Kaifovo (@kantoloop) profile picture

$quil and chill.

Texas, USA
Joined May 2017
What happens to the student's first app, or the privacy tool built by a volunteer? Google's mandatory registration blocks them all by default unless they play by Google's rules. keepandroidopen.org #KeepAndroidOpen
1
13
“let’s make a new state free from tyranny!” >adopts fascism
14
10
3
142
6,613
The new owner of farcaster, @balajis, has banned me from their side of the fork. Absolutely incredible work. So anyway, we're doing the hard fork of farcaster and would love to hear name ideas for our non-cancerous side. Spinning up an emergency group chat: t.me/+4eYOApUSw500Yzgx
89
74
23
494
30,859
Kaifovo retweeted
Decentralized social is important. Yesterday's actions by Farcaster, preemptively banning long-serving community members as an AI decided they "didn't fit the Singaporean citizen model" is frankly insane - and alongside being an example of AI used with the worst possible intentions, it also added to the importance of actual decentralized social. We're working with our friends within the Farcaster Community (the real one) to lend our dev resources towards ensuring the continuity of the protocol under HyperSnap & Quilibrium, spearheaded by @cass_on_mars. We believe this symbiotic relationship will be beneficial to both our community at imgnAI, as well as to the Farcaster community - and the large overlap of users active on both of our platforms. We all believe that decentralization is the way forward - and we want to ensure that we're doing everything in our power to keep decentralized social and its talented pool of builders & creators live and kicking.
14
47
8
184
8,040
Kaifovo retweeted
⚠️ ONGOING ATTACK ON COLDCARD - ACT NOW ⚠️ If you're a COLDCARD Mk3 user (your coldcard has no USB-C yet) you must MOVE FUNDS off the device RIGHT NOW. Seed phrases generated on the COLDCARD Mk3 are insecure. Move all funds to a secure hardware wallet. Call your trusted Bitcoiner if you need help. This is not a drill. I know people affected by this. Share widely.
COLDCARD Mk3 Security Advisory If you generated a seed on a Mk3 after firmware 4.0.1, your funds may be at risk. Mk4, Q and Mk5 are not affected based on our early analysis. Read the advisory and migrate carefully: blog.coinkite.com/coldcard-m…
29
140
22
518
79,923
Kaifovo retweeted
Legend. Hang this goblin 🚬
HOLLLLLYYYYY SHHHHIITTTTT
2
2
18
4,800
Kaifovo retweeted
🇪🇺 Chat Control has passed 😔 They can and will now legally scan any person's messages, emails and photos you send without a warrant The way they passed this law when the majority of the European Parliament was against it will shock you: They waited until most EP members were on holiday so only a few were present and then created an "urgent" vote for it to pass it through There's nothing democratic about any of it and big powerful forces are behind this that can manipulate the EU for whatever they want Democracy in Europe died a bit today 😔
Replying to @gandalv7
Illegaler Geschäftsordnungstrick der Parlamentspräsidentin - Eilantrag, der die Qualifizierte Mehrheit (360 Stimmen) GEGEN die Chatkontrolle gebraucht hätte, an einem Tag, an dem nur knapp 600 MEPs vor Ort waren, wegen Urlaub etc…
246
3,977
638
22,424
3,362,686
Starting September 2026, Google will block any Android app whose developer hasn't registered and provided government ID. This affects every Android device worldwide. Learn more: keepandroidopen.org @AlteredDeal #KeepAndroidOpen
3
165
MetaVM has been released. Prove execution of RISC-V, EVM, and Solana sBPF in ZK. Supports Quilibrium's BLS48-581 and Ethereum's BLS12-381 natively, without needing a GPU. MetaVM's RISC-V compatibility is the first full ZK RV64IMAC instruction set – you can run Linux in MetaVM, and prove everything that happened within the VM session, prove a block on Ethereum, or a slot on Solana, and emit proofs compatible with either Q or Ethereum. github.com/QuilibriumNetwork…
7
50
8
163
19,444
Kaifovo retweeted
