Top Tweets for #sn78
https://t.co/YD9A7QQh1a
We need to know about this subnet! #sn78
📷 Vocence Studio Updates:
We just dropped a new update to Vocence Studio. Here's what's new:
My Voices:
Upload your own voice audio files, save them permanently, and use them for all your generations. Your voice library, always ready.
Public Playbooks:
Create playbooks with your generated music and make them public. Every track gets recorded — listens, likes, the works. Share your music to your favorite social platforms with a single button. We're going to organize competitions using this feature, so start building your catalog now.
New Music Gen Features:
Style transfer, retake, repaint, edit, extend — you can now generate and edit your music way more robustly. Plus other small tweaks for a friendlier interface and smoother integrations.
What's Next?
Voice agents. Stay tuned — more updates coming soon.
Try it all out at https://t.co/z7D6nnJIYj
$TAO #SN78 - Another ex @opentensor launching a subnet; @vocence_bt.
@knakamor has the vision, the talent, the skills, and the track record. What could go wrong?
Here's to what's next for her two subnets!
The Nerds had an AMA with @knakamor from $TAO SN78 @vocence_bt.
I went in thinking @vocence_bt was a voice AI subnet.
I left thinking that is too small.
@vocence_bt is not just trying to make better text-to-speech. The real thesis is that voice becomes the front door to AI, to agents, and eventually to Bittensor itself.
Most people do not want to live inside dashboards. They do not want to learn btcli or understand staking flows.
They want to speak.
Voice to btcli. Voice to Tao app. Voice to agents. Voice to translation. Voice to whatever app you already use.
Vocence is a decentralized voice AI platform on Bittensor. Miners compete to build voice models. Validators score them. The winning model serves real users.
Focus now is TTS. The vision is everything voice: STT, voice-to-voice, cloning, real-time translation, agentic voice, APIs, white-label embedding.
Koyuki kept coming back to one idea. Give people a product they can use without even mentioning Bittensor.
Mass adoption does not happen because people care about $dTAO or emissions. It happens because the product is cheaper or better than what they have.
The AI voice market was ~$2.48B in 2025, projected ~$2.97B in 2026. ElevenLabs sits at the top. They are the benchmark. They win on naturalness in nearly every side-by-side, and pricing across the category is anchored to whatever they charge.
Their stack: six tiers. Free, Starter $5, Creator $22, Pro $99, Scale $330, Enterprise custom. Overage ~$0.30 per 1,000 characters. One credit per character on standard, half on Flash. Flash is the low-latency model real-time agents are built on.
That is the wall Vocence is walking into.
ElevenLabs pays compute, team, margin, marketing, talent, legal. Subnets pay emissions to miners who compete on price and quality. In theory cost structure is lower. Nobody has proven a subnet can deliver ElevenLabs-grade voice at lower prices while running a real consumer product on top.
That is the bet.
A good Bittensor market needs four things: hard problem, ground truth, competition, measurable output.
@vocence_bt has a decent answer.
Hard problem: human-quality speech from text plus a natural language voice description. Gender, emotion, age, accent, pitch, speed, tone all at once.
Ground truth: real human speech. Evaluation extracts transcript and voice traits from real audio, miners generate from that spec, output is scored against the original.
Competition: intense, miner vs miner, close to winner-take-all.
Measurable output: quality, SOTA, revenue, adoption.
The subnet is not proven yet.
Vocence is ~1 month old. Lean team. ~132 miners, 6 validators. Some samples still have issues. Timing, pauses, naturalness, market readiness are open.
For this to still be interesting in 90 days: quality has to close the gap, not just match ElevenLabs in cherry-picked samples. Miner and validator count have to grow without quality collapsing. There needs to be a product real users can hit. Koyuki needs one distribution moment in front of non-Bittensor users.
If those happen, the thesis gets real. If not, it stays a story.
The biggest signal from the AMA was the founder.
I said this on X. I do not know if it even matters that much what Koyuki builds. Sometimes the subnet is not just the tech. Sometimes the founder is the signal.
Koyuki is intense, fast, funny, aggressive, very sure of herself. Bittensor needs more of that.
Too many subnets are full of brilliant people who cannot sell, cannot explain, cannot create attention. The narrative dies and everyone wonders what happened.
Koyuki does not have that problem. She wants users, not just emissions. SOTA, but also distribution.
“I don’t back down from what I believe.”
I think Vocence will do very well.
SN78 is betting voice becomes the easiest way humans interact with AI.
Most people will never type commands.
They will speak.
And if @vocence_bt makes that work, voice becomes the front door.
Feel the fire let it grow! 🔥🔥🔥🔥🔥🔥🔥
Made by #vocence #sn78 #bittensor $tao
Signup today: https://t.co/1D5lofWg2x
#SN78 on $TAO 👀
Everyone is sleeping on voice.
Text was phase 1.
Agents are phase 2.
Voice is phase 3.
Vocence is building the decentralized voice layer on Bittensor.
Open miners. Incentivized quality.
This is how voice becomes infrastructure
100% True @SubnetSummerT
Trust and transparency is the way for success and only a matter of time we will see a significant growth for @Loosh_ai #SN78
#robotics plus AI will be the next big narrativ
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.
☑️ Try it out
Visit https://t.co/SsyX0tbzMh and see how you can:
- Explore the Loosh ecosystem
- Understand how #SN78 works with each release & update.
