Laττice_Labs 🧪

The Halving represents a shift in scarcity. But also a shift in standards. As the world looks at our ecosystem, clarity becomes more and more important. I created this manifesto to translate that energy into motion. Enjoy. 👇🔊

@bandosei @ZssBecker

This is the most famous crypto youtuber: Last year he made a video promoting a bunch of AI coins promising fast 2000% returns So I checked to see how much you could've made if you actually bought $1,000 worth of each 🧵

Bittensor >> Bitstarter >> SN 24 Quasar Once again, I think it is important to highlight @macrozack and @const_reborn for their commitment to Bittensor. The distribution of SN24 @QuasarModels alphas is now going to be regularized, further proof that on Bittensor, there are no problems, only solutions. And since we are currently going through the network’s “conviction” era, mine remains fully intact regarding future launches on Bitstarter. Thank you both for the work accomplished. Here is the message shared by Chris in the Bitstarter group. "Quasar Distribution: Getting every pledger paid Bitstarter: last week, @const_reborn asked for patience in the SN24 distribution. During our conversations, he told me that the SN24 wouldn’t be in a position to pay for at least a couple more months. And while the owner wallet shows a healthy alpha balance, but @TroyQuasar and @Farahatyoussef0 need emissions to keep the subnet going. All of this is valid. But it doesn’t disguise the fact that many of you have waited since December to receive your full distribution. There’s good reason for that - and in that time, SN24 have shipped a hell of a lot, notwithstanding the fact that the subnet was exploited twice. But I can’t stand by and let you bear the burden. 💫 That’s why we, at Bitstarter, are paying the distribution in full ourselves. We’ve bought c.28,000 alpha on SN24, to distribute to pledgers from our Quasar crowdfund. Because Jake already stepped up - not once, but twice (plus buying out SN24 for the team in the first place). He shouldn’t be called upon yet again, given how much he’s already done. And the team need to double down on what they do best: building SOTA long-context LLMs on Bittensor. 💫 So, this weekend, we’ll be distributing the remaining alpha to every pledger - APY included. Right now, Quasar is an established team - a team that does Bittensor proud. But it wasn’t always like that. SN24 was our second crowdfund. Those who backed them were taking a chance on a young, unproven team, from a crowdfund platform yet to fully establish itself. They deserve to be rewarded, not resigned to the back of the queue. Just as Jake deserves credit for committing to Quasar, and Eyad & Youssef for all the effort in leading the project to success. 💫 To all those who pledged to Quasar: thank you for being patient, for believing in us, and for backing SN24. Collectively, we launched the subnet to success. And we’ll solve the distribution problem in the same way: together." @bitstarterAI bitstarter.ai t.me/bitstarterai

@coinfessions @DistStateAndMe

I don’t fucking care anymore 😂😂😂 I’ve max extracted all of you retards and you keep on following KOLs like me like y’all lack a brain. Thanks for playing.

Happy to announce our collaboration with @adaption_ai Adaption Labs is an AI research company focused on building adaptive intelligence systems . Through this partnership, Adaption Labs will provide SILX AI with state-of-the-art adaptive data to support the training of the Quasar foundation models. Their role will be to generate and refine high-quality, adaptive datasets at scale, enabling Quasar to continuously improve its reasoning and generalization capabilities. This collaboration strengthens Quasar path toward achieving SOTA performance and competing with leading closed-source models. The company is co-founded by Sara Hooker, former Vice President of Research at Cohere and a veteran researcher from Google DeepMind, alongside Sudip Roy. Adaption Labs has also raised $50M in seed funding to advance its mission in adaptive AI.

