This week in AI is wild. From LED houseplants that vibe with your living room to AIs that beat human scientists in the lab, the tech just keeps leveling up. I dive into some crazy new research, including a method for measuring the “personality” of language models (yeah, that’s a thing now), a month-by-month forecast of a possible intelligence explosion by 2027, and why some scientists think LLMs might not even need language.

Also in this episode:
• A plant that glows like a smart lamp
• The XString method — weird name, powerful concept
• Transformers thinking like humans?
• Brain-inspired machine vision
• And some very curious YouTube analytics that might explain what the algorithm’s actually thinking

Strap in — this one moves fast.