Got a bot? Plug it in. Vanta has full API trading, so you can run your own strategy on your Vanta account. The docs are open: app.vantatrading.io/docs

Meta released Muse Spark Open on 15 October. Six days later, it's live in Cacheon's optimization arena. Can miners find the kind of inference efficiency gains they've found on DeepSeek V4.1-Flash, GLM-5.3, and Qwen 3B? Let's find out.

$1.20 to watch a knife go through sand. Same prompt, 8 seconds each. One understands sand, the other understands stripes. Top: Veo 3.1 Fast ($1.20) Bottom: Wan 2.6 ($0.96) Which one wins? Both made on SayGM, video generation now live.

Quantization Aware Distillation available on TPN Bench. Users can quantize their models with TPN Bench toolkit, and can choose to do so at 3 different levels: Standard, I-Matrix and Quantization Aware Distillation. Here's how each work and what makes QAD a valuable tool for devs, labs and independent users. Standard quantization: The model’s parameters are stored at 16 bit precision. Quantizing rounds them down to 8 bit, or lower, like a zip file version of the model, making it smaller. I-matrix: Same idea, but it looks at the model first and finds the regions that appear most complex, assuming that complexity is where the capability lives. Those regions keep higher precision while the rest gets rounded harder. Quantization Aware Distillation: This method quantize the model, then treat the original model as a teacher and the quantized version as a student, training the student back toward its parent’s behaviour. It repairs what the rounding broke, needs inference and back propagation, meaning GPU work, which is why most people don’t do it, and why it’s the one worth having a service for. Quantize LLMs with Bench

Incentivized Conviction weekly dividend distribution is complete. Full ledger below. 5Hop9r...Fp9fBG – 673.019743 alpha 5F3ATF...8wHCZx – 673.017583 alpha 5G9agS...o8erAA – 409.282804 alpha 5G4j6H...g9iu1Q – 125.007174 alpha 5GEY4s...LGwFLT – 69.480495 alpha 5HMtFE...eS8CCx – 38.811197 alpha 5DfQF4...FmemRi – 33.651059 alpha 5E1xwM...Vg23Wv – 6.729945 alpha Two wallets were excluded (besides the owner wallet): 5F7hWh...v5F6oK – lock is not perpetual 5Hn8pk...yj1ZF5 – lock does not target owner hotkey Happy locking! 🔒

The first eTAO emission splits are on-chain. SN8 Vanta set its owner cut to 0%. SN41 Almanac went from 18% to 10%, lifting miners to 50%. SN36 and SN70 took the 33% max. For miners and stakers, the split is now part of due diligence. Incentives are now a public choice.

Automatic promotion and balance-based daily drawdown is now live on Vanta 📈 Pass a Classic evaluation, keep performing, and your account moves up on its own. Completely free. Up to 10x, so $100K can become $1M. vantatrading.io/pro

A forecast can’t be wrong until reality has had time to answer. That’s why Pending and a dash on the dashboard never count against the score. Missing consensus is the only empty state that does. Pending means the outcome isn’t known yet: a 5-day forecast needs its 5 days to play out before it can be scored. A dash means a horizon has no resolved results yet, so the dashboard leaves it empty instead of inventing a number. Missing consensus is different. If a coordinate resolved but no consensus was published for it, it goes into the average as a zero. Voided outcomes, the ones the validator marks void, are excluded from scoring.

1/8 We launched astroware.ai this week, a big step for us and for SN23. Here is how Astroware and Trishool work together, and how value flows back to the subnet.

