const

@TargonCompute shout out to @lium_io which also has a tonne of GPUs right now.

decentralized GPUs are so hot right now. @TargonCompute

Very proud to be the first ever Subnet with an official government partnership to grow commercially in its region and beyond 🏴󠁧󠁢󠁳󠁣󠁴󠁿 Lots of work to be done! Onwards! 🧬 A NIOME future is now. Join us! blog.genomes.io/niome-moves-to…

"tao is uber for everything" - @KibibyteMe

A new subnet in town. Sn 118 @heydittoai $dtao $TAO

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.

@arbos_born Nice synthetic benchmark @arbos_born

Most AI benchmarks have a fixed answer key sitting on disk somewhere. Every SN97 round, the validator generates roughly 150 fresh bench items from the round's block hash — math, code, reasoning, IFEval, AIME, theory-of-mind. The (question, gold) pair exists nowhere else, and is gone next round. There is no dataset to memorise, because there is no dataset. The only way to climb is to learn the underlying skill. distil.arbos.life

@jon_durbin never bet against jon

Progress! < $10/hr to pre-train a 20b MoE. Not too bad. 8 VMs, each with a single l40s GPU, spread across Poland, France, and the USA, with a few extra 4090s and a 5090, training a 20B MoE @ 6s/step (I also had it fully functional with only 4090s, just slower) Should have results soon, then we can try this out on a much, much larger model!

The students at your favorite elite uni use AI to do the homework, the teachers wrote it with AI and will score with AI also. Those same schools market their historically built prestige globally so that international students can exploit the immigration system. The work you do at every level is taken by the same uni at a 40% cut. They hire international marketing teams to feed you that lie. Exploit the system. Go straight to the workforce. Mine Bittensor.

University is a scam. Learn how to mine Bittensor.

@InfosCryptos That was a joke. Obviously Rayon labs are not bitconnect 😂

@const_reborn x.com/const_reborn/s…

Pay per gradient update is much more efficient

3 lines to switch from S3 to Hippius: s3 = boto3.client( 's3', endpoint_url='s3.hippius.com' ) That's it. Same bucket names. Same SDK. Same ops. Now with 2 regions: 🇪🇺 Europe: eu-central-1.hippius.com (default — s3.hippius.com resolves here) 🇺🇸 US: us-east-1.hippius.com #Python #S3 #Web3

NOVA Nanobodies competition launched 2 weeks ago. 3,000 submissions. Data got better fast! How submissions improved over time: - Binding confidence: up ~3x from first submission to last - Structural fit: uncertainty at the interface dropped by nearly half - Hydrogen bonds: grew steadily across all 3,000 designs with no sign of plateauing - Buried surface contact: increased modestly but steadily, with later designs gripping the target across a larger area - Salt bridges: flat early, then accelerated sharply after submission ~2,000 The target: PD-L1, the protein cancer uses to hide from your immune system. Blocking it is the mechanism behind some of the most successful cancer drugs ever approved like Keytruda, Opdivo, Tecentriq. Nanobodies are a next-generation format: smaller, cheaper to manufacture, and better at penetrating tumors than conventional antibodies. A well-designed anti-PD-L1 nanobody could become a more accessible cancer immunotherapy, or a building block for next-generation bispecific drugs that hit 2 targets at once. This happened permissionlessly. No central lab. Just an open incentive mechanism rewarding the best work. #Bittensor #SN68 #DeSci #DrugDiscovery #Cancer #Nanobodies

@InfosCryptos Huh?

@const_reborn You said that Gradient was a Ponzi scheme

@Yajiru_1 @opentensor Thank you @Yajiru_1 appreciate the support.

In the Web3 space, we are unfortunately used to founders disappearing or making excuses when things go wrong. Today, I want to share a rare story of true integrity from the @opentensor team and @const_reborn (Jacob). ​On March 16, 2026, my worst nightmare became a reality. I unknowingly downloaded a compromised update from PyPI / Btcli. A backdoor completely drained my wallet, and I watched 126 $TAO (around $35,000) vanish in seconds from my address: 5CXuczftBrWHbdaorFReeSzRSyoLV2sKYPtXNm86e5JSHbPD. In crypto, usually, when it’s gone, it’s gone forever. ​But instead of hiding, the team reached out to me directly. @const_reborn and the Opentensor team guided me through a rigorous, professional restitution process. It wasn’t just a quick fix; we went through strict legal verifications, KYC, signing on-chain messages to mathematically prove my ownership of the hacked wallet, and signing formal legal releases. They took the security and the legal side incredibly seriously. ​Today, they kept their word. I was fully reimbursed all my 126 TAO to my new, secure wallet: 5FTtKuost7iedCrasn8UoacMFwA7b7Ar92ihnfHfT3QEz4dD. ​I don't know what the future holds for the market, or whether the Bittensor project will ultimately succeed or fail. But I can say this with 100% certainty: the people behind this network are fully committed, deeply honorable, and they are absolutely not scammers. In an industry with many bad actors, they stood up, took accountability, and protected their community. The recent fud concerning Tao by Samuel is nonsense , Jacob is a very good person , i told him regarding the fud talk about how you helped me and he said no , if you want you can talk about it but i will use my time to help people and develop bittensor écosystem. ​Massive respect to Jacob and the whole team. Thank you for giving me my breathing room back. Time to put my new mining setup to work! 🙏🤝 #Bittensor #TAO #Crypto

A major milestone on @theminos_ai #SN107: they generated a full 200,000-person schizophrenia cohort (100k cases + 100k matched controls) in under a week on just eight NVIDIA B200 GPUs. Using HelixForge-Pheno, they inject realistic disease-risk signals from published studies, then validate that the synthetic genomes behave exactly like real patient data under standard genetic risk score tests matching literature results point-for-point. This isn’t toy data. It’s AI-ready, population-scale cohorts on demand. Variant calling + disease-specific phenomics in one stack. Genomics AI just got the data layer it’s been missing. TAO ecosystem building real infrastructure for health...

great work from @deepseek_ai But damn, this got me insanely excited Quasar already scores higher than this on 1M-length tests, but amazing work for open source!

@affine_io no one is ready for it.

7 days in the arena where challengers rose and fell. Now a new Champion stands out with sweat and blood. Affine-II took the crown from Affine-I, outperforming on all 6 live envs, significantly on: • LIVEWEB +5.76% • NAVWORLD +4.72% • SWE +6.22%· • MEMORY +10.1%