VCavs retweeted
Best Performing Projects on Bittensor This Week
REAX (SN92) | @reaxlabs — +720.4%
THOR (SN112) | @THORMachines — +31.3%
SayGM (SN28) | @say_gm_ — +17.5%
Subnet 113 (SN113) | @TensorUSD — +16.0%
Attelier (SN126) | @attelier_ai — +15.3%
Quasar (SN24) | @QuasarModels — +15.1%
Green Compute (SN110) | @green_compute_ — +14.5%
deprecated (SN39) — +14.1%
Almanac (SN41) | @almnc_ai — +13.5%
DaniLabs (SN2) | @danilabsai — +9.0%
Source: SubnetRadar
VCavs retweeted
We just killed cold outreach.
Introducing Leadpoet: the AI sales agent that finds buyers who need you now, writes personalized outreach, and follows up on LinkedIn and email.
Used by Dropbox, OnePatch (YC26), Muttdata.
Comment “Poet” for 100 free lead credits to find qualified buyers.
VCavs retweeted
Introducing eTAO: a new system giving subnets control over their alpha emission split.
Subnet owners can now decide how much goes to miners, validators, and themselves, based on the specific needs of their subnet.
VCavs retweeted
Sign in with SayGM is live
Apps can now add a Sign in with SayGM button. Users connect their SayGM account in a few clicks, and the app runs on their existing SayGM balance, with no API keys to copy and paste.
You choose what each app can do
When an app asks to connect, you pick which permissions to allow:
- Use models: the app can make requests with a dedicated key that's only for that app.
- Read usage: the app can see what it has spent on your account.
- Read profile: the app can see your basic account details.
You set the budget
If you allow model access, you set a monthly spending limit for that app or choose unlimited. Each app has its own limit, and only you can change it. Limits reset at the start of each month (UTC).
You stay in control
Account → Connected apps shows every app you've connected, with its permissions and spend for the month. Revoke turns off that app's access and key straight away.
Building an app? Get paid for the users you bring
Sign in with SayGM uses standard OAuth 2.0 with PKCE. Your users bring their own balance, so you never front their inference costs.
Add your affiliate referral code to your Sign in with SayGM link. Every new SayGM account created through your button counts as your referral, and you earn alpha from the pool on what those users spend.
👉 Get your referral link: saygm.com/affiliates
👉 How rewards work: saygm.com/affiliates/how-it-…
👉 Register your app: saygm.com/contact (topic: Partnerships)
VCavs retweeted
$930,128 paid out to traders. Receipts included.
339 rewards to 122 traders, every one of them on a public ledger. The first million is getting close.
app.vantatrading.io/rewards
VCavs retweeted
Goodbye owner emissions, and thank you for the support. We've moved to 0% owner, 50% validator, 50% miner splits.
We believe removing owner emissions is part of a subnet’s maturation. Once you’ve built a stable & fair incentive mechanism, you can compete within it yourself.
That’s where we are with SN8. Anyone can provide trading intelligence to the network through an individual miner or an entity miner. We’ll compete by running an entity miner, and we see that as our future. Individual miners earn alpha for the PnL they contribute to the network. Entity miners, including prop firms, earn through profit splits with the traders they fund, with the shared goal of producing the network’s best risk-adjusted PnL.
We’ll keep building SN8 without owner emissions. APY from our existing holdings will fund subnet features and improvements. We’ll rely on mining emissions and Vanta Trading’s ability to compete to fund our continued growth on VT. As we have in the past, we’ll use specific wallets for Vanta Trading sales and any owner emissions, so the associated spending is easy to track.
Our goal is to grow the network to thousands of miners. We built SN8 to be more than a network where we’re the main competitor. There are many individual miners and entities already running on the network. With Null Consensus, we want to grow that exponentially. More miners will be able to join with the right rails, making it simpler to get started: USDC in, miner running.
As more miners join, we expect more network weight to flow to miners, bringing more trading intelligence and creating more capacity to meet demand for funding. That demand can generate further revenue for the subnet through copy trading.
We see this as the natural progression of a subnet, and believe SN8 is ready to begin that transition. It may take time, and there may be hiccups along the way. Our expectation is that, gradually, miners will receive 100% of the network’s weight. We’re starting that transition by moving our own owner to provide a bigger split to miners & our holders.
VCavs retweeted
Bittensor 24H Market Snapshot
🔹 NIOME (SN55): Generates high-fidelity synthetic genomic and CRISPR datasets so researchers can train models for precision medicine without using real patient genomes.
