My setup on GrapheneOS with all the exploit protections on except some off for apps with compatibility issues. Thoughts?

  • ZinQ@lemmy.mlOP
    link
    fedilink
    arrow-up
    1
    ·
    16 hours ago

    In NanoGPT You also got TEE (Trusted Execution Environment) models which are more private/secure from my understanding. From GPT-OSS 120B TEE:

    “TEE‑based AI models run their inference or training inside a Trusted Execution Environment (TEE), a hardware‑secured enclave that isolates code and data from the rest of the system. This provides data confidentiality, protects the model’s IP, enables cryptographic attestation of the exact model version, and satisfies regulatory privacy requirements, making AI services trustworthy and suitable for secure multi‑party or decentralized applications.” One downside is that they are usually pretty expensive to run