Two new services, Engy and SayGM, aim to address trust issues in using hosted AI models. Engy focuses on verifying that the specific model and its version being used are indeed the ones paid for, using Merkle roots and activation proofs to ensure model integrity without relying on trusted hardware. SayGM, on the other hand, provides a privacy-focused gateway that uses Intel TDX trusted execution environments to shield prompts from the gateway operators and host machines, offering end-to-end encryption for certain models. AI
IMPACT These services could improve developer trust and privacy when using third-party LLMs, potentially influencing how AI agents are built and deployed.
RANK_REASON The item describes new services that offer solutions to existing AI model usage problems, rather than a novel model release or research.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
