Distributed State
@DistStateAndMe
Followers
2K
Following
24K
Media
517
Statuses
7K
cult leader at https://t.co/rH6EoYxwlc || Bittensor Maxi
Joined April 2014
Const had to remind me never sell out. For 4 months , templar has , and will continue to do something that has never been done before. Permissionless, Incentivised Pretraining . LFG.
@const_reborn on why tplr.
4
2
62
I dont think its arrogance when we say you can build any AI start up on @opentensor . Our Incentive System , Yuma Rao is expressive enough to codify almost any system. Bittensor will power the world's super cluster. Yuma Rao subsumes all . #TAO.
12 hours ago @EMostaque asked me for crypto incentives which exchanged TAO for H100s. 4pm today. Done.
4
16
136
OPPPPEEEEENNNTTTTTENSSSSSOOOOOORRRRRRRRRRRRRRR πππππππππππππππππππ. @opentensor.
5
6
118
4
4
116
I am looking for Senior DeFi Engineers to come work with us at @akropolisio . We have got so many interesting things lined up and are looking for the right people to help execute. RTs appreciated!.
9
110
104
The date is 20/06/2022 and the super intelligence is 1 day closer to taking my job.
7
4
87
No one asked for it , so we built it. Starting a port of @eigenlayer to rust , starting with the sdk. Open to contributions.
10
4
89
Alpha leak: I am hiring shadowy super coders that are interesting in building on StarkNet. DMs are open.
My DMs are blowing up. Exact same convo every single time:.Dev: How can I get in on the STARK action?.Me: Go build on #StarkNet today. Start on the StarkNet discord:
4
14
81
Last week sucked balls. Slept about 10 hours combined. But the call today made it all worth it and more. We are the anti-fragile hydra , and with every cut we come back stronger. #TAO @opentensor .
5
7
80
1/7 Big news! The @opentensor team has released a key update to the protocol that solves the critical issue of weight copying. ππ #TAO.
1
7
80
Bittensor subnets are the garden of Eden. Each one a bold attempt at channelling incentives to produce super commodities .
2
6
76
90 subnets. 90 Narrative Violations. SUBNET SUMMER LOADING. LFG . ππ
2
9
66
Bittensor is proof that normal people can just do things . Invictus: βI am the master of my fate. Commander of my soul β . This message made my week β€οΈ
10
8
67
Why isnβt there a movie about βThe DAOβ yet? Tag line: βBlockchains darkest momentβ @netflix get at me I have your script!.
18
1
54
On Abracadabra Degenbox Strategies, dwindling Anchor Protocol Yield Reserves, and risk to Curve liquidity providers
2
14
58
Miners realising that there is a new guy outcompeting them. This is why we bittensor. Incentive Alignment ==> Hyper competition ==> TPS go brrrrrr. Beautiful. @manifoldlabs @0xcarro @0ximjosh #targon
1
9
55
What does number go up mean for #TAO? . Yuma consensus incentivises hyper competition. This hopefully draws new players into the ecosystem which will battle test each subnet's incentive mechanism , and reinforce the virtuous cycle of miners exploiting different mechanisms,.
4
6
55
10K Tao is fud.
0
3
55
Let's remember how ahead of their time @Bancor was. I'd admit I thought the idea was stupid initially. You can thank them for DeFi.
4
6
48
Today 64 subnet get the chance to walk the path of Yuma Rao. DTAO. Forward to Infinity!!!
0
1
52
That's my bull !!!. ". Bittensor will quickly leapfrog all centralised open source AI" . π.
1
5
41
Bearish.
2
3
43
Seriously the bitmind guys are killing it. Bittensor is a talent vortex . If you are relentless , step into the arena.
Over the weekend @BitMindAI created the best developer environment to train, deploy, and manage miners on #Bittensor. If you're an AI Developer, do you think you could be one of the top 15 developers on earning > $2k / day? . Try it out:
1
4
43
Spoiler alert : this is why the foundation never deployed STAO and is yeeting straight to DTAO. While DTAO doesnβt eliminate plutocracy , itβs steers it closer to an ideal state ; market based capital allocation. STAO is still cronyism .
