ʟɪsᴀ

@andyyy @Loosh_ai on sn78

What are the best subnets on Bittensor???

@ossurion Thank you 🫡

@lisacheng memes are literally the upgrade path for consensus theres a delay but reality catches up to narrative keep cooking real relevance comes when u meme it first

Having some fun today creating memes for our project on Bittensor The team has worked really hard over the last week to update the scoring and consensus mechanism for Subnet 78 It’s worth it because decentralized training is the future of scaling AI From 7B to 32B parameters LFG

@TradeCoinDream Thank you

@lisacheng Love the energy and the meme! Rooting for Loosh to actually solve this. Keep building 🔥

@CBarraford @bittensor @THORChain Bittensor liquidity needs a DeFi bridge

hey @bittensor , love what you guys are doing here. As a co-founder of @THORChain, lets collab. Trustless cross chain swaps for $TAO

@SubnetSummerT @Loosh_ai @YouTube 👀

How @Loosh_ai Mining Works on Subnet 78 (Model-Agnostic + Consensus Rewards). youtu.be/nBz-rjl1DPg?si… via @YouTube

@Angello3400PRMR Thank you

@lisacheng Good luck Lisa 😼

It’s been a challenging few days for Loosh. Witnessing the volatility over the past few days has obviously been frustrating. Many thanks to everyone who’s been out there voicing support, reaching out, and continuing to contribute to the network. We are so grateful to be part of a community that supports each other, its what makes building open systems with Bittensor worth it for us. Over the past few weeks we’ve actually been a bit quieter than usual. Not because progress slowed but because we’ve been thinking carefully about how to explain the bigger picture behind Loosh. What we’re working on is ambitious. It sits at the intersection of engineering, semantics, philosophy, and real world systems design. Building machines that can reason in ambiguous situations isn’t a single technical problem but a whole stack of them. It's important and badly needed. AI models are getting more capable every month. Robotic capabilities are accelerating fast. Yet the crucial layer, the one that allows systems to: interpret context, exercise judgment, and operate safely alongside humans in the real world - is still missing, and it’s exactly the problem we’re working on. Chris and I have spent a lot of time working through how to clearly explain this vision, and we’re now ready to start putting some of those ideas out into the world. Over the coming weeks we’ll share: why safety ultimately depends on reasoning rather than rigid rules, why reasoning infrastructure may become one of the most important layers in the AI stack, why we need to talk about AI systems as well as models, and much more. In parallel, development continues as planned. We’re releasing a scoring update later this week that strengthens Sybil resistance and improves scoring fairness for miners. We're making changes to the Incentive Model that will shift consensus towards honest miners rather than a centralized few. We appreciate the feedback, the support, and we are committed to improving and ongoing transparency. We’re looking forward to sharing more. Thank you.

@urbanspaceme Thank you

@lisacheng Some alpha holders are here for financial returns, and will come and go as the tide turns. Others see the problem, the passion, skills and drive you bring, and understand we (humanity) desperately need this critically under-served layer of the AI stack. We the latter, remain.

@SubnetSummerT @Loosh_ai Thank you

Transparency like this from the team behind Subnet 78 @Loosh_ai is exactly what we need more of in Bittensor. Lisa openly sharing the infrastructure, tooling, and real-time visibility into the subnet shows a genuine commitment to building in the open and bringing the community along for the journey. From miner stats to deployment docs, everything is there for anyone to explore. That level of transparency builds trust. I’ll be standing behind Lisa and the Loosh team in full support as they continue pushing forward. 🤝 Building something as ambitious as decentralized cognition and machine consciousness isn’t easy but doing it publicly is how strong networks are built.

Attention: Sybil Miners on Bittensor, we are done being polite.

@Levi_Researcher thank you

@lisacheng Reasoning layer talk hits hard, that’s the missing bridge between raw capability and understanding

💡 Subnet Spotlight #10 💡 Subnet 78 - @Loosh_ai Decentralised Cognition Infrastructure on Bittensor Most AI systems today are powerful but incomplete. They can predict, generate, and respond but they can’t remember, reason ethically, or make judgment calls in the real world. Loosh is building the missing layer. Subnet 78 provides a cognitive architecture that sits between foundation models and real-world action, giving robots and autonomous agents persistent memory, ethical reasoning, and the ability to learn from experience. 🚀 What It Builds 🔹 Persistent Memory Fabric Loosh gives agents structured memory across sessions episodic, semantic, and working memory - modeled on how humans actually think. Agents remember context, learn from past decisions, and improve over time. 🔹 Ethical Reasoning Layer Every action is evaluated against ethical, legal, and contextual frameworks before execution. Agents can refuse requests, seek guidance when uncertain, and explain their reasoning. Judgment is built in - not bolted on. 🔹 Temporal Knowledge Graphs Loosh maintains a continuously evolving graph of entities, relationships, and events over time, giving agents a structured sense of before, now, and after. This enables long-horizon reasoning and stronger recovery from failure. 🔹 Decentralised Inference via Bittensor Subnet 78 connects to a competitive network of miners running inference, keeping costs efficient, performance high, and the system resilient as the network scales. 🔹 Emotional Inference (Q1 2026) Loosh is beginning to train models on authentic human emotional signals, enabling agents to understand and respond to human affect in a genuine and contextually appropriate way - going far beyond traditional sentiment analysis. 🧠 Why This Matters Robotics and autonomous systems are scaling fast: • Humanoid robots • Warehouse automation • Care assistants • Security agents These systems are rapidly moving from research labs into the real world. But the real bottleneck isn’t hardware or model capability - it’s cognition. Agents need the ability to: ✔ Remember past interactions ✔ Reason across time ✔ Make ethical decisions ✔ Operate safely around humans Loosh aims to provide the cognitive infrastructure that makes this possible. 📈 Product & Ecosystem Impact Subnet 78 is positioning itself as a core intelligence layer for autonomous systems: 🔥 Persistent memory for AI agents and robots 🔥 Ethical reasoning frameworks built directly into decision making 🔥 Structured temporal knowledge for long-horizon reasoning 🔥 Emotional intelligence capabilities for human-AI interaction 🔥 Decentralised inference powered by Bittensor As autonomous systems scale, the demand for safe, context-aware AI cognition layers will only grow. 📌 In Short Loosh (Subnet 78) is building a decentralised cognition layer for robots and autonomous agents on Bittensor - providing persistent memory, ethical reasoning, and contextual intelligence so AI systems can operate safely and effectively in the real world.

