An open-source “AI agent” (yes, the lobster one) didn’t just go viral — it helped push infrastructure conversations into the mainstream, triggered copycats/scams, and raised serious security questions for any organisation experimenting with agentic AI.

In this episode, FirmBrain hosts Jerrie Yang and **Jessica Wu break down:

• What the Clawdbot/Moltbot/OpenClaw project actually is (“AI that acts,” not just chats)

• Why the market paid attention — and what it signals about enterprise AI infrastructure

• The dark side: exposed instances, prompt injection, and why “useful” can also mean “dangerous”

• The scam wave: fake tokens and fake extensions riding the hype

Who this is for: CTOs, IT/security leaders, founders, and business leaders trying to understand why “agentic AI” is suddenly everywhere.

Key takeaway: We’re moving from AI that advises to AI that acts — and that shift demands new thinking about governance, access controls, and auditability.

Listen on Spotify and share it with a colleague who’s evaluating AI agents this quarter.

(General information only — not financial or security advice.)