Introducing Root Reborn: a new market for optimizing TAO yield across Bittensor. Validators compete to build the best-performing subnet baskets. Successful strategies can attract more stake, so more root yield goes toward the subnets that create the most value.
48
133
604
102,754
SOMA cut a prompt’s token count by 76% live, reducing the text you pay AI models to process. On the latest Novelty Search, Subnet 114 presents its compression tool for coding agents, already in early access for Copilot and Codex. The team reports ~15% lower inference costs on average across coding tasks. @say_gm_ has integrated SOMA’s open-source compressor, and Dendrite uses it in daily development. Hosted by @const_reborn Full episode in the first comment
9
20
117
29,061
Novelty Search :: Bittensor Subnet 114, SOMA :: Open, Competitive Context Compression nitter.net/i/broadcasts/1DGleVWkm…
8
19
98
13,045
Openτensor Foundaτion retweeted
The full Exploit lineup is now LIVE⚡ Two days of launches, live demos, adversarial debates, workshops + experiments across AI training, robotics, agents, compute, privacy and more. New products. Hard questions. Big bets. The pioneers of distributed AI, all in one place. See what’s coming ↓ exploitsummit.com/agenda/
1
16
66
12,607
.@openroboto just built a decentralized alternative to @Figure_robot’s Index for collecting robot training data, with real-world capture hardware. Open robotics on Bittensor now has its data network to challenge the industry’s biggest labs. Bittensor is coming for physical AI.
Today we’re launching OpenRoboto Shift, opening a new chapter for OpenRoboto. Shift is a decentralized network for collecting egocentric robotics data: first-person video of real people doing real work. Only Possible on Bittensor. Explore Shift → shift.openroboto.ai/
15
41
240
24,072
A Harvard research team and @chutes_ai just released a public dataset covering one year of real-world LLM inference on Chutes: 6.12B requests across 9,174 models. Technical usage data from a Bittensor subnet is now open to the wider AI research community.
Announcing one year of LLM inference metadata traces, with 6.12 billion requests. We hope this dataset can support research on real-world LLM serving workload understanding, system design and infrastructure optimization. Explore the dataset and learn more: data.agentic-system.org Driven by our great graduate student William Nixon and in collab with @jon_durbin @airesearch12 @chutes_ai
18
61
376
27,102
Openτensor Foundaτion retweeted
Welcome to the era to decentralized post trained models that beat the frontier in their classes.
Introducing Reliquary-4B. A 4B math & code model trained with reinforcement learning. Anyone could join the network and contribute rollouts. Independent miners chose the prompts and generated the rollouts. The protocol verified them and trained the model. Here’s the model and the research behind it.
19
71
499
37,456
This Thursday on Novelty Search :: SOMA, SN114 @SomaSubnet is building a decentralized marketplace for AI services, starting with context compression. Miners compete by submitting compression algorithms that make AI workloads more efficient. Validators test & score miners on the balance between compression and retained performance. Thursday :: 5PM EDT / 9PM UTC Live via Youtube and X
4
27
163
74,548
.@near_ai’s confidential models are now available through @say_gm_ on Bittensor. GM is becoming a gateway for private AI across ecosystems.
NEAR AI Cloud's confidential inference is now live on @say_gm_'s confidential tier. SayGm reaches dozens of models through one API key, and runs its own routing inside an Intel TDX enclave rather than on ordinary servers.
12
56
357
20,852
Openτensor Foundaτion retweeted
The Hippius mobile app is live on Android. Your photos and files, backed up automatically and end-to-end encrypted on your phone before they ever upload. The same account and storage you already use on desktop and web. Your whole cloud, now in your pocket. iOS is coming next.
27
77
331
55,222
“You can literally build your Discord on top of Beam.” On the latest Novelty Search, @b1m_ai demonstrates how Bittensor’s Subnet 105 can power live video apps without developers running their own streaming servers. The team also previews Beam Studio, its tool for automating data transfers, and explains how parallel transfers can move 1 TB across Cloudflare R2 regions in under two minutes. Hosted by @const_reborn Full episode in the first comment
14
24
128
10,633
Openτensor Foundaτion retweeted
Pareton miners found it, vLLM merged it. An optimization from our Qwen campaign on #Bittensor SN10 is now upstream in @vllm_project: ~4% more throughput at batch 4–8 for Qwen3.8 with MTP speculative decoding. Open competition → open-source wins. PR: github.com/vllm-project/vllm… 1/5
4
18
64
13,580
Openτensor Foundaτion retweeted
Today, we’re announcing a solution found by our miners to both parts of Erdős Problem 14, open for over 34 years. The result proves a square-root lower bound on exceptions to unique representation as a sum of two elements of any set of natural numbers. Verified in Lean through Conjectures. Full proofs below.
18
56
353
22,450
.@webuildscore just launched Score Studio, a Bittensor-powered platform where independent engineers can build and deploy vision systems using Score’s models. Through Manako, Score brings that technology to enterprises like AVIA and has signed more major clients in recent weeks.
12
56
271
14,206
Bittensor gives Score access to top computer vision engineers from around the world, competing 24/7 to develop state-of-the-art models for its products. Manako and its partners are taking that work into paying enterprises. Score Studio now opens the same research capability to independent engineers. As its commercial reach grows, Score can bring new customer problems to that network rather than keeping every research challenge inside one company.
1
1
7
1,356