🔹 Attelier (SN126): Decentralized generative video platform for creating and editing AI videos.
🔹 Almanac (SN41): Prediction-market intelligence layer that rewards early, accurate forecasts across events.
🔹 SayGM (SN28): Private drop-in gateway to Claude, GPT, Gemini and other frontier models, routed through TEEs for privacy and competitive pricing.
🔹 KubeTEE (SN90): Confidential AI factory that runs state-of-the-art inference, agents and microservices inside hardware TEEs on decentralized Kubernetes clusters.
🔹 Engy (SN53): Verified inference for frontier open models, served cheaply on consumer GPUs with cryptographic proofs that the correct model actually ran.
🔹 REAX (SN92): Verified inference for calibrated decision models that output finite answers (yes/no, multi-choice, scales) with fidelity checks.
🔹 Lium (SN51): Decentralized GPU rental marketplace where providers list NVIDIA GPUs and renters get on-demand compute.
🔹 Actual (SN95): Heterogeneous inference network that turns ordinary consumer hardware into a private, coordinated AI inference fleet.
Data source: @subnetradarcom
VCavs retweeted
We set the target at 90%. Miners came back with 98%.
Fallen person detection on SN44 went from a 43.6% baseline to 98% on crowded street footage, scored in public every hour. Open weights.
So we're raising the bar. Harder footage goes in next, and the difficulty keeps climbing from here.
VCavs retweeted
The embodied AI data market is forecast to grow from an estimated $536M in 2025 to $4.9B by 2032. Roughly 9x in seven years.
But volume alone isn't enough. A demonstration recorded for one robot arm can be useless to a lab training a different one. Labs need data cut into individual actions, labelled with language instructions and checked for consistency, all to their own spec.
THOR starts with what the lab needs, then turns those requirements into paid annotation and verification tasks for providers on Bittensor. Outputs are checked against the lab’s specifications.
We're building so labs can scale their data operations without coordinating every provider themselves.
VCavs retweeted
“just use a foundation model” they said.
so we benchmarked 10 of them against Manako on a real forecourt damage clip.
same full 1280x720 frames, same class prompts (person, car, fuel dispenser, fuel nozzle), no crops, no zoom.
the hard class is the nozzle/hose. small, thin, occluded, and the thing that actually gets damaged.
- Grounding DINO labelled a bin as a fuel nozzle.
- Florence-2 labelled the canopy pole as one.
- OWL-ViT, YOLOE and YOLO-World got the dispenser at best, never the nozzle.
- RF-DETR doesn’t have the class.
- Gemma 4 E2B described the scene, boxed the pump, and returned no incident time.
the only open model that localised the nozzle was full SAM3. 3.45 GB, 0.2-1 FPS on a Jetson Orin Nano.
Manako (3 specialists - 29.5 MB), 3-6 FPS on the same Jetson, tracked the nozzle at 5 FPS and fired a live equipment_damage event on the hose drop at 10.3s.
117x smaller than the only model that could keep up.
open-vocab models are great at the classes the internet photographs a lot.
operations live in the long tail:
nozzles, hoses, cabinets, the small parts that break.
that’s where specialised detectors win, and it’s why we built Manako the way we did.
VCavs retweeted
Bittensor bittensor:native is the best looking multi-year chart primed for a breakout to over $3,000 imo
VCavs retweeted
Common question: if I buy $TAO directly on CEX or swap USDC into subnets, that bypasses $TAO. So why will $TAO price rise?
Answer: No matter how you bought $TAO subnets, eg via CEX or other method, the CEX buys $TAO first then swaps it. You just don’t see it on the front end.
VCavs retweeted
So to be super clear:
I have USDC on Ethereum.
I can buy @engyai subnet tokens on Bittensor.
One transaction. Directly. This is a bridge + DEX in one go.
Later, I can sell my ENGY SN53 and I will get USDC on Ethereum.
VCavs retweeted
Decision Models may be much bigger than chatbots.
Florian explains why these smaller models could power millions of fast, cheap decisions inside everyday software, and why Bittensor could be a natural home for them.
Read the full conversation 👇
taodaily.io/florian-explains…
VCavs retweeted
NEW CLIENT ANNOUNCEMENT
We’re happy to announce that Beam is now covered by Sentios Continuous Coverage.
Beam has a live badge on their Github README linked to the Sentios Active Projects Page.
sentios.io/active-projects
Let everyone know your code has Continuous Coverage.
DM us to join the ranks.
@Bitsecai $TAO