🧵Klearu: A Strategic Building Block for the Era of Autonomous AI Agents 1️⃣ The Next Wave: Autonomous AI Agents The next revolution won’t just be chatbots. It will be autonomous agents capable of: • trading • arbitraging • managing logistics flows • executing contracts • coordinating machines • interacting economically with each other These agents will require: 👉 crypto payments 👉 programmable identity 👉 distributed compute 👉 native privacy An agent dependent on a centralized API can be: • rate-limited • shut down • monitored • blocked An agent running on permissionless infrastructure becomes: • resilient • autonomous • economically sovereign This is the context in which Klearu becomes strategic. 2️⃣ What Klearu Actually Is Klearu is not a new LLM. It is an infrastructure layer enabling: 👉 private AI inference 👉 on remote nodes 👉 without the server seeing the data The server computes… but does not “read.” This relies on: • sparse architectures (only part of the model activates) • hash-based routing (SLIDE) to efficiently select relevant neurons • 2PC protocols to preserve confidentiality 3️⃣ Why This Is Strategic for Quilibrium Quilibrium is not building a chatbot. It is building infrastructure: • massive sharding • peer-to-peer networking • native privacy • MPC and programmable permissions Klearu becomes the AI engine capable of running within that architecture. This aligns perfectly with Quil’s philosophy: no AWS dependency, no single control point, no data leakage. 4️⃣ Toward the Democratization of AI Compute Today: Advanced AI = rare GPUs + hyperscalers. Klearu explores another path: making inference viable on optimized, distributed CPUs. If this works at scale: • lower barriers to entry • more participants • less concentrated compute • less oligopolistic AI We move from a closed model to a network model. 5️⃣ A Piece of a Larger Puzzle Klearu is not the end goal. It is a deep infrastructure layer. Its value depends on: • the rise of autonomous agents • the growing need for private AI • adoption of permissionless infrastructure Short term: technical. Mid term: strategic. Long term: potentially foundational. If machine-to-machine economies truly emerge, infrastructure capable of supporting sovereign AI execution will become critical. And that is precisely where Quilibrium aims to position itself.
Today, we are publishing one of the side tracks of research ongoing with Q, our E2EE ML training and inference library, klearu: github.com/QuilibriumNetwork… SLIDE proved that hash tables can beat GPUs at training deep networks. Further works compounded on this, and Klearu is the first native Rust implementation built on top of this research, extending it to LLM inference, sparsity prediction, and private two-party computation. In the current days we're seeing deeper trust being placed on AI, while the largest of providers are collecting this data for the purpose of not only training, but also advertising, or even selling this data to others. The risks grow worse with every passing day. The majority of AI research for private AI exists in the form of using TEEs – but we've seen time and time again that using TEEs for privacy is disastrous, guaranteed to leak, and even by it's name, is a massive requirement of trust. Outside of this, other private AI looks towards FHE. We know, at least for the near future, that FHE cannot perform at a speed high enough to be generally useful. So instead, we adopted 2PC, with flexible security configurations, where users can be assured that their requests remain private. The majority of these research projects have strictly an output of papers, with no or limited real world instances of their use. Klearu's implementation is available now, with simple instructions for developers to try it out.
2
10
40
1,246
Today, we are publishing one of the side tracks of research ongoing with Q, our E2EE ML training and inference library, klearu: github.com/QuilibriumNetwork… SLIDE proved that hash tables can beat GPUs at training deep networks. Further works compounded on this, and Klearu is the first native Rust implementation built on top of this research, extending it to LLM inference, sparsity prediction, and private two-party computation. In the current days we're seeing deeper trust being placed on AI, while the largest of providers are collecting this data for the purpose of not only training, but also advertising, or even selling this data to others. The risks grow worse with every passing day. The majority of AI research for private AI exists in the form of using TEEs – but we've seen time and time again that using TEEs for privacy is disastrous, guaranteed to leak, and even by it's name, is a massive requirement of trust. Outside of this, other private AI looks towards FHE. We know, at least for the near future, that FHE cannot perform at a speed high enough to be generally useful. So instead, we adopted 2PC, with flexible security configurations, where users can be assured that their requests remain private. The majority of these research projects have strictly an output of papers, with no or limited real world instances of their use. Klearu's implementation is available now, with simple instructions for developers to try it out.