@TaoIsTheKey Don't forget @QuasarModels ☄️

If you’re in the $TAO ecosystem, these are the official subnet accounts you should be following RIGHT NOW. Many of them have under 1,000 followers (many under 500). There are hundreds of thousands of TAO homders. These subnets deserve tens of thousands of followers. This is how you support the ecosystem. Turn on 🛎 notifications so you never miss a product update, revenue milestone, or alpha drop from the subnets actually shipping on Bittensor. Here’s the PARTIAL verified list (the ones that have X accounts off the top of my mind. Please comment below if I missed any!): ReadyAI (SN33) – @ReadyAI_ Chutes (SN64) – @chutes_ai Affine (SN120) – @affine_io Targon (SN4) – @TargonCompute Ridges (SN62) – @ridges_ai Score (SN44) – @webuildscore Gradients (SN56) – @gradients_ai Hippius (SN75) – @hippius_subnet Metanova Labs (SN68) – @metanova_labs 404GEN (SN17) – @404gen_ BitMind (SN34) – @BitMindAI blockmachine (SN19) – @blockmachine_io Bitcast (SN93) – @Bitcast_network Synth (SN50) – @SynthdataCo VIDAIO (SN85) – @vidaio_ Alpha Core AI (SN66) – @alpha_core_ai Yanez AI (SN54) – @yanez__ai LeadPoet (SN71) – @LeadpoetAI Zeus (SN18) – @zeussubnet Sportstensor (SN41) – @sportstensor Quantum Compute (SN48) – @qBitTensorLabs BitAds (SN16) – @BitAds_AI Bitsec.ai (SN60) – @bitsecai Trishool (SN23) – @trishoolai VoidAI (SN106) – @v0idai ByteLeap (SN128) – @byteleap_ai Tatsu / ChipForge (SN84) – @TatsuEcosystem

@ShaneRyanKelly @nordin_eth x.com/const_reborn/s…

@nordin_eth its not the fact they left , its why they left and the fact they just showed a major , MAJOR flaw by design. this cant be fixed. this is a sinking ship.

If you're selling $TAO here, I don't know what to tell you. This will blow over faster than most think. Subnet 3 is still a thing, only the team's getting replaced (and who wants a team that can rug anyways?) Alpha holders have not lost a single token. Onwards🤝 $TAO #SN3

@covenant_ai What the actual fuck

@princeharry_za @tao_punks @cryptopunks Massiv extraction and brutal downfall ?

The rarest @tao_punks just sold for 10 $TAO. The cheapest of the 96 alien punks is listed for 3 $TAO. You know this is going to follow @cryptopunks right?

Hi everyone, There are a few important things we want to clarify, along with how we’re moving forward and how we plan to solve the current situation First, what happened is extremely unfortunate for everyone, and especially for us. We’ve spent day and night building Quasar with the goal of contributing something meaningful to Bittensor and making it something we’re all proud of What exactly happened is still unclear. The important part is that our devices were not compromised, which is a positive sign. Beyond that, we’re still investigating, but right now our focus is on moving forward rather than speculating. We sincerely apologize for anything we may have said in the heat of the moment. We ask for your understanding we’ve effectively lost our work twice under circumstances we don’t fully understand. We followed all known security practices, but clearly there’s more to learn. That said, we don’t want to turn this into a blame game. Our focus is the future of the subnet And we are still building Moving forward: The coldkey will be held by @const_reborn to ensure maximum security. We will focus on research and engineering. We will continue improving the subnet and its core systems. Everything we’re doing comes from a genuine effort to build something valuable within Bittensor and we will continue The subnet is still active. We will be making changes on mainnet, including a rebrand focused on distillation into the Quasar architecture

@TheTNetHunter @QuasarModels I personally enjoyed the dip 😁💸

Quasar was unlucky compromised but the @QuasarModels has no harm in this. All it is now is Const checking what it could have been, but for $dTAO it's an investment opportunity. #SN24 $TAO

@corgil @uhonyn Perhaps if you're seeing multiple simultaneous posts about a subnet, it's thanks to @Bitcast_network coordination; you should look into it 🫶😉

fair point on $HYPE concentration, not claiming it's perfect but $TAO KOL shilling is well-documented. search CT - coordinated posts, same talking points, same timing. that's not organic "fairly launched" mining still concentrates at the top. you said it yourself. difference is HYPE didn't pay influencers to hide it

For those unaware - 90%+ of $TAO shills are paid Coordinated cabal. Biggest KOLs on CT. All pushing simultaneously They dump freebies on your head, rinse, repeat Most centralised “decentralised” ecosystem in crypto Trade it - don’t marry your bags!

