryptocurrency, which also has 21 million cap that only exists inside of that network." "it only has value if at the end of the day, people are going to want to buy that token to get access to the computers in this network."

28:51 JASON on Engy's half-price tokens. JACOB: "behind that product is this liquid swarm of anybody in the world that can enter and try to reduce that price over time." JASON's rack-idle-for-100-days example; JACOB: "Precisely."

31:09 History: "People have been building mechanisms on Bittensor for two years." The art of the incentive mechanism: "how do you build the permissionless adversarial game where even the worst person in the world, if Hitler and the devil were to have a child and that child were to mine on your subnet, they would just produce value." "you can take away intent."

33:30 How consensus works: validators all run the same checking code; "if more than 50% of them are running the right code" they reach consensus. "the honest majority will determine the direction of the incentives." "we can build a network that's permissionless and verifiable and auditable." Weight is proof-of-stake, not proof-of-work.

35:02 Engy example: a miner passing off an old model as GLM 5.2 gets caught. JASON: a built-in SLA? JACOB: "instead of a contractual SLA, it's a programmatic SLA." Names TOPLOC (Prime Intellect) which hashes hidden states. "it's one thing to write a paper, it's a completely other dimension to write it on Bittensor because we have the smartest people in the world and we have the smartest hackers in the world and they're just right next to each other."

37:23 JASON: you are the police, the Jedi. JACOB: "This is the role that we play, reluctantly, while it's still required." Subnet life: "they build the mechanism, and it breaks, and then they fix it, and then it breaks, and then they fix it, and then they get it right, and then it starts working, and then they make a million bucks." Adversaries are "effectively nation state and militias." People who call about starting a subnet: "you're really not coming at it with the right level of intensity because your adversaries on this network are very serious. They could literally be a nation state. It could be North Korea." Praises subnet 51 for GPU attestation without TEEs.

40:31 dTAO: "let's apply ourselves to ourselves. We're really good at building these permissionless mechanisms. Let's build a permissionless mechanism which selects permissionless mechanisms." Launched "just over a year ago."

41:17 On the builders: "the entrepreneurs in this chain are so much better than I am, and many times smarter. And I listened to as many of them as I can because they all often are mentors and guides for us."

42:51 JASON: who comes to build? JACOB: "A good number of them are scammers. And you kind of can't avoid that because you have the doors open." Image: "we can like open our eyelids the most, that's what is going to make us win. But we're staring right at the sun." Then: "there's a lot of stigma in crypto, well earned, honestly." The people who come are those who can "break past these fiat-funded companies." "we need to do something different and better and more powerful than what they're doing. Otherwise, they'll just beat us with more cash."

46:12 JASON: you are playing the grandmasters. JACOB: "And yet, the largest supercomputer in the world is Bitcoin. So this is the technology that has the example of the only thing that's beat them in size." On scaling: "a mechanism that's just described by really just a code base and a permissionless market, you just pump more money through it and it just scales." A compute subnet "tripled their amount of compute in two months."

47:49 The goal: "not just to do more compute and more inference... Our goal is actually to truly go head to head with the elite centralized labs at intelligence. This is, I think, the pinnacle commodity that we can measure." "measuring intelligence is still possible. It's just very, very, very abstract and hard to get at." Affine, subnet 120, is "my personal mission."

49:22 How to start a subnet: "if we had a board, then we wouldn't really be permissionless." Registration is a Dutch auction; currently 608 TAO, about $121,000, "which is a lot, too much, and we want to lower that a lot." The burn goes into the initial TAO/alpha liquidity pool. "It's like the US dollar... it has all of these amazing companies inside of the US system and it has a massive network effect."

53:16 Why dTAO exists: "we still have 10 million TAO to distribute." "we can actually elicit the internal machinery of capitalism, which is thousands of individuals from degens to scientists and the like, to long-term investors." "if you were to sit down right now, Jason, and go, how can we order all of the companies in the United States, and you didn't have the NASDAQ, how would you do that?" Warning to newcomers: "there's some rotten milk and there's some cream and the cream is more expensive." "you're actually governing Bittensor. In that way, Bittensor is highly decentralized... it's tens of thousands of nodes contributing and making informed bets on the truth."

55:39 LON asks about Affine. JACOB: last year they produced a model better than Qwen's best at 35B, then Qwen shipped a better one. "it's not an understatement to say that we are up against nation states because this is truly like China versus the United States." Affine's motto is "mining reasoning." The reflexive test: "you would think that I'm smart if whenever you talked to me, you felt a lot smarter yourself." Miners produce models whose reasoning makes other models answer correctly; could act as adapters so a big model runs "like 30 times faster." End goal: "mining that latent space of thought is gonna be the prerequisite for us training incredibly good models."

