Sabitlenmiş Tweet
macrozack (τ, τ) 🚀
535 posts
macrozack (τ, τ) 🚀
@macrozack
Builder @bitstarterai Director @MacrocosmosAI Editor @taotimesdotai Convening Bittensor's @ExploitSummit 🧩 https://t.co/9k3wunNQwr 🚀
London Katılım Mayıs 2025
720 Takip Edilen1.5K Takipçiler
@KeithSingery With fingernails that shine like justice
And a voice that is dark like tinted glass
English
I want a founder who gets up early
I want a founder who stays up late
I want a founder with uninterrupted prosperity
Who uses a machete to cut through red tape
Bittensor $TAO
English
macrozack (τ, τ) 🚀 retweetledi
Bitstarter's next launch: Live in the Louvre 🏛️
x4 finalists pitching.
x4 judges on the panel.
x1 subnet for the winning team.
...and who decides? You: by pledging to the team you believe in.
🇫🇷 Join us June 3rd @proofoftalk
⚡ Livestream via X
English
macrozack (τ, τ) 🚀 retweetledi
☑️ Machine Learning support, activated
Our ML teams need to do more than win on Bittensor - they need to succeed beyond it.
revela.io can make that happen.
An elite ML developer ready to build production-grade AI architectures - and scale our teams to success.
English
Launching a start-up live in the Louvre, June 2-3 - together with Bittensor’s best
Proof of Talk@proofoftalk
Proof of Pitch, Bittensor Track - Live Crowdfunding with @bitstarterAI Subnet teams, this is your launchpad. For the first time, Proof of Talk and Bitstarter are bringing live Bittensor crowdfunding to the Louvre Palace. Selected teams will pitch on stage in front of some of the world’s leading investors, capital allocators, and Bittensor ecosystem leaders, with the opportunity to receive funding through TAO pledges live from the room and beyond. Jury lineup: @mogmachine, Founder & CEO, @taostats @mikecontango, Co-Founder & CEO, @generaltensor @JacobSteeves, Co-Founder, @opentensor @macrozack, Founder, @bitstarterAI For subnet teams preparing to move to mainnet, this is the ultimate stage: visibility, capital, ecosystem validation, and live community backing. June 2–3, Bittensor Track at Proof of Talk.
English
Bringing together Bittensor's builders across machine learning - and asking how we find more people like them
Publishing articles like this reminds me of how much talent there is in our network 👌
TAO Times ⚡️@taotimesdotai
There's more to Bittensor than AI. But for Bittensor to succeed, we need the best machine learning teams building right here on TAO. For our first Roundtable, we ask how to get the best AI start-ups on-chain - and on-track to the top. So how can we do it? Read on 🧵
English
@TheAlbWriter @elonmusk …it was a joke. But given his track record lately, I doubt @Jason would miss the opportunity
English
@macrozack @elonmusk I have liste to the whole podcast and there is zero mention of $TAO. Why are posting fake shit like this ?
what kind of man does this ?
English
Ah, @elonmusk, thanks for joining us. What took you so long? bittensor:native
Elon Musk@elonmusk
English
@CadeONeill @CadeONeill you may find this interesting
bitstarter@bitstarterAI
Every team, evaluated 💯 Our first Subnet Update is here Read the deck for April's portfolio perspective featuring @QuasarModels @djinn_gg @handshake_58 and @adtao_ppcrebel Monthly metrics, top achievements, and more - link below 🔽
English
Which $TAO Subnet deserves more attention?
Targon? lium? Affine? Score? Vanta? deprecated? iota? Gradients? Hippius?
English
macrozack (τ, τ) 🚀 retweetledi
Every team, evaluated 💯 Our first Subnet Update is here
Read the deck for April's portfolio perspective featuring @QuasarModels @djinn_gg @handshake_58 and @adtao_ppcrebel
Monthly metrics, top achievements, and more - link below 🔽
English
macrozack (τ, τ) 🚀 retweetledi
Chris Zacharia @macrozack is the founder of Bitstarter @bitstarterAI, Bittensor’s first decentralised crowdfunding platform, built to help new subnets connect with the right capital, people, and ecosystem support. This summer, Chris will be showcasing Bitstarter's accelerator by launching a new subnet team live in the Louvre.
Before Bitstarter, Chris was part of the founding team at @MacrocosmosAI. Now he is building one of the key platforms helping subnet teams go from idea to launch, with the support they need to get off the ground properly.
At the Louvre, he will be bringing some of the strongest subnets in the ecosystem to pitch at Proof of Talk, another reflection of the work Bitstarter is doing behind the scenes to back the next generation of teams.
This summer Chris joins us at the Louvre.
June 2–3, Bittensor Track at Proof of Talk.