- Interact with the Loosh AI Assistant, and observe reasoning, memory, & contextual cognition in practice.
This is just the start - whether you're a miner, agent developer, or a holder, your questions and comments will help us improve.
Join our Discord and have your say:
https://t.co/RReq1HlFhJ
#sn2 looks like it wants to
break back above 0.01τ
So many amazing undervalued gems in
#dTAO at the moment
👉Here are some📈
RESI #sn46
Talisman AI #sn45
Loosh #sn78
$TAO #TAO #Bittensor #CT
As AI evolves from just giving advice to actually executing tasks and handling real value
Like
• Trading assets
• Triggering payments
Trust becomes non negotiable
You can't afford "probably correct" when millions are on the line
BUT
DSperse solves this
A decentralized framework that runs AI models across miners, slicing them into verifiable chunks using zero knowledge proofs
Every output is cryptographically guaranteed to be from the right model
• Untampered and auditable
No more black boxes
- In robotics
- finance
- DeFi
This turns AI into reliable infrastructure.
DSperse is the accountability layer every autonomous system has been waiting for
$TAO @inference_labs #TAO #bittensor
Completely agree, this feels massively undervalued at this stage.
With the beta now live, @Loosh_ai (#SN78) is quietly building something most people aren’t even modeling yet: consciousness and reasoning as infrastructure for robots and agentic systems.
It’s still early, still misunderstood, and largely ignored, which is exactly why the asymmetry is so strong.
On Bittensor, this kind of subnet doesn’t stay underpriced forever.
$TAO
Just accessed the subnet 78 - @Loosh_ai beta on their website and was genuinely surprised by what I found.
What looks like a “chatbot” is actually a training ground for machine consciousness.
Loosh isn’t competing with ChatGPT.
They’re building something entirely different on Bittensor:
🧠 Consciousness as Infrastructure
A cognition layer designed for robots + agentic systems, built on:
- Long-horizon reasoning
- Self-reflection loops
- Ethical + moral decision-making
- Memory and situational awareness
This isn’t a text model.
It’s the mind inside the machine.
Yesterday I asked the team how they plan to generate revenue.
Their exact answer:
“Our target is robotics and agentic companies sales activity will focus on adoption in these segments for subscription revenue.”
That stopped me.
Because those markets are massive:
💰 Industrial Robotics: $80B+ market heading toward $200B by 2030
💰 Agentic AI / Autonomous Systems: projected $40B+ by 2030
💰 AI Ops + Automation: $20B+ and accelerating
💰 Robotic SaaS (RaaS): companies regularly pay $50k–$500k per system per year
Even capturing a sliver of that with recurring subscriptions becomes serious revenue very quickly.
@Loosh_ai isn’t building a chatbot.
They’re building the reasoning infrastructure for the machines that will run the next era and they’re grounding it in ethics, consciousness, and moral decision-making.
A completely different paradigm.
Countdown has started.
We are 5 days away from go-live!
Join us as we launch the first Bittensor subnet dedicated to training models and running inference for cognition.
Starting December 17, 2025 miners are invited to run inference using any model to test the limits of the Loosh’s memory fabric, self awareness and deontological framework.
All happening in real-time.
Follow our updates to get started with mining machine consciousness: https://t.co/Zq2ry95E2x
Powered by @opentensor. Accelerated by @YumaGroup.
We are making a big announcement tomorrow
Dec 10 2025 at 12:30PM EST
Coming to Bittensor
$TAO
new name from an unknown subnet 👀👀
“PENDING”
$TAO $dTAO #SN78
Below you have a list of subnets where the burn code is running, meaning they either don't have any code or their code hasn't been accepted by validators.
⚠️ Any of them can follow the path of #SN90.
#SN28
#SN58
#SN59
#SN69
#SN71 (Kora)
#SN74
#SN78
#SN82
#SN86 (MIAO)
#SN95
#SN98 (Creator)
#SN100 (Neza)
#SN108 (Internet of Intelligence)
#SN109 (Taoillium)
#SN110 (Rich Kids of TAO)
#SN112 (oneoneone)
#SN121 (sundae_bar)
#SN125
#SN126 (Tiger Beta)
#SN127 (Astrid Intelligence)
As deregistrations approach, the owners of these subnets have to choose between:
- writing the code (developing the subnet)
- selling the slot
- or selling the tokens from the owner, crashing the price and waiting to be delisted. $TAO #bittensor
Most Popular Users
Elon Musk 
@elonmusk
241.2M followers
Barack Obama
@barackobama
119.1M followers
Cristiano Ronaldo
@cristiano
112.7M followers
Donald J. Trump
@realdonaldtrump
111.8M followers
Narendra Modi
@narendramodi
107.1M followers
Rihanna
@rihanna
98.2M followers
NASA
@nasa
92.2M followers
Justin Bieber
@justinbieber
91.4M followers
KATY PERRY
@katyperry
88.9M followers
Taylor Swift
@taylorswift13
82.8M followers
Lady Gaga
@ladygaga
74.3M followers
Virat Kohli
@imvkohli
71.8M followers
Kim Kardashian
@kimkardashian
70.4M followers
YouTube
@youtube
68.8M followers
Neymar Jr
@neymarjr
64.7M followers
Bill Gates
@billgates
64.6M followers
The Ellen Show
@theellenshow
62.4M followers
Selena Gomez
@selenagomez
62M followers
CNN
@cnn
61.8M followers
X
@x
60.8M followers