English
Fezicles.eth
1.1K posts
Fezicles.eth
@Fezicles_eth
I like to make cool stuff and watch Football | Community and Marketing @Chutes_ai | Decentralised AI Maxi
Katılım Ocak 2022
221 Takip Edilen198 Takipçiler
OpenRouter has updated our provider status after verifying our privacy policy thanks to our recent updates
Chutes is in their default routing now!
openrouter.ai/provider/chutes
English
OMG, the whole @chutes_ai source code just leaked.
The API, the internal SDK, everything.
github.com/chutesai/chute…
github.com/chutesai/chutes
github.com/chutesai/chute…
And even the inference code for the most important LLMs, e.g. here Kimi-K2.5:
chutes.ai/app/chute/2ff2…
English
Top 20 Trending Coins on CoinGecko 🦎
This week, $TAO (@opentensor) leads the pack, followed by $SIREN (@genius_sirenBSC), and $BTC (@Bitcoin).
Are you keeping an eye on these?
coingecko.com/en/highlights
English
if you’re marketing a crypto project and your ads look professional, you’ve already lost
I’m serious
the audience has been conditioned to scroll past anything polished. they assume it’s a rug.
make it ugly. make it funny. make it look like a person made it at 2am.
that’s the playbook. there is no other playbook.
English
- Decentralised Infra no one company controls the network
- 100% Uptime
- TEE and Confidential Compute with programmatical guarantees
- End to End Encryption
- 60+ top of the line Open Source Models
- Pretty much the cheapest Inference you can find
- Built on Bittensor $TAO
What else could you possibly ask for?
Florian S@airesearch12
ever thought about the advantages of decentralized inference providers? @chutes_ai
English
Not really lol, data is pretty much correct
While we did drop of off the emissions list (stupid TAOflow)
fact is we did our all time high in terms of revenue on March 25th 🤷🏻♂️
It’s all up on our revenue page for the public too
chutes.ai/app/research/p…
While some subnets just burn miner emissions, Chutes actually has real revenue and customers
Which is growing by the day 🪂
English
@Karamata2_2 @chutes_ai @rayon_labs @bittensor No they dropped out of the top emisisons list you’re looking at old data
English
🔥 @chutes_ai (SN64) is one of the standout subnets in the Bittensor ecosystem that is already running real flow.
This is a subnet built by @rayon_labs currently leading @bittensor ( $TAO) in emissions with around 14.4% of the entire network, equivalent to roughly $52M per year in TAO subsidy. But what matters more is not the emissions, it is the actual usage.
SN64 operates as a serverless AI inference layer. Developers can deploy and run models like:
- Llama
- DeepSeek
- Qwen
- Mistral
- diffusion
- speech
- and embeddings
through simple API or CLI access, without touching infrastructure, GPUs, or scaling complexity.
The product is straightforward: lower cost, high performance, and immediate usability.
The numbers reflect that clearly:
- Processed over 9.1 trillion tokens cumulatively
- Throughput reaching over 100 billion tokens during peak periods.
- Serves over 400,000 users, including more than 100,000 API users.
Pricing is another key factor: In many cases, inference costs on Chutes are approximately 85% cheaper than AWS and still 10–50% cheaper than providers like Together AI. That pricing advantage is driven by its supply-side incentive model.
Chutes has a public revenue dashboard and a clean flywheel when Miners provide real GPUs:
-> compete on inference
-> devs pay
-> revenue + $TAO rewards loop back
The strongest part of Chutes is the incentivized supply side via TAO.
Miners compute and compete aggressively on speed, uptime, and cost, creating a supply floor that VC-funded infra struggles to match long-term.
They stay early with new models (DeepSeek V3, Qwen3…), support TEE for privacy,
and the design is highly composable -> fits well with future AI agents.
English
Fezicles.eth retweetledi
TIP: If you want private inference where even the provider can't see prompts or completions, you can use TEE-supported providers like @PhalaNetwork or @chutes_ai
Many providers also provide contractual guarantees about privacy, but TEEs will provide programmatic guarantees.
soulman 🎮@Web3GameMaster
Throughout March, models running through @PhalaNetwork on OpenRouter have been processing over 1 billion tokens per day, all with TEE-GPU guarantees. That means the inference it’s happening inside secure enclaves where your prompts and data stay private and aren’t exposed to the underlying provider. If you’re building something on @OpenRouter you can plug these models into your app with a standard API key. Agents, copilots, data pipelines, automated workflows, it all works the same way you’re used to, just with confidentiality baked in by default. no extra setup, no tradeoffs on privacy. Worth exploring openrouter.ai/provider/phala if data protection matters in what you’re building
English
AI inference should not require trust in infrastructure.
chutes.ai/news/end-to-en…
On March 2nd we shipped end-to-end encrypted transport on Chutes. Here's how it actually works under the hood.