1
2
42
Definitely picked up an additional mental illness or two building @tplr_ai . Would do it all over again in heartbeat. Bittensor is a narrative disruption that breeds narrative disruptions. LFG ππππ.
The most interesting thing about Bittensor is that if you look closely, there are tiny teams are doing the same thing as well-funded, much larger VC startups. and even outperforming them. Running on pure adrenaline + passion and some mental illness.
1
1
41
Pretty exciting to see independent community initiative start to flourish. A month ago this was an idea. Today we have subnet onwers doing a Q & A with the @tao_minersunion . May a thousand gardens flourish! #FuckMiners
1
8
37
βLouder for the people at the backβ. The inflow of Google talent into @opentensor has being unreal. web3/acc is only just getting started.
a few years ago i wouldβve killed to work at google, netflix, vercel, or apple. now those places seem dull and uninteresting and there are dozens of blockchain teams iβd rather work with.
0
5
35
This is madness. Github Autopilot is actually autocompleting @CairoLang code. Definitely going to be out of a job soon π°. It doesnt seem to think its @rustlang but not too far off
3
5
37
Jon Durbin is the best of us . Excited to be helping with this.
I'm working with the legendary @DistStateAndMe to soon convert my share of chutes into a what is essentially a non-owned shared entity with multi-sig via bittensor EVM smart contract. I will be giving up all ownership claim of any of the accrued alpha, the funds can only be.
3
3
37
Previously I tweeted this ELI 5 of #YumaRao , the subjective utlitity consensus powering @opentensor . Today I delve a bit deeper for the more adventurous.
So we are going to start with an ELI5 version . 1/10 Imagine a game where you and your friends vote on how many candies each person gets. The number of candies you get depends on the votes you receive. π¬π.
2
7
32
"The convergence is strong here" - Master Yoda
Only Bittensor peeps understand how hard incentives are, and how crazy it is to corral miners into a smooth and stable loss curve. Incentives are to computer science as fusion reactions are to physics β at first they blow up in your face, and then they click and unlock infinite.
0
5
37
To the guy about to scam me $700 dollars on a token2024 ticket. I just lost 5k on a meme coin this morning. You can't hurt me. What is dead can never die.
11
2
30
Akropolis is going to be pumping out bee products over the next couple of months. 1/n
10
4
33
Super stoked to begin this new join with some amazing people. Canβt wait to start rolling in on the milestones!!!.
6
1
32
"Babe wake up , @const_reborn just dropped π₯π₯π₯".
@WayneVaughan It is basically correct. We built the API and consensus algorithm to be as general as possibleβnecessary for the first digital commodity we were evaluating which focused on miners producing information-theoretic value. Pure information is so fuzzy and high-dimensional that we.
0
4
33
Bear market got you feeling down? That meme coin down 99% ? . Then fear not. I'm hiring! . No agencies at this point. Shares would be appreciated.
4
14
29
1
1
29
βShow me the incentives and Iβll show you the outcomesβ.
focus on incentive design, nothing else.
0
2
32
DTAO , @latentholdings , @CrucibleLabs , tao . com. For AGI will be built on antifragile foundations , and the process is accelerating .
"The purpose of power is to give it away".
2
1
31
Resources should not be a barrier to talent . If you are GPU poor , and want to take part in the only permissonless training run , reach out . Will pay for your GPUs . @tplr_ai.
Places for the GPU poor to work in AI (after asking a bunch of people at ICLR).1. Evals / datasets / benchmarks.2. Trying things on small models and then trying to scale them up.3. Interp / explainibility / robustness.
1
2
30
People will soon realise how DTAO smashes this power dynamic to bits. It has to evolve from "my subnet" to "our subnet" for it to survive. That "we've got revenue" narrative is silicon valley ponzinomics. You make 5 bucks , so obviously you command at 15 B valuation.It takes.
Did you know that a 5% total bittensor emissions subnet lost its DB, and users need to fill out a form to get refunded? π€‘. It's not about the tech but who owns the subnet.
0
0
30
TemplarNovaaaaa.
also s/o to @DistStateAndMe for supporting @metanova_labs he's one of the hardest working people i've ever met, building an ambitious and technically impressive project, and somehow still finds time to help.