58:45 Who can participate: "you don't need to be a machine learning engineer... You could literally just get your OpenClaw and send them the website." Remote work through games: "anybody in the world can contribute almost without communicating by just sharing the best work they can." His sales pitch: "I don't actually really sell TAO the token. I sell that you can be part of this." "for some people mining Bittensor is like the most fun game they've ever played ever."

60:22 LON: how much of a subnet is adversarial 4D chess? JACOB: "that is the entirety of the project. All the other stuff is fluff." "if you can't solve the adversarial problem in the network, then the network produces no value... it's just lipstick on a pig." At the core of a good mechanism there is "a very simple, compressed idea... that is like elemental." Inference verification is really "information checking."

61:57 JASON: the Jensen moment and the Templar controversy. JACOB explains why Templar is named Templar: the holy grail of the field is decentralised training, "where you have a computer and I have a computer, Lon has a computer, and they are making the same machine learning model at the same time." "in order for the rest of us to organically self-organize and to take on OpenAI, we need to come up with a way that we can work together because only together we will have enough compute." The bandwidth wall: merging weights every step means terabytes per step. Fragility: "If you have one person there that's fucking around with my language model, it can destroy the whole thing."

67:22 JASON: candidly, what happened with the "rug pull"? JACOB: "in any company, if a CEO starts a company, they can just leave." The person "sold their shares, more or less" and "wanted to cover up that they were leaving with a reason for leaving." "they wrote an article that basically said hey we're not bad for leaving, it's the mean network and Const for being an evil dictator because he sold some of our token." "the project fell apart immediately... you can't burn your reputation like that." JASON compares it to founders who take YC money and stop shipping.

71:19 JACOB on the double edge: "it's a double-edged sword to allow anybody permissionlessly to have access to early stage startups." "crypto has become a lot more prickly and has thorns now." "overall in the Bittensor ecosystem, we've made people a lot more money than they've lost." Why it is worse in crypto: founders can sell into a liquid market, "hence rug pulls." The fix, conviction: teams can lock tokens so "if they want to go and sell them in one big sell, that's a public event." "a lot of the teams go up to the plate and be like, yeah, you know what? I'm locking this thing perpetually for years."

74:27 JASON: the foundation's role, your stake, and how far are you from making yourself irrelevant? JACOB: "I think it would definitely continue if I went away, most certainly... I hope that people would miss me." 128 teams understand the tech. Two foundations: Opentensor in Canada, "no longer involved with development," and a second, "more shielded" one "purely about upgrading the chain," with on-chain governance ("the triumvirate") to be expanded this year. "I'm no longer involved with Opentensor. I work just in the [second] Foundation and I program. That's my language... I am a programmer at heart, that's how I speak."

76:49 JASON's "buy one TAO" philosophy: a ticket to watch the experiment, a vote of confidence, and maybe Bitcoin-like upside.

78:25 JASON: how do speculators and pumping affect the project? JACOB: "it's inevitable... It's the nature of markets, but it's not the goal. The reason why we have TAO in the first place is so that everyone can have some, and that everyone can join." "There's always people that are gonna say shit on both sides. I can't stop them." He prefers people "that see it... as a really, really powerful technology" and "the third path for what artificial intelligence can be born out of." "I'm in all the price chats as well."

80:01 JASON closes; hosts discuss Jason's own position and transparency about his trades (not Jacob).

## Jacob's stage-worthy lines (verbatim, Whisper medium)

- "How do we connect the most powerful computer in the world to the most important computational problem in the world?" (13:40)
- "Oil is not computationally defined. It's physically defined. But a Bitcoin hash is actually computationally defined." (6:30)
- "a permissionless market is purely blind." (9:30)
- "we just measure the output. And when you just measure the output, it means you can optimize the output." (11:40)
- "Think of it like a contract but it mines a particular commodity." (20:30)
- "each of the networks on Bittensor has this relegation system." (26:00)
- "instead of a contractual SLA, it's a programmatic SLA." (35:40)
- "we have the smartest people in the world and we have the smartest hackers in the world and they're just right next to each other." (36:40)
- "Let's build a permissionless mechanism which selects permissionless mechanisms." (40:40)
- "we can like open our eyelids the most, that's what is going to make us win. But we're staring right at the sun." (43:00)
- "Otherwise, they'll just beat us with more cash." (45:20)
- "you just pump more money through it and it just scales." (47:40)
- "we can actually elicit the internal machinery of capitalism." (54:00)
- "there's some rotten milk and there's some cream and the cream is more expensive." (55:00)
- "I don't actually really sell TAO the token. I sell that you can be part of this." (60:00)
- "that is the entirety of the project. All the other stuff is fluff." (60:40)
- "only together we will have enough compute to compete with the big guys." (63:50)
- "I am a programmer at heart, that's how I speak." (76:50)
- "The reason why we have TAO in the first place is so that everyone can have some, and that everyone can join." (79:20)