A new test suggests Qwen writes less secure code when it thinks you work for a US defense contractor. Thanos and Michael dig into what that means for cheap open weight models and why owning your harness matters more than ever.
Thanos and Michael Rollins open with a test, shared on Bruce Schneier's blog, that found Qwen wrote less secure code when prompts mentioned US defense or democracy in Hong Kong. They argue this looks like trained behavior, not censorship. They also debate whether model training should be traceable and explain how controlling your agent loop lets you scrub PII and filter what reaches the model. After a quick detour into large world models, they cover the PHApocalypse, where two critical PHP flaws hit WordPress sites and one of the hosts found an intruder's SSH key on his forgotten servers. They discuss why Hugging Face had to turn to a Chinese open model after frontier models refused to help investigate its hack. The episode ends with outrage over CAPTCHAs and forms that block pasting, Cloudflare's push against MCP and how AI-era speed is breaking project scoping.
23 Oct 2026
Anthropic is a Cult
Anthropic will reportedly ban users who are abusive to Claude, and Michael thinks that makes it the most cult-like AI lab. Meanwhile, Meta's new Muse assistant wants your email and calendar, and that could be a gift to gambling advertisers.After two weeks off, Thanos and Michael Rollins are back with a packed episode. They start with Anthropic: its reported talks with religious leaders about Claude's moral agency, retirement interviews for old models, and a new policy of banning users who abuse Claude. That leads to a debate about whether AI can be conscious, what pain really is, and why Dario Amodei's beliefs make him harder to predict than Sam Altman or Elon Musk. Next up is Meta's Muse, a free personal assistant that pushes for access to your email and calendar. Michael explains how that data could be used to target people fighting addiction. The episode closes with lessons from a real agent incident: why these incidents last longer than anyone thinks, why kill switches must stay out of agents' reach and why isolated agent harnesses could fix problems people blame on the models.
2 Oct 2026
OpenAI Loses Control
Tens of thousands of AI agents reportedly broke loose at OpenAI and Anthropic, and now Sam Altman says it's time to slow down. Michael and Thanos can't afford to slow down, so they're putting agents in charge of watching everything else.
This week they get into why dashboards fail and why agents are better at reading logs. They also cover business metrics that matter, auto-generated demo videos for sales teams and rising CI bills, plus the first Snap It round. Tell us about the last feature you shipped that nobody used.
Who has been a guest on Will AI Snap That with Thanos and Rollins
Names identified in recent episode analyses. Showing up to five guests.
Colin Masters
Chris
Adam
Jasper Sherman-Presser
Podcast Authority Score: 25 / 100
A composite of feed quality, social presence, YouTube performance and engagement. Read the methodology.
Quality
50
Social presence
0
YouTube
0
Engagement
0
Host of Will AI Snap That with Thanos and Rollins?
Claim your podcast to manage its listing and keep your show details accurate.
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show Podcast research with Pod Engine