Your data is encrypted on your machine, directly to the GPU instance running inside a Trusted Execution Environment.
It stays encrypted through our API, load balancers, and the network. Decryption only happens inside TEE-protected hardware where memory is isolated from the host.
Impossible for anyone to see including us.
The key exchange uses ML-KEM-768 — a NIST-standardized post-quantum key encapsulation mechanism. Every request gets a fresh ephemeral keypair. Forward secrecy by default. Resistant to future quantum attacks.
Full technical breakdown in the blog:
chutes.ai/news/end-to-en…
If you want to try it:
→ Python: pip install chutes-e2ee
→ Any language: docker run parachutes/e2ee-proxy:latest
github.com/chutesai/chute…
github.com/chutesai/e2ee-…
English
Best in class privacy at a fraction of a cost
What else could you possibly want?
$TAO
Chutes@chutes_ai
AI inference should not require trust in infrastructure. chutes.ai/news/end-to-en… On March 2nd we shipped end-to-end encrypted transport on Chutes. Here's how it actually works under the hood. Your data is encrypted on your machine, directly to the GPU instance running inside a Trusted Execution Environment. It stays encrypted through our API, load balancers, and the network. Decryption only happens inside TEE-protected hardware where memory is isolated from the host. Impossible for anyone to see including us. The key exchange uses ML-KEM-768 — a NIST-standardized post-quantum key encapsulation mechanism. Every request gets a fresh ephemeral keypair. Forward secrecy by default. Resistant to future quantum attacks. Full technical breakdown in the blog: chutes.ai/news/end-to-en… If you want to try it: → Python: pip install chutes-e2ee → Any language: docker run parachutes/e2ee-proxy:latest github.com/chutesai/chute… github.com/chutesai/e2ee-…
English
Fezicles.eth retweetledi
Indeed, the decrease a few days ago looked bad on the surface, but this is the "why", preparation to ramp up paygo usage. Like pulling back a slingshot.
~$22k/day rev currently, 10m ARR pace just a breath away, then...
Victor VL@Victor_crypto_2
Yesterday, @chutes_ai hit a new ATH in revenue! Same for the daily token stats Chutes processes, over the past few days they’ve been reversing the downward demand trend from February/March These past few days, we’ve seen a lot of speculation around multiple high market cap subnets (and it's great), but don’t forget who’s been carrying the network’s adoption for the past year
English
Fezicles.eth retweetledi
We're now powering inference for @GenLayer.
GenLayer is a chain where validators use LLMs to reach consensus — interpreting natural language, pulling live data, and resolving outcomes on-chain that traditional smart contracts can't.
Their validators now have access to 60+ open source models through Chutes. Scalable compute, no infrastructure to manage, powered by the Bittensor network.
Decentralized inference meets decentralized consensus.
More to come 🪂
English
Fezicles.eth retweetledi
For what it's worth:
1. the number includes validator/staker rewards which you can basically remove from the equation.
2. it also includes owner emissions, which actually get locked into a smart contract, irrelevant.
3. Assumption that every dime of emission is sold, not true.
4. Our ARR projections are much closer to 8 figures now.
So, basically fake news, but DYOR lol.
Pine Analytics@PineAnalytics
Bittensor's biggest subnet receives $52M a year in TAO emissions. It generates $2.4M in actual revenue. Without the subsidy, it would cost more than AWS.
English
@WellerOlaf @chutes_ai Much appreciated! Feel free to reach out if you have any questions btw!
English
@chutes_ai Very interesting guys! In June I will give a new webinar in the Netherlands about privacy and AI, targeted at everyday consumers. I will be sure to include your service in it.
English
Fezicles.eth retweetledi
He's not wrong.
Like BTC, except proof of "useful" work, and in which you can build a flywheel where that useful work is even paid for instead of just expense to sustain the ledger (which itself is also valuable of course).
@jason@Jason
$tao > $btc
English
@airesearch12 @chutes_ai Still not over its pfp, Whatta cutie
English
