@chutes_ai: Powering trillions of tokens monthly with the most secure open-source decentralized AI inference (TEE post-quantum E2EE)

Joined October 2024
224 Photos and videos
Pinned Tweet

5
27
91
13,805
i had a feeling. 30 mins later seems accurate 🎯 @const_reborn wen subnet shorting? i fucking hate the amount of intellectual dishonesty in SO many of our subnets. there is an insane amount of grift. we need true punishment, harsh ones - via a quorum
unfortunately every subnet is a scam until proven otherwise. you cannot just trust anyone for their word especially in this industry. have we not learned from the covenant bullshit? hold people accountable. this is why i simply chutes n chill 🪂
5
2
1,023
@peytonspencer and @omarzanji hopefully you can show me i’m wrong here.
2
198
@heydittoai would love to hear you speak on this
98
unfortunately every subnet is a scam until proven otherwise. you cannot just trust anyone for their word especially in this industry. have we not learned from the covenant bullshit? hold people accountable. this is why i simply chutes n chill 🪂
the intellectually dishonest about this is lmao “Over 1000 users already running real workflows. No ads. Just word of mouth.” is misleading then it comes off as x number of users acquired solely through word of mouth, the “no ads” aspect is bullshit the post says paid sponsorship… unless that’s a mistake
1,192
arkhet.hl retweeted

2
3
454
arkhet.hl retweeted
Fiat used to print cash. Now it prints GPUs. Either way, it all ends up flowing into the protocols.
9
44
384
46,482
arkhet.hl retweeted
I can’t wait for the day CT realises this: @chutes_ai is perhaps the only scaled confidential inference provider in the world whose GPUs can’t be switched off by a government entity. Not Venice. Built on Bittensor. $TAO
1
11
82
1,979
arkhet.hl retweeted
Replying to @_0xghost_
10
48
308
9,166
arkhet.hl retweeted
Presumably, so long as there are network participants, it would be roughly equivalent to trying to exterminate all ants on planet earth. Good luck.
9
18
126
5,222
arkhet.hl retweeted
Access to intelligence should not depend on a handful of companies or governments. This is why open, decentralized, permissionless AI matters. This is why Bittensor matters.
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees. The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance. Access to all other Claude models is not affected. We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible. Read our full statement: anthropic.com/news/fable-myt…
68
364
1,234
77,044
arkhet.hl retweeted
Curious about what this Bittensor $TAO thing is all about? Take some time this weekend to watch The Incentive Layer, a new documentary by @evert_scott youtu.be/71rvASmXUN8?si=nKIT…
106
334
1,382
190,408
day 5 of posting an update every day till convergance
day 4 of posting an update every day till convergance
9
183
🪂 Chutes AI serves quantum resistant inference 🤺
Replying to @arkhet
Here’s an idea of how they’re doing real ML-KEM-768 encapsulation client-side. You grab the instance’s ML-KEM pubkey (generated inside the TEE at startup), make your own ephemeral keypair, encapsulate to get the shared secret, derive keys with HKDF-SHA256, then encrypt the prompt with ChaCha20-Poly1305. It stays ciphertext through the whole API path and only decrypts once it’s inside the Intel TDX NVIDIA CC GPU. Responses come back the same way with fresh keys each request, so you actually get forward secrecy. Nonces are single-use with atomic enforcement too.
3
21
783
arkhet.hl retweeted
Chutes is running post-quantum E2EE inference in production. ML-KEM-768 key exchange client-side encryption that only decrypts inside TEE GPUs. Not even the provider can read your prompts. While everyone else is still on classical crypto, Chutes is already quantum-resistant by default. This is the bar now. @chutes_ai 🪂
2
13
60
1,981
arkhet.hl retweeted
The Battle against Centralized AI: Florian Standhartinger
Replying to @arkhet
youtu.be/BIsgQGYHze8 watch this for an insight
1
5
380
.@chutes_ai is the most secure open-source AI inference (TEE post-quantum E2EE). Powering trillions of tokens monthly 🪂
2
22
244