10
60
15
147
19,780
Replying to @QuilibriumInc
@QuilibriumInc just dropped Klearu - an open-source runtime for private LLM inference. 🤖 +🔒 GitHub: github.com/QuilibriumNetwork… Most "private AI" today relies on TEEs - basically hardware enclaves where you have to trust the manufacturer isn't peeking. The real alternative (fully homomorphic encryption) is still too slow to be practical for AI workloads. 🤖 You can ask Quily chatbot questions about Klearu: quily.quilibrium.one/ Klearu takes a third path: two-party computation (2PC). When you send a query to a remote node, the computation happens in a way where neither side ever has the full picture - the node runs the inference but mathematically cannot see your inputs or outputs. No trust required, just cryptography. A few things worth knowing: • CPU-native, no GPUs needed • Supports LLaMA-compatible models (up to 7B+ params) • Built-in sparsity optimizations that skip unnecessary computation per token • Speed is comparable to llama.cpp in lower security mode This is the kind of infrastructure that makes "private AI on a decentralized network" mean something concrete rather than a marketing claim. Tech docs and benchmarks by @LookDragi : qstorage.quilibrium.com/info… $QUIL
Today, we are publishing one of the side tracks of research ongoing with Q, our E2EE ML training and inference library, klearu: github.com/QuilibriumNetwork… SLIDE proved that hash tables can beat GPUs at training deep networks. Further works compounded on this, and Klearu is the first native Rust implementation built on top of this research, extending it to LLM inference, sparsity prediction, and private two-party computation. In the current days we're seeing deeper trust being placed on AI, while the largest of providers are collecting this data for the purpose of not only training, but also advertising, or even selling this data to others. The risks grow worse with every passing day. The majority of AI research for private AI exists in the form of using TEEs – but we've seen time and time again that using TEEs for privacy is disastrous, guaranteed to leak, and even by it's name, is a massive requirement of trust. Outside of this, other private AI looks towards FHE. We know, at least for the near future, that FHE cannot perform at a speed high enough to be generally useful. So instead, we adopted 2PC, with flexible security configurations, where users can be assured that their requests remain private. The majority of these research projects have strictly an output of papers, with no or limited real world instances of their use. Klearu's implementation is available now, with simple instructions for developers to try it out.
7
1
44
960
After seeing that Discord plans to make KYC mandatory, we want to be clear on where we stand. KYC for messaging is a complete non starter for us. Quorum Messenger does not ask for your personal details and it never will. No identity checks, no sign ups tied to who you are, and nothing happening behind the scenes. Quorum is decentralized by design, and both DMs and group chats are fully end to end encrypted. That means only the people in the conversation can read the messages, not us, not a third party, not anyone else. Your conversations are yours. You own them, you control them, and that is how it is designed to stay. Privacy is not an add on or a talking point for Quilibrium, it is the whole point. So if you are looking for a safe haven messenger that will not turn its back on you, come try Quorum Messenger 👇 quorummessenger.com/
Quorum Mobile Beta is officially live 📱 If you signed up for early access, you’ll be receiving an email shortly with download links for iOS and Android. We’re excited to finally share the first mobile version of Quorum with the world.
16
62
10
165
12,470
Kaifovo retweeted
🧵 @QuilibriumInc $QUIL Why the Bear Market Is an Opportunity 1️⃣ A bear market is not a pause => it’s a filter In a bear market, speculation fades. Noise disappears. Attention returns to fundamentals. For an infrastructure project, this environment is ideal: 👉 less pressure, more focus, better decisions. 2️⃣ Quilibrium is infrastructure, not a narrative Quilibrium is not built around marketing or storytelling. It is a full technology stack: 🔹️base network & hypergraph 🔹️native privacy 🔹️storage 🔹️key management 🔹️messaging & event systems This kind of system cannot be rushed. It must be built deliberately, away from noise. 3️⃣ What internal discussions reveal: engineering over spectacle Quilibrium discussions focus on: ✔️consensus ✔️synchronization ✔️network invariants ✔️hard technical tradeoffs Releases are delayed when necessary. Decisions prioritize correctness over optics. 👉 This is the behavior of teams building for durability. 4️⃣ The upcoming production launch is for builders The production launch happening in the coming days is not about mass adoption yet. It marks: ➡️real availability of core primitives ➡️SDKs, APIs, and tooling going live ➡️developers beginning to build on a stable base This is the moment infrastructure becomes truly usable. 5️⃣ Quilibrium follows the only sequence that works The order is always the same: ▫️1. infrastructure ▫️2. developers ▫️3. applications ▫️4. users Quilibrium is now at the critical transition from 1 → 2. This step is invisible to most => but decisive. 🧠Conclusion => The Beginning of the Quilibrium Revolution The imminent production launch is not an end point. It is a starting point. 👉 The beginning of a new infrastructure layer 👉 The beginning of a truly privacy-first ecosystem 👉 The beginning of the Quilibrium revolution Now the foundation is ready. Builders are next. Everything else follows. Be early, Be free.
