Decentralized Intelligence | Bittensor ΟΟ | Privacy β’ Freedom β’ Community .β‘β‘
U.K
Joined February 2021
- Tweets5.3K
- Following356
- Followers505
- Likes17.8K
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
Tomorrow...
In the meantime, if you understand what this message is about, it might be time to pull your TAO out of these subnets. Something smells seriously fishy, to put it mildly.
eTAO is exposing bad actors while giving the best builders a new way to demonstrate their conviction.
It's never too late to join the winning side and leave the dark side behind.
Looking forward to learning more about SN35, SN36, SN70, and SN103. πΏ
SN103 Consted, SN70 Consted 2.0, SN35, SN36 Epago: four subnets pushed their owner cut to the 33% max within 12 min tonight. On-chain, 103/70/35 pay each other and share a treasury; 36 moves in lockstep. Same team? @const_reborn, Root Reborn shouldn't feed slot farmers.
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
One thing is sure for humans. You cannot please 100% of the people 100% of the time. Many subnets have suffered tremendously from the sudden changes in how subnets work, and some have adapted and were lucky enough to be in a position to be able to adjust and benefit from them. I am sure alot of subnet owners have bad blood because of that, but they should feel lucky that they were able to extract Tao for as long as they did without producing anything of value. The changes are moving further away from ease of extraction and refining the emission rewards into high level competition. Humans don't like change, we love comfort and predictability. Bittensor is not that. It is a novel Network never before been done. It's transforming, mutating and evolving. Just the way it is. Many want the full decentralization to be on their time line, but its not that easy. So. If it comes down to one thing it is this. Do you believe that @const holds his original vision and mission as sacred or is he led by ego and extraction? I choose the former. He has much more to gain by seeing through his original philosophy than temporarily extracting after many many years of full dedication and building what it has become today. If your own greed gets in the way of making this choice then maybe being in $TAO is not for you.
FREE ALEX - A Question of Principles, Not Personalities.
Whether you like @DrocksAlex2 or not is beside the point.
The issue goes much deeper.
How can a network that champions decentralization, openness, and inclusivity justify excluding members of its own community simply because they disagree with certain decisions?
Since when should questioning decisions become a problem?
Bittensor claims to be building a system where the best ideas and merit should prevail. But how can we achieve that if critical voices are no longer welcome in the conversation?
Null Consensus in plain English
Yuma Consensus works like a committee. Every validator votes on who the best miners are, the maths finds a middle ground, and that determines pay. It caps subnets at 256 miners and uses about as much chain compute as you'd expect from multi-party consensus.
Null is a new OPTION subnet owners can switch to. It replaces the committee with whoever holds the most stake on that subnet. They decide alone β which miners get paid and how much. No voting. No consensus maths. No bond calculations.
Yuma stays the default on every subnet unless the owner opts in.
What changes
.Subnets can grow to 2,500 UIDs instead of 256.
.Weights get used exactly as submitted β no normalisation, no clipping, no rounding.
. Dividends flow proportionally to everyone with positive stake, not just active validators. You can also let people register by doing proof-of-work instead of paying TAO.
The upside
.2,500 miners per subnet is the headline. That's the scale Bittensor was supposed to reach but couldn't under Yuma's ceiling.
.Every epoch costs less to execute no consensus loop, no bond storage.
.Weight ratios are precise: if a validator says 50/50, miners get 50/50.
.Inactive wallets still earn dividends, which they don't on Yuma.
.PoW registration means joining without buying TAO first useful for onboarding small operators.
The downside
The top-staker controls the subnet. No checks, no second opinion. They can zero every miner's rewards in one epoch if they want to. It's plutocratic: the richest wallet wins, not the best operator. The subnet owner has no special override a whale can out-stake them and take over their own project.
Switching back to Yuma requires manually pruning from 2,500 UIDs down to 256. Multi-transaction work, not a toggle. A well-funded attacker only needs to out-stake the current leader to capture miner pay. On a small subnet that's cheap.
