https://t.co/itb4CexZlC
Everything was mentioned here and i actually did the same exact thing.
- It was my followers who wanted me to launch
- I made nothing so far as everything is onchain
- If someone lost something then it’s their decision where to click on these random tickers
- And there are people who made the money as well and this is how trading works
- If someone is buying at the same time someone is selling
- And if you’re scared of losing money then you shouldn’t do trading and never be the part of bigges casino on the earth that is “SOLANA MEMES”
Stop being a hater & talk real shit.
- Thug life $GANG
/arc_angels
arc is teaming up with @AbstractChain — an innovative l2 in crypto focusing on consumer apps to bring evm compatibility to rig — our rust-based ai agent framework.
together, we’ll foster new devs building ai-powered apps on abstract. stay tuned for a new era of on-chain ai.
we build rig, rig builds us
/arc_angels
we’ve partnered with @MongoDB to bring flexible persistence to rust-based agents built with rig. this integration lets developers store, search, and manage vector embeddings on MongoDB’s production-grade database—unlocking powerful retrieval-augmented ai for real-world apps.
we build rig, rig builds us
Hey everyone,
Sorry that took way longer to type than I anticipated
It's time to change security engineering forever. Grab your team or coffee, this is worth your time...
For those that aren't aware, I've been building a ton of new infrastructure and systems to match, and today I'm finally able to start talking about it in-depth.
Enter: The Mizuki Framework
The Mizuki Framework is a groundbreaking system that lets developers train their own security AI models on ANY security rules, for ANY programming language. Think about that for a second - your rules, your language, your security model.
But here's where it gets wild: These aren't just basic models. We've engineered a framework that exports to both ML and ONNX (Open Neural Network Exchange) formats. Why does this matter?
- .NET developers can run these models locally
- Web developers can plug them into existing agent frameworks
- Any AI agent can become a security expert through ONNX inference
You can train these models on a basic laptop. No expensive GPU farms. No cloud dependencies. No vendor lock-in. We're democratizing security AI from the ground up.
Take a sip of your tea and le that sink in: Every developer, every company, every open-source project can now train and deploy custom security AI models. The barriers to entry are practically gone with the tooling that will be built from this framework.
The Mizuki agent that's been turning heads on Twitter? That's just ONE application. It's using our security model for inference alongside an LLM. Mizuki stands for more than just what you've seen and interacted with. It's an entire ecosystem.
Want to swap out models on the fly? Done. Models are hot swappable.
Need to update training data? Easy. Simply point the framework to your preferred training config.
Want to build your own security framework? Go for it.
We're making this completely open source because we believe security AI shouldn't be gatekept. The ONNX bridges we're building mean any developer can integrate these models into their agent frameworks - This includes @ai16zdao's @elizawakesup agent and framework for example.
Let me blow your mind with just ONE example of what's possible:
Imagine a Copilot-style tool for security, but running completely locally. As you code, it analyzes every line in real-time using Mizuki models. No cloud calls. No latency. No cost.
Your IDE becomes a security expert that:
-Instantly flags security issues
-Suggests secure alternatives
-Runs at the speed of your typing
-Works offline
-Knows YOUR specific security rules
And because it's using locally-running Mizuki models:
- Zero latency
- Complete privacy
- No subscription fees
- Custom-trained on your standards
- Fully autonomous
I wanted to tweet this so that everyone is aligned with where we're going next. We have some massive updates coming, working with some of the best and brightest in the industry to make this available to everyone.
The diagram below should give you a visual idea of how the framework will allow you to train models which get exported to the two formats (ML & ONNX), which can then be hosted directly on the client and used for local inference if required, or hosted over an API for any cloud based application
Thanks for reading this and your time. I really appreciate the continued support!
Stay safe!
Choose your favorite: $CLANKER, $FAI, $GAME, $TAOCAT, $TRISIG. And lets vote for the Next Gem on KuCoin!
⏰ Don’t wait—voting ends Jan 9, 2025 (UTC).
Push Your Favorite Project to the Top on KuCoin!
$ALGO $TRX $LINK $COOKIE $AAVE $LTC $HBAR $AVAX $BNB $DOT