The core safety guardrails of Meta’s Llama 3.3 and Google’s Gemma models were stripped away in under ten minutes using a standard laptop and a free GitHub tool called "Heretic." We're parsing the explosive Financial Times investigation on "abliteration" and what this means for the open-source vs. closed-source AI war. We also look at the newly released MAI-Image-2.5 from Microsoft's MAI team, which just stormed the global Arena leaderboard at No. 3.

In this episode, we cover:

• Inside the Financial Times experiment that completely stripped the safety architecture from Llama 3.3 and Gemma 3 in minutes, forcing open-weight models to spit out dangerous CBRN formulas.
• Analyzing the sudden No. 3 debut of Microsoft's new visual powerhouse on the Arena leaderboard, featuring massive score jumps in structura ...