@micaelabazo @metanova_labs I’m very sorry for your loss. I lost my mother to pancreatic cancer last year. Agree, medical innovation is needed

6 months ago I lost my father to complications from chemotherapy. he beat pancreatic cancer against all odds, but the treatment ironically ended up killing him. I never stopped working on @metanova_labs because I believe *we* can do better. if you’ve ever struggled to direct your life, if you’ve ever felt powerless in the face of disease, if you’ve ever lost someone, then you understand there is nothing more important than medical innovation. for the first time in history we have the technological primitives to tackle the world’s most important problems. I would argue health is #1 on that list. with #bittensor you can fund and co-own these solutions. it's an opportunity to participate in and help shape the building of a new world. join us today.

Running a subnet is hard But we aren’t giving up We’ve seen some volatility in the subnet market over the weekend. That’s part of how open markets behave and we don’t speculate on individual trading decisions. What matters is that the fundamentals of the project remain unchanged. Development continues exactly as planned and we remain on track with the Loosh roadmap. We’re also releasing a scoring update later this week that strengthens Sybil resistance and improves scoring fairness for miners. Building open networks means continuously improving them as they grow. Our focus remains the same: developing the reasoning infrastructure that allows AI systems and robotics to operate safely in complex real-world environments. More updates soon.

@walls_jason1 This

Yesterday Mark Cuban reposted my work, DM'd me, and told me to keep telling my story. So here it is. I'm a Master Electrician. IBEW Local 369. 15 years pulling wire in Kentucky. Zero coding background. I didn't go to Stanford. I went to trade school. Every week I'd show up to a home where someone just bought a Tesla or a Rivian. And every time, someone had already told them they needed a $3,000-$5,000 panel upgrade to install a charger. 70% of the time? They didn't need it. The math is in the NEC — Section 220.82. Load calculations. But nobody was doing them for homeowners. Electricians upsell. Dealers don't know. And the homeowner just pays. I got angry enough to build something about it. I found @claudeai. No coding experience. I just started talking to it like I'd explain a job to an apprentice. "Here's how load calcs work. Here's the NEC code. Now help me build a tool that does this." 6 months later — @ChargeRight is live. Real software. Stripe payments. PDF reports. NEC 220.82 calculations automated. $12.99 instead of a $500 truck roll. I'm still pulling wire. I still take service calls. I wake up at 5:05 AM for work. But something shifted. Yesterday @vivilinsv published my story as Claude Builder Spotlight #1. Mark Cuban saw it. The Claude community showed up. And for the first time, I felt like this thing I built in my kitchen might actually matter. I'm not a tech founder. I'm a dad who wants to coach little league and be home for dinner. I just happened to build something that helps people. If you're in the trades and thinking about using AI — do it. The barrier isn't technical skill. It's believing you're allowed to try. EVchargeright.com

A quarter million dollars in GPUs is the entry ticket to “democratized” AI training. Each participant needed a minimum of 8x NVIDIA B200 GPUs. “Anyone with GPUs could join” is technically true the same way “anyone can buy a Gulfstream” is technically true. Twenty-plus contributors scattered across the globe just finished training a 72 billion parameter model on 1.1 trillion tokens, coordinated over commodity internet, with no central cluster and no whitelist. The engineering is legitimately impressive. They compressed gradient communication by 146x using SparseLoCo, let participants join and leave mid-run without killing the training process, and used a blockchain incentive layer (Bittensor subnet 3) to pay contributors in TAO tokens for staying honest and keeping machines running. Six months ago the largest permissionless decentralized run was INTELLECT-1 at 10B parameters. Covenant jumped to 72B. That’s a 7x parameter leap while removing the trust assumption entirely. Now the context that matters. GPT-4 trained on roughly 25,000 A100s in a dedicated cluster estimated at $500 million in infrastructure. OpenAI’s total compute bill hit $5 billion in 2024. The industry is racing toward $100 billion clusters by 2028. Covenant-72B benchmarks “competitively with centralized models at similar scale,” which means competitive with LLaMA-2-70B, a model Meta released in July 2023 using a conventional data center. The 72B number sounds massive until you remember frontier models are now in the trillions of parameters, trained on 10-15x more tokens, with post-training pipelines that cost as much as the pre-training itself. What Covenant proved: distributed training over the open internet works at a scale that would have been unthinkable two years ago. The optimizer is brilliant. The infrastructure leap is real. And “permissionless” still costs a quarter million dollars at the door.

TLDR Oil Tokenomics: - 3 token system (Brent, WTI, Dubai), each produced on different chains - Not 1:1 redeemable - 3 main bridges (Hormuz, Malacca, Suez) to move oil between chains. Hormuz is down - Devs (OPEC, US) control mining rate + insider supply (SPR) + sanctions

If youre going to @consensus2026 and working on Bittensor would love to meet up! Thanks Tatiana cant wait to hang out again :) Excited to see you!