2
3
29
We cannot have permissionless training without incentives, and the only way to really test incentives is in production. From day 1, @tplr_ai has been permissionless, iterating on this problem. For as long as it's a VPN, training might be distributed, but is not decentralized.
Templar (Bittensor subnet 3: @tplr_ai) is the only live decentralized training network in the world . permissionless and incentivized. And raised zero VC. Meanwhile > $100M has been invested in other βdecentralizedβ startups experimenting with algos and white papers β with no.
2
7
30
Study @const_reborn / @shibshib89.
Study L1 founders abandoning their project once they get the bag. Study Vitalik still being heavily involved in Ethereum 10+ years on from when he first published the whitepaper. Built completely different.
0
4
29
Look ma. Iβm famous.
#VISION2022 . "#Ethereum's L2 will continue to mature, offering increased efficiency and cost reductions. This will usher in a Cambrian explosion of "#DeFi native" structured products offering sustainable, risk-adjusted yields.β. - Samuel Dare @DistStateAndMe, CTO of @akropolisio
8
1
28
Being in the space for about 8 years and nothing blows my mind as much as what @khalani_network is building . Obviously biased . Position yourself accordingly.
0
4
26
Correction: @opentensor cuts across all these categories and then some. More on this soon. Spoiler alert: Each of these starts ups could be a subnet.
With @nvidia earnings coming in hot yesterday, it is clear that demand for compute is higher than ever. Attention is turning towards AI x Crypto projects, ushering in a new realm of incentivisation for AI. Here's an overview of the current landscape π§΅
1
5
29
Seth writes. I share without reading it. No regrets.
3
0
28
With @OpenZeppelin and @HardhatHQ building tooling for Starknet itβs obvious that StarkWare is committed to developer experience. Canβt wait to try this out!.
The powerful @shard_labs team is building a @HardhatHQ plugin for #StarkNet. Another small step in making sure developers have access to the tools they know and love!.
1
3
26
"All those uids belong to us".
3
1
26
Miner is GOD.
$TAO Bittensor devs are among the smartest I've ever seen.
1
1
26
βBe the Tao Maxi you want to see in the world β - Confucius / Yuma Rao. Shout out to @LukitaTao and TM merch.
6
0
27
@maclusters @const_reborn @Crypto_McKenna add me on discord and i will pay for an H100 core for you. discord name: formalised_tensor. Be part of history.
3
2
27
@ow @Steven_McKie This is something we should be really scared of and not celebrate. He's a single point of failure and we have to hope he doesnt die or become evil.
3
0
23
Today's Novelty Search was fire. It reminded me of the good times , two capable teams pitching the pros and cons of their pretraining approaches. @Coldint_ (sn 29) , @MacrocosmosAI (sn 9), you are both winners in my books!.
0
2
25
5/6 Huge shoutout to the brilliant research and engineering teams at @OpenTensor who worked around the clock to analyze this problem and come up with a solution, and the feedback from the community. Your hard work and dedication are moving Bittensor forward! #YumaRao is nothing.
4/7 We ran simulations using real subnet data, and the results are clear: copiers' rewards are significantly reduced. In most cases, copying is no longer more profitable than honest validation or nominating one's stake. ππΈ.
0
2
25
So when @tplr_ai finishes pretraining , and needs. what would we do ? π€. hold my beer πΊπ».
Gradients 56 now supports endpoints where you can one-click a query like this and an incentivized p2p network of miners will compete to instruction finetune an LLM of your choice.
0
0
26
Somebody tell him that every subnet on Bittensor is a fair launch now.
Listened to the coffeezilla interview. Pretty bizarre. Also incredible how all blockchains (solanaβs turn now) are βrediscoveringβ fair launches lol.
0
1
24
@lior_eth @Ar_Douillard @gensynai @PluralisHQ @PrimeIntellect @NousResearch FYI. @tplr_ai already has permisionless, incentivised training runs today. This is our training current run. Over 100 miners , who dont care about the loss curve , but care about the incentives. We used the incentives to drive down loss reduction. 0 guard rails.
2
0
24
To set the record straight. Yuma 3 is massive , and a net positive for horde. FOR THE HORDEEEEEEE.
2
2
24
The Optimistic Roll Up Fuel.sol is 2k+ Lines of Code π€―π€―. While this might not be correlated with how good it is . hats off to @IAmNickDodson / @fuellabs_ for the gargantuan effort !.
