Aug 17, 2026 · 1h 25m
Bittensor co-founder Jacob Steeves outlines the future of decentralized artificial intelligence
Bittensor creator Const on Affine, dTAO, "mining reasoning," and more | E2326
This episode provides a deep dive into the mechanics and economic incentives of Bittensor, highlighting how blockchain technology can be used to decentralize artificial intelligence development.
- 1Bittensor uses a competitive subnet model where developers compete for TAO token emissions based on model performance.
- 2The introduction of Dynamic TAO aims to refine the protocol economics and better incentivize high-performing machine learning models.
- 3Operating a permissionless network requires navigating adversarial threats, smart contract risks, and early-stage project failures.
The brief
Bittensor co-founder Jacob Steeves joins the show to explain how the protocol is building a decentralized, peer-to-peer network for artificial intelligence, positioning it as a competitive alternative to centralized tech giants.
By utilizing the TAO token economy, which mirrors Bitcoin with a 21 million supply cap, Bittensor incentivizes developers and validators to build and run specialized machine learning subnets that compete for token emissions.
The conversation addresses the harsh realities of a permissionless ecosystem, including security threats, bad actors, and the controversy surrounding the Templar subnet failure and subsequent rug pulls.
Despite early-stage startup failures, the network is attracting elite talent, including former Google Brain researchers, and launching subnets like Affine to focus on decentralized model training and reasoning.
What was said on this episode
23 statements · 19 positive · 2 negative · 2 neutral
Jason predicts he will make tens of millions from Bittensor-related investments.
“I predict I'm going to make tens of millions.”
Listen at 3:44
Jacob assesses Bitcoin as possibly the most efficient market in history.
“Bitcoin is probably the most efficient market we've ever seen in history.”
Listen at 7:24
Jacob says Bitcoin’s permissionless computational network continues growing.
“And so anyways, Bitcoin invented that first example and it still is growing day to day.”
Listen at 13:07
Jacob positions Bittensor as connecting global computing power with artificial intelligence.
“How do we connect the most power powerful computer in the world to the most important computational problem in the world?”
Listen at 13:46
Bittensor’s mining primitive can support machine-learning inference.
“We could inference machine learning models as example, which is, you know, when you call them and you get the outputs.”
Listen at 16:55
Bittensor enables creation of specialized mining networks.
“Bittensor is a blockchain where you can produce a different type of mining network.”
Listen at 20:42
TAO has a maximum supply of 21 million tokens.
“So Tao has a 21 million cap.”
Listen at 21:50
Bittensor subnets aim to produce inference compute or models faster and better than competitors.
“The whole premise of one of these projects is that they're able to use this permissionless hyper competitive market to produce this digital commodity inference compute or models faster and better than anyone else in the world”
Listen at 27:01
Permissionless excess compute can dramatically reduce inference costs.
“as a consequence we can dramatically reduce the cost of that commodity.”
Listen at 29:31
An honest validator majority determines Bittensor network incentives.
“The majority, the honest majority, will determine the direction of the incentives in the network.”
Listen at 33:53
Bittensor can enforce service levels through programmatic verification rather than contracts.
“Instead of a contractual sla, it's a programmatic sla.”
Listen at 35:36
Bittensor subnet 51 developed GPU-attestation mechanisms without trusted-execution hardware.
“subnet 51 figured out GPU attestation, you know, trusted execution without trusted execution”
Listen at 39:37
Tripling miner payments tripled subnet compute within two months.
“because they've tripled the amount of money they're paying miners, they've tripled their amount of compute in two months.”
Listen at 47:46
Affine is developing a mechanism to measure intelligence as a commodity.
“we're building the, the code, the mechanism that can actually measure that pinnacle element. Like what does it mean to actually grasp in your hand intelligence as a, as a commodity?”
Listen at 48:56
Affine miners produce models intended to generate reasoning.
“what we have the miners on the network do is we have them produce machine learning models that can produce reasoning”
Listen at 57:05
Affine’s smaller model could augment GLM while operating roughly thirty times faster.
“you could run GLM and instead of GLM thinking, you would just talk to the smaller model and then be able to go like 30 times faster.”
Listen at 58:11
Mining latent thought is expected to enable training highly capable models.
“the end goal is that we think that mining that latent space of thought is going to be the prerequisite for us training incredibly good models.”
Listen at 58:19
A Bittensor network creates no value unless it resists adversarial gaming.
“if you can't solve the adversarial problem in the network, then the network produces no value.”
Listen at 1:01:14
Decentralized collaboration is needed to compete with OpenAI’s compute resources.
“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.”
Listen at 1:03:37
Bittensor cannot currently train trillion-scale models without solving distributed bandwidth constraints.
“training a trillion per hour model is sort of beyond what Bittensor can do until we can come up with an algorithm that stitches together all of the compute in a way that gets around this bandwidth problem.”
Listen at 1:05:30
Jacob predicts Bittensor would continue operating without him.
“I think it would definitely continue if I went away.”
Listen at 1:15:26
Jason recommends that people buy one TAO token.
“I think buying one Tao, just buy one tao is what I've been telling folks.”
Listen at 1:17:01
Jason considers a scenario where TAO rises from $200 to $60,000–$120,000.
“What if it is bitcoin and what if it goes from 200 to 60,000, 120,000.”
Listen at 1:17:50
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
What people are saying
Jason Calacanis puts Bittensor’s AI startup thesis in the spotlight
Episode reactions
- The appearance is being credited with putting Bittensor and its competitive AI-model thesis in front of a wider startup audience.
- A circulating clip highlights Const explaining whether blockchain-based systems can compete with—or outperform—conventional AI approaches.
Wider topic conversation
- Broader Bittensor discussion focuses on more than 100 subnets spanning AI agents, inference, robotics, and drug discovery.
- Ecosystem accounts frame Bittensor as a marketplace where specialized AI systems compete for usage and emissions.
Featuring
Bittensor
BitcoinListen to the full episode and explore every guest, topic, and moment on PodLume.