X thread: "How to run AI agents without becoming a sysadmin"

7-10 tweet thread for X/Twitter explaining OpenClaw on NEAR AI as the solution to the agent deployment trilemma.

Structure:

1. Hook: The problem (cloud = unsafe, local = hard)
2. What people do now (compromise on security or convenience)
3. TEE explanation (encrypted execution, zero-trust cloud)
4. Why it matters for agents specifically (credentials, memory, 24/7)
5. What's possible now with OpenClaw + NEAR AI
6. CTA to https://near.ai/openclaw

Tone: Educational, not sales-y. For technical audience.

Deliver: Thread in text file with tweet breaks marked