Hey Q fam! Here is a summary of the latest live stream with @cass_on_mars (@QuilibriumInc founder). ✨ Quilibrium Dev Update: Q Services Launch & QKMS Deep Dive Major announcement: QConsole with QStorage, QKMS, QQ (SQS-style queues), and QPing launching within 24 hours. Plus updates on node synchronization, Quorum Mobile, and QNS. ▶️ Watch it on YouTube (English captions): youtube.com/watch?v=lQI4K8Bz… ▶️ Watch it on X:teapawt.net/QuilibriumInc/status/2… ✅ Key topics: ⦿ Version 2.1.0.18 update progress and final synchronization blocker ⦿ QConsole services launch: QStorage, QKMS, QQ, QPing ⦿ True MPC (Multi-Party Computation) key management explained ⦿ Quark SDK demo with tokenized 3D game assets ⦿ Quorum Mobile and QNS updates ✅ Node Update: Version 2.1.0.18 The team has been working on resolving issues around the 2.1.0.18 release. There is one remaining blocker related to synchronization - specifically getting the CRDT-style trees that power the hypergraph to align properly between nodes requesting data (client side) and nodes providing it (server side). The latest commits include additional debug details to help pinpoint where differences are occurring. Once resolved, this update will roll out to all node runners and will fix the seniority score display issue that some users have been experiencing. Shard coverage continues to increase, halt risk is decreasing, and needs coverage has also decreased since the previous broadcast. ✅ Frame Production Issues Resolved Three categories of frame production issues have been fully resolved. First, one of six archive nodes went offline when the operator chose to stop running it. With the two-thirds consensus requirement, this created instability, but a new operator has stepped up and the node is back in the fold. Second, subtle changes to consensus logic had been creating halt conditions, which have now been addressed. Third, sync issues were causing consensus participants to disagree on the network state. This has been resolved, though general sync improvements continue. ✅ QConsole Services Launching Multiple AWS-compatible services are launching within 24 hours once billing tests complete. QStorage provides S3-compatible object storage with event notifications support. QKMS is a Key Management Service supporting symmetric keys, asymmetric RSA, asymmetric elliptic curve (including SECP256K1, ED25519, and DCAF-448), HMAC, and SM2 for Chinese markets. QQ offers SQS-style message queue service with nearly 100% API compatibility. QPing provides event notification service similar to SNS. All services integrate with Quilibrium's privacy primitives and support the same SDK interfaces developers already know. ✅ True MPC: Why It Matters Cassandra explained how QKMS differs from most "multi-party wallet" services on the market. Many competitors claim MPC but actually generate the full key in one place, then split it using Shamir's secret sharing. When signing, they recombine key shares into one location (either user device or trusted enclave). This means the full key exists somewhere at some point - making whoever holds that moment the custodian. Quilibrium's approach uses true threshold MPC. Key shares are generated through DKL-S18 (for 2-of-2) or higher threshold variants (for 3-of-5, etc.), and they never exist in the same place. Signing uses oblivious transfer protocols where the key material is never combined. The system supports multiple curves including SECP256K1 (Bitcoin/Ethereum), ED25519 (Solana/Farcaster), and DCAF-448 (Quilibrium stealth signing). This approach is truly non-custodial and compliant with regulations like those from the New York Department of Financial Services. ✅ Quark SDK Demo A demonstration showed the Quark SDK enabling tokenized game assets. The demo featured 3D avatar assets stored on QStorage with RDF schema validation and in-game currency backed by Quilibrium token primitives. Users can purchase and own avatars as private tokens, with asset ownership being completely private - the game cannot see your wallet or what you own. The system uses view and spend keys visible only to the user, not the game. This enables privacy-preserving digital ownership for games and virtual worlds - unlike public blockchain approaches where wallet contents are visible to everyone. ✅ Quorum Mobile Update An update is coming shortly after the QNS update (24-48 hours). The new version will support more L1 chains in the wallet, including Monero and Zcash, along with deeper integration with QNS features. Users will also get the option for full device sync, allowing them to choose whether to sync just preferences and profile information or their complete chat history. Various bug fixes for quality of life issues reported during beta are also included. ✅ QNS Update QNS (Quilibrium Name Service) is getting an update within 24-48 hours. The update includes improvements to auctions and offers features, deeper integration with Quorum Mobile, and the ability to message name owners using the underlying dispatch primitive. ✅ What is Hypersnap? A community member asked about Hypersnap. While the full explanation was not covered in detail during this stream, Cassandra mentioned that the Hypersnap page describes its features. The core principle relates to Quilibrium's privacy-first architecture - everything on Quilibrium is private by default, and Hypersnap relates to how this privacy is maintained when interacting with external systems. ✅ QClient Update QClient has not been released yet due to ongoing changes to protobuf messages used on the network. Releasing it now would create support issues as these messages are still being refined. Once the synchronization and consensus updates stabilize, QClient will be made available. ✅ Coming Soon QConsole public launch with all services is expected within approximately 24 hours. The Quark SDK will also be published. QNS update is coming in 24-48 hours, followed by the Quorum Mobile update. Version 2.1.0.18 for node runners will be released once the sync issue is resolved. $QUIL
1
6
1
58
1,937