Bittensor Keynote @ Cambridge
Bittensor directs the power of markets towards society's most important digital commodity - artificial intelligence. Not only to build the most powerful intelligence network, but also to ensure that the benefits and the ownership of machine intelligence are in the hands of all, not just the few.
Bittensor essentially generalizes bitcoin style mining primitives to solve a wide variety of problems. Turning energy into digital commodities or intelligence, rather than into currency. It was purpose built for AI, and houses 128 varied companies which tackle problems such as distributed confidential compute, pre-training, fine tuning, RL, inference, computer vision, drug discovery, finance applications, coding agents, distributed storage, deepfake detection, and many more.
Some key recent achievements:
chutes.ai currently serves over 1 trillion tokens per month on TEE enabled machines, across 100+ open source models.
tplr.ai trained the largest decentralized model in the world at 72B parameters, innovating new training techniques to make this possible - SparseLoCo https://x.com/tplr_ai/status/2014075166238621982?s=20
Grail AI released research proving we can conduct decentralized RL runs which match centralized methods in speed https://x.com/grail_ai/status/2021608045705310406?s=20
Jacob Steeves, co-founder of Bittensor, will discuss the philosophical foundations of the protocol and why decentralized AI matters in a world where AI inherently centralizes power. Drawing on the lessons of Bitcoin, he will show how mining primitives can be expanded into a general framework for coordinating valuable work on Bittensor, and why open, permissionless incentives are more orders of magnitude powerful than traditional organizational models for humanity. Through concrete case studies of frontier research being built on the network, Jacob will illustrate how Bittensor is advancing the state of the art today, before closing with a practical look at how students and new contributors can participate in building the future of open intelligence.
📍 Erasmus Room, Queens' College Cambridge
📅 Sunday, 26 April
🕕 4:00 – 5:30 pm
Seats are limited.