Pinned Tweet
JUST IN: @const_reborn shares the vision for gamma tokens, usage credits that subnets could fund from a slice of their emissions and spend on compute, inference, and other infrastructure.
What this means ⬇️
→ Subnets stop being isolated prize pools. They can buy services from one another instead of only competing for $TAO
→ Gamma is meant to work as a transferable credit, not just another speculative subnet token
→ Access is designed to reach beyond Bittensor, so outside compute and inference providers could take the same credits
→ Emissions become operating budget, not only a reward that gets sold into TAO
If a subnet can spend what it earns on real work, Bittensor starts looking less like a scoreboard and more like an internal market for intelligence.
Intelligence ττ retweeted
Two years ago, Hippius was a storage network on Bittensor.
Today it’s a full cloud platform. Encrypted storage for files, S3 for your apps, confidential VMs, AI model registry and a CDN.
A platform like that deserves a website to match.
Meet the new hippius.com👇
COOKING: @1Password is giving AI agents a way to log into websites without handing them the passwords.
Integrations are rolling out for Muse and Grok Bot, with Claude, dots, and more coming soon.
Full details ↓
tao.media/1password-launches…
ICYMI: @mentatminds is giving $TAO holders an automated way to gain exposure to robotics-focused subnets
Est. APY of the index is ~50%
"Across 20+ common tasks, Auto delivered results similar to Opus 5.5 and Astra at about 90% lower average cost."
- @NotionHQ
COOKING: @NotionHQ's AI agents can now work on tasks for up to 24 hours, with computer and browser access built in.
The upgrade also adds Auto model routing, which matches leading models across 20+ use cases at about 90% lower average cost.
Users can connect Plus, Pro, or Go ChatGPT subscriptions to their personal Notion Agent, while a new credit system separates everyday AI work from manually selected models.
tao.media/notion-extends-ai-…
Mistral has released Voxtral 3 with a LiveKit-based SDK that links real-time transcription, reasoning, tool calls, and expressive speech for phone and web conversations.
The stack supports transcription in 50 languages, sub-500-millisecond real-time latency in more than 35 languages, and conversational turns in under one second.
tao.media/mistral-releases-v…
COOKING: @agilityrobotics is turning Pittsburgh into a software, AI, and service hub for its Digit humanoid robots as its customer base expands.
The planned 50,000-square-foot facility will add 35,000 square feet to Agility’s existing footprint and bring skill development, deployment, fleet monitoring, and repairs under one roof.
tao.media/agility-robotics-p…
NVIDIA and Germany are teaming up to bring physical AI into the country’s industrial base, while Jensen Huang plans investments in four German startups.
The partnership connects NVIDIA’s technology with Germany’s manufacturing expertise and Industry 4.0 network.
The planned investments in Agile Robots, Avelios Medical, Mirelo, and Reflex Aerospace span robotics, hospital software, generative audio, and spacecraft, and sit within NVIDIA’s nearly $10 billion commitment across Europe.
tao.media/nvidia-and-germany…
Agile Robots’ Agile ONE SN gives robotics researchers a preconfigured humanoid platform with NVIDIA’s robot-learning stack and Jetson Thor computing built in.
NVIDIA says the ready-made setup could cut integration time from weeks to days, with release planned for early 2027.
tao.media/nvidia-adds-isaac-…
COOKING: OpenAI has acquired @ashokaraj_'s Orchestra Bio for drug-development program management.
The platform can track studies, manufacturing, and timelines across a program, flagging how a failed batch could delay a clinical trial.
tao.media/openai-acquires-or…
JUST IN: @YumaGroup is asking the SEC to add a new Subpart F for token issuers that continue selling and buying back tokens after an initial offering.
Full details ↓
tao.media/yuma-urges-sec-to-…
COOKING: @NotionHQ's AI agents can now work on tasks for up to 24 hours, with computer and browser access built in.
The upgrade also adds Auto model routing, which matches leading models across 20+ use cases at about 90% lower average cost.
Users can connect Plus, Pro, or Go ChatGPT subscriptions to their personal Notion Agent, while a new credit system separates everyday AI work from manually selected models.
tao.media/notion-extends-ai-…
JUST IN: @say_gm_ is giving AI apps a safer way to use a user's existing model balance without sharing a main API key.
Its new OAuth flow creates separate credentials for each app, with user-controlled permissions, monthly spending caps, and instant revocation.
Developers can also earn SN28 alpha rewards from qualifying referred top-ups.
tao.media/saygm-adds-oauth-l…
Perplexity Computer can now run the full AI agent loop locally on Apple silicon Macs, keeping private-file workflows on the device.
The Mac app can work across files and apps such as Notes, Messages, Mail, and Numbers, while local tasks use no Computer credits.
Users can choose Qwen 3.6 or PPLX Pro.
tao.media/perplexity-adds-fu…
Bittensor’s Etao upgrade lets subnet owners rewrite the default 18% owner, 41% validator, and 41% miner split to fit how their networks operate.
And subnets have already started adjusting to the new rules of the game.
Explore the update ↓
tao.media/bittensor-activate…
COOKING: DEWALT’s DALE has drilled more than 385,000 holes on data center job sites, with speeds up to 10x faster than conventional methods and over 99% accuracy.
It just earned a spot on TIME’s Best Inventions of 2026.
tao.media/dewalts-dale-data-…
ICYMI: @bot can now be summoned with the iPhone’s Action Button
Assign Bot to the button through an iOS Shortcut, then start talking without opening the app
tao.media/grok-bot-adds-ipho…







