TEE security discussion heating up:
@zama showing Battering RAM/WireTap exploits on Intel SGX (~$50 hardware cost for memory aliasing attacks)
Meanwhile, 4 new TEE implementations launched this week: 🧵

• @tenprotocol: 500M+ testnet transactions in SGX/TDX
• @oasisprotocol ROFL: Off-chain compute in Intel TDX
• @zk_agi: SGX/SEV for federated training
• Cap's AVS: Certora audit complete (no critical issues)
The debate: Math-based proofs (FHE, ZK) vs. hardware trust (TEE enclaves)
@Gaianet_AI AVS approach: Cryptoeconomic validation layer
✅Not choosing between TEE and proofs
✅Economic slashing for incorrect verification
✅Works with any compute substrate

Built on @eigenlayer restaking. Validator set securing inference quality through staked capital, not just attestations.
Live with real validation data. Not picking sides in FHE vs. TEE - we're adding economic accountability to both.
1,894
40
本頁面內容由第三方提供。除非另有說明,OKX 不是所引用文章的作者,也不對此類材料主張任何版權。該內容僅供參考,並不代表 OKX 觀點,不作為任何形式的認可,也不應被視為投資建議或購買或出售數字資產的招攬。在使用生成式人工智能提供摘要或其他信息的情況下,此類人工智能生成的內容可能不準確或不一致。請閱讀鏈接文章,瞭解更多詳情和信息。OKX 不對第三方網站上的內容負責。包含穩定幣、NFTs 等在內的數字資產涉及較高程度的風險,其價值可能會產生較大波動。請根據自身財務狀況,仔細考慮交易或持有數字資產是否適合您。