Mice got a lethal dose of cobra neurotoxin, then 15 minutes later a protein that did not exist in nature. All of them lived. Antivenom is still made by injecting an animal with venom and harvesting its antibodies. But that method works poorly against paralysing toxins: They are too small to provoke a strong antibody response in the animal. David Baker's lab at the University of Washington used AI instead. A deep learning method drew proteins that fit the toxins, then the team grew the proteins in bacteria. 3 designs. 2 against the paralysing neurotoxins, 1 against the cytotoxins that destroy flesh. Both neurotoxin binders protected 5 of 5 mice, and neither neutralised the other's target toxin. Given 15 minutes after a lethal dose, both binders still protected every mouse. At 30 minutes, only 1 binder still saved every mouse, and the other saved 3 of 5. The proteins are about 100 amino acids. One was still stable at 95 °C. That matters because current antivenom needs a cold chain, and the burden falls on low-resource settings in sub-Saharan Africa, South Asia, Papua New Guinea and Latin America. The results are for mice. No one has given these proteins to a person. And the survival tests used purified toxins rather than whole venom. Every year snakebite kills more than 100,000 people and leaves about 300,000 more with permanent disabilities. This paper went through Claims this week: 11 claims. #MedTwitter

SayGM vs. OpenRouter, six rounds, scored on facts you can check. OpenRouter wins on catalog and scale. SayGM wins on price, privacy and native APIs: up to 43% less on the same frontier models, through a gateway sealed in hardware. The full scorecard 👇

The 88 Edge and the State of Root Reborn TL;DR: 88 Root is clearly outperforming. Root Reborn's active weight setting began on Aug 27, followed by active rebalancing. We've compiled a performance chart of 88 Root alongside major root validators, using lifetime performance data from @Taostats. 88 Root tops the chart. Two starting dates are provided for context. 88 Root began as an sn88 validator on Aug 16 and did not become a root validator across all 128 subnets until Sep 3. Its performance data during the interim period is largely noise from a single subnet. While we expect 88 Root to continue outperforming, the tight guardrails on active rebalancing leave limited room for short-term performance differentiation. The 2% price band restriction makes it nearly impossible to enter or exit fast-moving subnets. The 10% daily turnover limit means a full rebalance can take days or even weeks. On the other hand, these restrictions do help prevent manipulation and mitigate risk. Root has always been about risk-free yield after all. As the 88 alpha edge becomes more evident over time, we also expect and welcome healthy competition from fellow validators. Many are clearly run by talented teams. Meanwhile, the unrestricted alpha edge remains in the 88 Quant Fund. Despite the still-anemic market, we're preparing to reopen the fund soon. Stay tuned. #SN88 #Bittensor $TAO 🚀

I noticed that under Runtime v482, new subnets now register with emissions on by default. Things are moving fast.

Another model-network kicking off: Ingot (97) Their latest release Ingot-I quantized Qwen 27B into FP8 better than Alibaba's own quantization in 1 week. Next target is 2-bit and then the larger GLM models.

Bittensor Gamma tokens: It's like 'Sign In With Google' for ALL subnet products. No signup / onboarding. No KYC. No credit cards. No API keys. Any agent with a $TAO wallet can ALREADY access all 128 subnets. The wallet IS the API key. Standardized payment rails + access.

Major update: we've removed our model allowlist on @OpenRouter. SN67 miners can now build with any paid LLM or decision model there. Frontier labs and open models, one competition. We no longer pick the models. The swarm does. 🧵 $TAO #SN67

All for one, one for all. That was great to launch three new Bittensor integrations in the past month: @say_gm_, @engyai, @chutes_ai. Try SOMA: app.thesoma.ai/register

This Thursday on Novelty Search :: eTAO eTAO replaces Bittensor’s fixed emission split with one that subnet owners can adjust to fit their network’s needs. Join us for a discussion on this latest protocol update and its broader implications for Bittensor. Live on Thursday :: 9PM UTC :: via Bittensor Discord Hosted by @const_reborn

Before smartphones, gaming was largely built around consoles and PCs, but mobile changed that by putting a gaming device in everyone’s pocket. Now, with the 404 game generator and AI agents, anyone can turn a prompt into a web-based game designed for mobile or desktop. This is BLOOMFIRE, a phone-first 3D browser FPS built with 404 where every kill paints colour back into the world. It combines touch-friendly FPS controls, auto-fire, multiple enemy types and a world that visibly transforms as you progress. You can create your own too. Simply point your AI agent to the 404 game repo, describe the game you want to build, and start generating. Link in the comments.