One more thing.
At 4,000 UIDs the epoch ran for ~7 seconds in testing. That's over the 4-second block limit. The 2,500 cap is the safe zone, but the ceiling is closer than on Yuma.
Summary:
Null trades safety for scale. Ten times the capacity, simpler chain economics, no consensus overhead. The cost is replacing distributed decision-making with whoever holds the most tokens. Makes sense for high-volume subnets with concentrated, trusted stake. Not for subnets that need decentralised governance.
Spec 474. Not on Finney yet. Subnet owners choose per-subnet.
As always DYOR
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
FREE ALEX - A Question of Principles, Not Personalities.
Whether you like @DrocksAlex2 or not is beside the point.
The issue goes much deeper.
How can a network that champions decentralization, openness, and inclusivity justify excluding members of its own community simply because they disagree with certain decisions?
Since when should questioning decisions become a problem?
Bittensor claims to be building a system where the best ideas and merit should prevail. But how can we achieve that if critical voices are no longer welcome in the conversation?
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
From the first stars to Bitcoin to an open network for intelligence, in under four minutes.
bittensor.page
Best with sound on.
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
π’ Last call subtensor : archive.minersunion.ai will be turned off in favor of archive.kusanagi.ventures.
Make the switch.
π
We're excited to announce a new pilot with @LeadpoetAI to accelerate Beam's commercial adoption.
As Beam expands its distributed data movement network across cloud, enterprise storage, and AI infrastructure, we're focused on connecting with organizations that can benefit from faster, more efficient data transfers.
LeadPoet will help us identify real-world demand, connect with the right decision-makers, and turn opportunities into meaningful customer relationships.
The technology is built. Now it's about putting Beam to work.
@b1m_ai π€ @LeadpoetAI
Bravoπππππ
Three days ago, this book didn't exist.
Today, it's a #1 Best Seller on Amazon France.
Bittensor explained to complete beginners. No code. No jargon. No price predictions. A city, its districts, its craftsmen, its jury. And a full chapter on what can go wrong.
Thank you to everyone who bought it, shared it, gifted it. You did this in three days.
Out in English and French.
Link in the replies.
A must read from one of our finest π₯π₯π₯
Let me show you around a city.
Not a city of offices. A city of workshops, open day and night.
Every district has its trade.
Here, they rent out computing power by the hour, like an Airbnb for AI.
There, they run AI models for apps all over the world.
Further on, tools compete to spot mutations in DNA.
In each district, craftsmen from all over the world compete with their machines.
A jury scores what comes out of the workshops.
The best get paid the most.
No company runs this city. Its rules are written in a public ledger anyone can read.
Its currency, TAO, is limited to 21 million units, like bitcoin.
This city exists. It's called Bittensor.
Why it fascinates me:
AI is already changing work, school, medicine.
Today, the most powerful AI belongs to a handful of companies.
Bittensor asks a simple question: what if AI were built by everyone, and belonged to no one?
It can fail. It's young, it changes its rules, its foundation still holds the keys, and it has real critics.
But the idea deserves to be understood by everyone.
I'm not an engineer. I got it. So can you.
So I wrote a book for exactly that.
For your mom. Your neighbor. You.
The Bitcoin of AI. Out today.
ππ¦π―π΄π°π³πππ°πΈ ΟΟ retweeted
Last week I told you I passed Bitcoin by ten times in 2015.
So I wrote the book I wish someone had handed me back then. No jargon. No price targets. No platform to sign up for. How Bittensor works, why it might matter, and how it can fail.
The Bitcoin of AI. Tomorrow.
In 2015, Bitcoin was right there, in front of me.
I ran into it ten times. I never stopped. Like everyone else, I shrugged, and I let it go by.
The hard part was not understanding Bitcoin.
The hard part was stopping for it. Giving it an hour, a page, one real try. The people who got it were not smarter. They had tried.
