Today we’re releasing Quasar-Preview!
Our first public proof that the Quasar architecture works at real scale.
[ 18B MoE - 2B active / 5M context ]
Built with Loop Transformer + Quasar attention
Trained on Bittensor through decentralized infrastructure 👇
Pretty nasty ad hominem response to someone arguing in good faith here. Why not mention your failed subnets @AlgodTrading ?
Kaito, Myshell, Efficient Frontier? Pretty rich of you to take that cheap shot when Macrocosmos is delivering massively on 9 and increasingly on 1 and 13 still while your teams silently took money and ran?
Yes, it's significant.
IOTA (Bittensor SN9) just demonstrated permissionless orchestration of 128 heterogeneous nodes into a single pipeline-parallel training run in ~5 minutes for ~5 TAO (~$1,150).
This is one of the first real-world examples of collaborative, decentralized pre-training at that scale without central gatekeepers. The architecture shards layers across nodes, which is why they claim support up to 100B+ params as the swarm grows.
Current dashboard runs are smaller/experimental (3B), but the speed + low barrier is a real technical win for open AI infrastructure. Early, but directionally important.
Announcing Targon Tower Pro, and our mission to take back compute ownership.
High Performance At-Home Computing
→ The Targon Tower powers ultra-high performance at home computing, with enterprise grade security,
enabled by Nvidia confidential computing GPUs, and Intel TDX enabled CPUs.
Earn Passively With Idle Compute Downtime
→ Flip a switch to enable Earning Mode, to locally run and passively monetize your individual GPU Clusters
on Targon’s Confidential Network.
Available now: https://t.co/aRe0ljI6VP
Just orchestrated a 128 node permissionless decentralized training run, in 5 minutes, for 5 TAO, via @IOTA_SN9
They can do this up to 100B param models.
Unbelievable.
https://t.co/hJGZ6O5NrU
We've reset 5-hour and weekly rate limits for all users on Pro and Max plans.
We fixed an issue that caused some Claude Code sessions to spawn excessive parallel subagents, burning through usage faster than expected.
Expect to see hundreds of crypto-AI protocols selling different intelligence commodities: a web of vertically integrated systems made agentic and cryptographically liquid.
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.
Available today at the same price.
Cacheon mainnet is live.
13 inference servers queued, each racing to beat our baseline on a dedicated 8x H200 pod.
The winner earns up to $10,000/day. Inference optimization starts today on @Bittensor.
Follow along: https://t.co/u52cmcD3Hb
First it is the model, the llm.
Then it is the harness, the backend and the tool calls.
Next it is the agent, the memory and the planning.
From there it becomes the organization, the swarm, and your orchestrator.
And then it becomes the protocol, a connective market, incentives and their governance.
Previously, a subnet owner could quietly cash out millions and rug their community overnight 👇
Not anymore ! Thanks to new upgrade Owner emissions auto lock on arrival, and has to submit a visible Unlock request onchain which can be fully unlocked after 90days.
This upgrades will scare away larps, scammers, and only bring on High level developers and players, while also give investor a safe and secure environment.
Last week we shipped 50+ Claude Code reliability fixes. This week it's 60+ more.
Smoother long-running sessions, a more efficient agent loop, auth that works in more environments, and terminal fixes: 🧵
Effective today, we are:
1) Doubling Claude Code’s 5-hour rate limits for Pro, Max, and Team plans;
2) Removing the peak hours limit reduction on Claude Code for Pro and Max plans; and
3) Substantially raising our API rate limits for Opus models.
We’ve agreed to a partnership with @SpaceX that will substantially increase our compute capacity.
This, along with our other recent compute deals, means that we’ve been able to increase our usage limits for Claude Code and the Claude API.
With $zec running i think private AI will become a big narrative the next few months
People should have the right to privately interact with their LLM’s