Use code TAO for 30% off tickets 👇
English
‘On Bittensor, we’re speed-running humanity -corporation, platform, federation, religion, nation state’
@const_reborn spoke at two of the world's top universities last week - and after attending both, I wrote an article sharing my take on Jake's perspective
TAO Times ⚡️@taotimesdotai
5 key takeaways from @const_reborn's talks at Imperial College London and University of Cambridge ⬇️ Full article linked below
English
We might be close to proving that @stitch3_ai actually scales.
We proved the model worked in the Bittensor ecosystem.
What we didn’t know was whether we could repeat it for new ecosystems.
Now miners are joining from our new prediction markets eco with increasing speed - 55 so far.
If it works in one new eco, no reason it can’t work in many. And there are MANY.
We built the tech to scale profitably.
Time to scale it is approaching.
English
macrozack (τ, τ) 🚀 retweetledi
Machine Learning on Bittensor. Incubated by Bitstarter.
Announced live on @TWiStartups featuring @alex, @Lons, @macrozack, and @mccrinbc, our new track is:
☑️ Backed by @const_reborn
☑️ Built for the best ML start-ups
☑️ Subnet registration, development, and infra included 👀
English
macrozack (τ, τ) 🚀 retweetledi
Just dropped the “Shifting Frontiers” talk from the Bittensor Breakout conference
🎥 $TAO
Chris Zacharia - @macrozack
Mark Jeffrey - @markjeffrey
Karl Osis - @ktotheohh
Chris Sorel - @AgenticToaster
English
macrozack (τ, τ) 🚀 retweetledi
Very interesting take from @macrozack at the end of this extract:
(paraphrasing) Networks are the most natural environment for artificial intelligence to develop, not corporations.
Just look at how intelligence emerged in nature, through survival of the fittest and natural selection, in a distributed system of predator and prey.
My take is that every intelligence we know of emerged from a network, and every future one will too.
Even the LLMs we use daily, built inside closed labs, only exist because they were trained on the internet, aka the largest open network humans have ever built.
So if AGI emerges the way every prior intelligence did, it won't come from a lab. It'll come from a network.
Bittensor is the first one built specifically for it.
This Week in Startups@twistartups
This is BitTensor’s biggest weakness 👇 Users (miners/validators) are incentivized to compete. This means gaming the system, exploiting every possible advantage. Subnets need to be built assuming people are gonna cut corners. Builders have to assume exploitation is the default. They know it’s gonna be hard to build, because, here, cheaters do in fact win. Welcome to building startups on hard mode.
English
@simioape @Alfarius77 @macrocrux @opentensor I agree that harnessing miner intelligence is optimal - but IOTA does support decentralization in another way, by lowering the barrier to entry
Covenant's 72B was trained <50 miners - effectively mini datacentres
IOTA opens access to consumer compute, scaling to thousands
English
@Alfarius77 @macrozack @macrocrux @opentensor something that leverages miner intelligence and that uses bittensor. IOTA doesn't.
They run a server that tells validator what weights to set (no real yuma consensus) and they distribute rewards manually. You could literally do this with any other token.
English
Macrocosmos resuming the lead for distributed training on Bittensor
High talent, high conviction team led by @macrocrux, formerly CTO of @opentensor itself
If you're looking to back decentralized model training, this is our very best shot
Macrocosmos@MacrocosmosAI
Today, we present ResBM (arxiv.org/pdf/2604.11947), a 128x activation compression technique for achieving SOTA training results in low-bandwidth, distributed communication settings for pipeline parallel training across the internet. This technology underpins @IOTA_SN9 - our distributed training platform.
English
@simioape @macrocrux @opentensor OK - but innovations like ResBM are what allow IOTA to conduct decentralized pretraining on a global scale, with thousands of participants
English
@macrozack @macrocrux @opentensor this is not decentralized, there is no MoE. Macrocosmos just rewards compute provided but all innovation remains on their side
let me know if this is wrong
The fact they send rewards to a wallet controlled by them to then distribute rewards manually is a story for another day
English
@twistartups Might want to contact this guy
x.com/kodjima33/stat…
Nik Shevchenko@kodjima33
Spent 4 months and built Omi for Desktop, your life architect It sees your screen, hears your conversations and tells you what to do next It’s like having a second brain that actually pays attention Open source, local, link below
English
We're offering a 5 THOUSAND dollar bounty to the best app that can give us real-time insights as we record the pod. 💰
Think you have what it takes?
Learn more: 🧵
English
Wake the world's sleeping compute.
Look at the Mac nearest to you. What's it doing?
Probably nothing.
There are 100M+ Macs with Apple Silicon out there. Apple quietly made them *really* good at inference. A $3k Mac runs a 60B model at 30 watts.
Most sit idle most of the day.
Meanwhile every AI API call passes through three layers of margin before reaching the hardware. We call this the Inference Tax.
We got curious: what happens if you connect idle Macs directly to inference demand?
This is Darkbloom. Private inference network for idle Macs.
darkbloom [dot] dev -- paper + code open.
Reply for invite + free credits ↓
English