2
5
22
"We would like to extend our gratitude to @OpenAI, @GoogleDeepMind, @AIatMeta. We exist because of your policies, and appreciate your attempts at pseudo-open source. We will double down on our mission of building the eternal AGI federation.". - Signed @opentensor , @tplr_ai
Les Schwab on why @opentensor / bittensor matters. #TAO.
0
3
24
TAO consumes all.
So we've been claiming most projects out there could work as a subnet within $tao just to get to a point where 3 months old subnets have achieved much more than 5 years old projects, worth close to 1bn, and we have to hold back from comparing them?.
0
2
24
. people realised that DTAO == fair launch. every subnet gets their eth, bitcoin. BITTENSOR moment?. LFG :rocket:.
What if.
1
4
23
It was fun auditioning for @brodydotai βs role on taotalk. Thanks for having me !.
Thanks again for joining me this week @DistStateAndMe! Great vibes.
1
1
23
The whole stack or nothing.
In the end, the goal is ***decentralized*** AI. That means interlinked commodity systems where we mine each otherβs protocols instead of relying on centralized infra at any layer. Bittensor will internally mine (build in-house storage, compute, and inference) and externally.
0
4
23
Had the pleasure of meeting the @metanova_labs team , and instantly feel I love with their passion and deep expertise. This is the calibre of team we want to draw into the ecosystem. Pretty excited to see their subnet unfold!.
bittensorβs incentive design prioritizes long-term growth, value creation, and resilience. value-add exchange > ponzis. post dTAO, itβs the network best positioned to promote PPP plays. join us, if you havenβt already.
2
2
20
And to think I "farmed" this just by using my browser? Download the Brave browser to enjoy 0 gas yield faming @brave
3
2
21
What is it about these ex-Celo guys that makes them world class buildersπ€.
OP rollups with built-in interoperability π΄β©. @0xAlex_ built an epic OP stack mod that includes Hyperlane out-of-the-box. Developers can now launch rollups with interop on day 1. Learn more here β¬.
1
1
21
Templar 70B cooking.
βOpen Source AIβ.
2
2
22
We must protect him β₯οΈ.
I'm working with the legendary @DistStateAndMe to soon convert my share of chutes into a what is essentially a non-owned shared entity with multi-sig via bittensor EVM smart contract. I will be giving up all ownership claim of any of the accrued alpha, the funds can only be.
2
2
22
@jbrukh @AnthropicAI @jackclarkSF At @tplr_ai.( , we conduct permissionless , incentivised training runs on mainnet. This is a modest 1.2B , and we plan on bumping it soon. We are about 25% through this run , and the downstream metrics look good .
1
3
21
"Everything is a subnet" - @const_reborn.
@EMostaque @DistStateAndMe @Awaki18 @SchellingAI if the crypto application that requires a token is incentive for compute provision, that part can be a bittensor subnet.
0
0
20
βEverything is a subnetβ - @const_reborn.
Been looking at new #AI infra decks. Think rather waiting for their growth of ecosystem from scratch, why not invest in $TAO . a) 38 teams building on top already .b) full liquidity without vesting period. Donβt get distracted investing in sth cheap, invest bc of its greatness.
1
1
21
We're still so early. You don't have a chief java officer do you? We when βblockchainβ fades into the background. Not because it's irrelevant (or a terrible like Java π) because it's ingrained into the fabric of every business
3
1
21
Bittensor = Collective Intellegence.
Ethereum = Verifiable World Computer .ETH = World Computer Money .ZKsync = Ethereum's Elastic Compute Cloud.Fermah = Ethereum's Prover Engine.
0
0
19
The garden is growing.
β
Registered subnetwork with netuid: 52.
0
0
20
But. you've invested in Ethereum based companies?π§.
Ethereum is a technological dead end. The more itβs used, the faster it dies. Fortunately, USDt is also available on the #LiquidNetwork which is more scalable and later will allow Lightning Networks to be created for assets like Tether. πβ‘οΈ.
1
1
19
No longer will chains be subjected to the Tyranny of liquidity bridges , operators or validator sets. Khalani is here is provide permissionless liquidity to any chain.
0
4
17
