Omer Goldberg
@omeragoldberg • 10,975 subscribers
founder @chaoslabs | eng @instagram, @meta, https://t.co/0TaNkknFrg | cs @TechnionLive
Videos
Chaos AI analyzed the vaults impacted by the USR exploit. A ton of interesting weekend transactions. Here, we build a knowledge graph, filtering for the Gauntlet USD Alpha Vault and the Gauntlet Resolv USDC vault: Pre-exploit: • USD Alpha had allocated ~438,440 USDC to the Gauntlet Resolv USDC vault • That vault was deposited into the impacted Morpho Resolv markets via the Gauntlet Resolv USDC vault. • USD Alpha held receipt tokens representing this exposure Post-exploit, March 23: • 00:30 UTC, Gauntlet USD Alpha sends ~438,440 USDC worth of resolvUSDC receipt tokens (~405,439 shares) to a Safe. • 00:35 UTC, Gauntlet USD Alpha receives USDC ($438,401) from Coinbase. Receipt tokens out, USDC in, 5 min apart, w/ roughly the same notional. A ton more to dig into. Analyzing and contrasting vault curator operations and allocation patterns in real time is one of the use cases we’re building Chaos AI for.
Omer Goldberg34,883 次观看 • 3 个月前
Chaos AI zero-shotting a 19-step sequence to open a 6x levered stETH/ETH loop on Aave V3 E-mode. In a sea of crypto x AI hype, the most differentiated and valuable use case is building agents that can act fully autonomously. A wallet, capital, and permissionless access to any onchain financial instrument/asset class are unique to our industry. In closed beta. DM me or sign up for early access.
Omer Goldberg19,864 次观看 • 3 个月前
The best change to 10x Claude Code is adding verification skills. Verification is domain specific; here's ours. Chaos Labs is a financial portfolio agent. I spin up users that simulate user/agent convos. With cc hooks, this runs on every pr. Code gen needs verification!
Omer Goldberg20,974 次观看 • 5 个月前
Most of the crypto × AI narratives you hear are backwards. They start with a conclusion and work their way to a demo. The real story is simpler and more interesting. LLMs proved that intent can be compiled into code. You describe an outcome in natural language, and software figures out the implementation details. DeFi proved something orthogonal but just as important: Code can directly express and execute financial actions. Lending, borrowing, repayment, and liquidation; not as products, but as callable primitives. Put those two facts together, and you get a new class of system: GenFinance. Where intent itself becomes the instrument. This only works onchain. Not because blockchains are trendy, but because they expose every financial primitive as a permissionless API. There is no equivalent environment in TradFi. Once you see this, many “UX problems” in finance reveal themselves as coordination failures. We’ve been forcing humans to operate software because software couldn’t reason about states, risks, or intents. The video below shows a concrete example of this: A Chaos agent takes a user’s natural-language intent to go long on GMX 🫐, inspects the user’s wallet state, and generates the required sequence of actions to make that position possible. For this wallet, it meant: - borrowing on Aave - bridging from Ethereum -> Arbitrum - finally, going long ETH on GMX 🫐 The result isn’t a nicer interface, but rather the removal of the interface entirely. This is what finance looks like when it’s native to software: Not something you operate, but something that operates on your behalf.
Omer Goldberg19,667 次观看 • 5 个月前
没有更多内容可加载