Today’s episode is wild. Robots failed to pass butter. OpenAI dropped new safety models. And Meta’s AI ad tool swapped a model for a grandma. It’s funny, weird, and kind of important.

We’ll talk about:

• OpenAI’s new “open(ish)” safety models and why they matter
• Butter-Bench: the robot test that exposed AI’s physical world struggles
• Meta’s AI ad tool going totally rogue
• Why Grammarly rebranded to Superhuman (and what it changes)

Keywords: GPT-5, Claude, Gemini 2.5, Meta AI, OpenAI safety, Butter-Bench, Superhuman, ChatGPT ads, AI tools, AI fails

Links:

1. Newsletter: Sign up for our FREE daily newsletter.
2. Our Community:  ...