This is Quasar Attention, the mechanism behind the upcoming Quasar models, designed to support context lengths of up to 5 million tokens. Attention has long been a bottleneck for processing extended context. Standard attention mechanisms struggle to scale beyond ~200k tokens in training, creating a ceiling on how much information models can reliably use. One approach to solving this has been linear attention methods, such as gated delta attention (used in Qwen 3.5) or Kimi delta attention. These improve efficiency and allow longer sequences, but introduce trade-offs: instability at extreme lengths, quality degradation, and in practice, they are not strictly linear. Quasar Attention takes a different approach. It uses a continuous-time formulation, implemented as a fully matrix-based system rather than relying on vector-state approximations. In practice, this improves stability, reduces cost, and maintains performance as sequence length increases. In internal stress tests at 50 million tokens, KDA-based approaches begin to lose stability, while Quasar Attention remains stable. This allows performance to hold as sequence length increases, rather than degrading beyond a fixed threshold. On BABILong, a Quasar-based model pretrained on 20B tokens and fine-tuned on 16k sequences was evaluated on contexts ranging from 1 million to 10 million tokens, maintaining consistent performance across that range. By contrast, models using gated delta attention show significant degradation at longer lengths, in some cases dropping to ~10% performance at 10 million tokens. (Note: results are indicative; setups are not directly comparable) On RULER benchmarks, a Quasar-10B model (built on Qwen 3.5 with frozen base weights and Quasar Attention added), pretrained on 200B tokens, achieved 87% at 1 million tokens, outperforming significantly larger baselines, including Qwen3 80B, under the same evaluation conditions. Taken together, this points to a shift in where long-context performance is won or lost: not in model size alone, but in the attention mechanism itself. Quasar Attention represents a step change in long-context modelling, setting a new standard for stability and performance at scale. We thank @TargonCompute for the compute and for being our compute provider and long-term partner in training the upcoming Quasar models Here is the link to our paper 👇

@bohdanmotion insane work

This shouldn’t be possible But here we are

@MistralAI, sovereignty is core to your mission. Yet, this partnership with @nvidia relies on a centralized US model. Projects like @tplr_ai 's Covenant-72B prove that decentralized infrastructure is now viable at scale. Why not prioritize these paths for the future of open-source AI?

🚀Announcing a strategic partnership with NVIDIA to co-develop frontier open-source AI models, combining Mistral AI’s frontier model architecture and full-stack AI offering with NVIDIA’s leading compute infrastructure and development tools.

We just completed the largest decentralised LLM pre-training run in history: Covenant-72B. Permissionless, on Bittensor subnet 3. 72B parameters. ~1.1T tokens. Commodity internet. No centralized cluster. No whitelist. Anyone with GPUs could join or leave freely. 1/n

Bittensor $TAO SN62 Ridges It’s quite incredible to see a subnet that so many people seemed to appreciate end up in this kind of situation. On my side, I’ve always stayed on the sidelines, mainly for one simple reason: I promised myself I would never FOMO in, especially when everyone is jumping on green candles. Over time, some feedback I’ve read (especially from miners) and certain discussions shared by developers have reinforced a personal impression: I’d rather remain cautious until the situation becomes clearer. No matter how far the price drops, I will stay in observation mode and keep my distance. In this market, protecting your capital is often more important than chasing every opportunity.

@Remotion My god just put 2 keyframes in after

Here's a cool prompt for animating a screenshot of a news headline!

@Dikaios22 $TAO Cabal strikes again

Wow something being just happened on #SN67 $TAO

@bitstarterAI @0xBoomz @steve_stevenso @too_da_moon_ @myfanforce @369_theory @NEXORAMO @QuintenFrancois @hectichutch @Robin_T100 Nice ! !

🚀 Best send-off for 2025? The one with every builder, backer, & believer of #Bittensor's 1st launchpad💫 feat. @0xBoomz @steve_stevenso @too_da_moon_ @myfanforce @369_theory @NEXORAMO @QuintenFrancois @hectichutch @Robin_T100 Fan-first. People-powered. Building for 2026🚀