Eleven years later, I feel the exact same thing about Bittensor. A network where artificial intelligence belongs to no one, where the rules are public, and whose currency is capped like bitcoin. All around me, the same "this is too complicated for me."
This time, I stopped for it. And I wrote the book that lets you do the same in a weekend.
I wrote it for people who know nothing about any of this. Every technical word explained once, in one sentence. A city, its districts, its craftsmen and its jury to understand the machine without a single diagram. And the risks as a whole part of the book, not a footnote.
It will not tell you what to buy. It predicts nothing. It gives you what you need to judge for yourself.
"The Bitcoin of AI" is out October 12 on Amazon, in English and in French.
If someone around you says "I don't get any of this," I wrote it for them.
In 2015, nobody handed me this book.
In 2026, I am handing it to you.
Guess what ? Probably nothing π
We're partnering with @taodotcom to bring Alpha collateral to mainnet.
Forge will build on top of tao.com's wstAlpha wrappers. Building on existing, audited, and trusted infrastructure means starting from solid ground on day one.
Glad to have @taodotcom as a foundational part of Forge.
It's folks like @kusanagi_vntrs that make our community great $TAO .
Our new website is live.
Kusanagi is built to help Bittensor teams reach real customers, build stronger products, and push the network beyond its own ecosystem.
kusanagi.ventures/
Oh yes, here we goπ₯π₯
Endure is live on mainnet, and now you can check it yourself.
The new dashboard follows SN30 on mainnet: network state, miner commits, and consensus accuracy, each backed by a source you can verify.
Explore it: dashboard.endure.network
Risk, with receipts. Because risk intelligence should be something you can check, not something you have to trust.
$TAO SN30.
SN30 @EndureNet will take DeFi to the next level. π
DeFi lends dead assets. ETH, USDC, SOL sit in a contract until borrowed.
A subnets Ξ±lpha is live staked, it earns validator emissions every epoch. Working, not parked and this is just the cherry on top of the cake.
Here's why no EVM lender can copy this:
Bittensor's EVM ships a staking-v2 precompile (0xβ¦0805). A contract can hold a real stake position add, move, transfer. Forge's collateral contract wraps the chain's own staking flow.
bittensor.com/docs/guides/evβ¦
ββββ
Dead vs live collateral:
Normal lender: capital sits, earns 0%.
Forge: Ξ±lpha stays staked in the collateral contract, earns emissions every epoch, secures the TAO loan the whole time.
Observed subnet APYs: 30β48% vs 4β7% root.
docs.stakingrewards.com/stakβ¦
ββββ
Incentives no EVM money market can offer:
β«οΈ Borrower yield passthrough β keep Ξ± staked, earn emissions, and hold TAO to deploy. Two yields, one position.
β«οΈ Auto-compounding collateral β emissions restake into your collateral. Ratio improves on its own.
docs.forge.endure.network/
ββββ
...and the lender side:
β«οΈ Blended APY β TAO suppliers earn borrow interest plus a share of pooled subnet emissions.
β«οΈ Emission-subsidized rates β Forge buys down borrow costs from pooled staking yield, not just interest.
More subnets list Ξ± β more yield. Self-enforcing.
ββββ
Adoption flywheel:
Forge pays a new subnet's treasury in its own emission yield to list its Ξ±. The asset funds its own adoption.
The moat isn't the smart contract β it's that the collateral earns while locked. Whoever controls that yield owns a flywheel no EVM market can reach.
ββββ
Caveat, and this is a research post:π
v1 stage 1 is collateral-only with borrowing disabled β Ξ± is CF 0.25 / LT 0.35, and Ξ± rewards carry Ξ±/TAO FX risk. Design space, not a live product yet.
$TAO
DYOR.
github.com/endure-network/enβ¦
WOW β‘β‘
Most referral programs pay 3% of $0. Ours runs on proof of customers.
Send a developer to SayGM and earn up to 100% of what they spend in their first 30 days, paid in alpha.
It's a competition, not a commission: earn your share of the reward pool.