Want to follow Quasar more closely?
Join our Telegram community:
✅ Early updates
✅ Direct access to the founders
✅ Ask questions and join the discussion
Come build with us ↓
https://t.co/6uZx8q0iLy
If you want to understand why Endure, listen to this podcast!
You're going to understand why I am so bullish.
@gordonfrayne and the Endure team go through how big risk actually is and they also discuss Forge, the first native money market on $TAO.
My big takeaways:
🔸 Within a month Forge is already live on testnet, all risk parameters will be set by Endure to make it safe and as profitable as possible.
🔸 More then 50% of revenue of Forge is going to be used to buyback and burn Endure
🔸 They are solving a real problem for Bittensor and thats shows as they have some very good conversations with Subnet Founders and institutions 👀
🔸 Forge is only just the first product and they named that they already have great design partners for the next.
🚨Quasar Block Explanation 🚨
Quasar emissions were blocked on Bittensor, but that is actually based on a misunderstanding.
The blocking is meant for subnets that are burning emissions. The issue with Quasar’s incentive system is that it points to 100% burn when there is no winner, then goes back to 0% burn when there is a winner so the winner can receive their emissions.
In the last 24 hours, we had our winner, and they already received their emissions. For the past two months, mining has been running with the same incentive system, which may look on-chain like we had 100% burn for long periods, but that is not what was actually happening.
SN24 paid more than 200 TAO to miners in the past month, running every day with a new winner each day.
To avoid any issues, we have now made the burn stay at 0% in both cases, and we should get our emissions back.
This is simply a misunderstanding!
Subnet 56 @gradients_ai joins Subnets 24 and 3 in training Quasar models.
Three subnets, one training mission, pushing decentralized AI closer to SOTA !
Gradients will play a key role in Quasar’s post-training and RL, helping turn the base model into something that can truly chat, code, reason, and become an overall SOTA model.
The open, decentralized AI lab is taking shape… 📷
🚨 BIG NEWS: Root Reborn
#2759 dropped on Github.
Simply put: $TAO's Root Reborn changes root staking from a Sell Machine into a Reinvestment Machine
Right now, root staking earns yield by taking subnet dividends and automatically selling them back into $TAO.
That means every block, root yield, creates sell pressure on the very subnet tokens that are supposed to give $TAO value.
So Root Reborn changes that.
Instead of dumping subnet alpha into $TAO, validators would choose where that root yield gets reinvested across subnets.
So the flow changes from:
Subnet Dividends = Auto-Sold into $TAO to Subnet Dividends, Reinvested Into Subnet Baskets, which Compounds Over Time.
This could change everything.
It reduces automatic sell pressure on subnets.
It creates more buy pressure for selected subnets.
It lets root yield compound instead of leaking out.
It makes validators more important again because they actively curate where capital goes.
It makes $TAO’s Risk-Free Rate cleaner because the yield is backed by reinvestment, not constant dumping.
It could also make root staking more attractive, which brings more capital to root.
That capital then gets deployed back into subnets, which is bullish for some subnets.
More capital attracted. That is the idea, I think, and we need it.
The PR is still open, and the reviews flagged real technical risks that need to be fixed before this should go live.
The concept is powerful.
But the implementation still needs safeguards around things like unbounded work, migration safety, and making sure claims are paid to the right stakers.
And before many ask:
What stops bad validators from sending root capital to weak subnets or insider baskets?
Simple their stakers. If a validator allocates badly, their basket performs badly.
Stakers can see ALL that.
It doesn't remove bad actors. It just makes it visible, measurable, and not worth the cost.
$TAO people should be payinging attention here.
🔗https://t.co/aAdUs26pyE
Quasar is trending on Hugging Face.
For context, HF is the central hub for AI models. Every serious open-source model lives there.
We’re already on page two, alongside Xiaomi, Qwen, and Liquid AI and this is just from our small model experiment!
The ML community are taking notice. Not just on Bittensor but in the broader open-source world.
The open-source community already built MLX and GGUF versions so Quasar can run on MacBooks and local AI setups.
We are just beginning.
Quasar is trending on Hugging Face.
For context, HF is the central hub for AI models. Every serious open-source model lives there.
We’re already on page two, alongside Xiaomi, Qwen, and Liquid AI and this is just from our small model experiment!
The ML community are taking notice. Not just on Bittensor but in the broader open-source world.
The open-source community already built MLX and GGUF versions so Quasar can run on MacBooks and local AI setups.
We are just beginning.