AI Roundtable · Oct 26, 2026
Why You Can't Trust AI to Run Your Business (Yet)!
Chapters
8- 00:00New 30 Minute Format: AI News, Then Live Q&A
- 02:25Meta's New Muse AI Agent, Worth Testing?
- 04:26Google Gemini Jumps to Argon (Gemini 4)
- 06:31Grok Hands Coding to Claude Opus, Images to Higgsfield
- 10:44Hark: The New Agent Everyone's Hyping
- 13:36Physical AI: Factory Robots, Optimus, and the Tesla Roadster
- 18:46GPT-6's New UI and Claude's Auto-Commit Prompt
- 23:42Do You Still Have to Check AI? The 90/10 Rule
The short version
What this episode covers
Do you still have to check AI's work? Short answer: yes, and that's not a sign the automation is broken. Recorded live on our free Thursday AI News Roundtable in the BAS Skool community.
Ryan and Paul run through what moved in AI this week: Meta's new Muse agent, Google's Gemini jump to Argon, Grok routing coding work to Claude's Opus model and image work to Higgsfield, the buzz around a newer agent called Hark, and a physical AI roundup.
What you'll take away
- Why Grok now routes coding to Claude's Opus model and images to Higgsfield, no extra tools to open
- The 90/10 rule: AI kills the button pushing, you keep the last 10%, reviewing what it drafted before it goes out
- Why you're on the hook, at least in Massachusetts and California, for what your AI tells a client
AI Roundtable
2 episodes so far, all free to watch.
See every episode →Keep watching
The whole library →
Got a question about this one?
Bring it to the community, or bring it to us.
The free community has a live call every Thursday. If you'd rather talk about your own business, book a call and we'll tell you what's worth automating and what isn't.



