Robyn Anderson tests Google's new Gemini 3, calling it the smartest LLM available with PhD-level reasoning that outperforms ChatGPT 5.1. She used the accompanying Anti-Gravity platform to debug an app and get business suggestions, not just code fixes.
Chapters
OpenAI and Anthropic are probably doing some all-hands meetings this morning because Gemini 3 came out last night and it's awesome. Let me tell you about it. First of all, it's the smartest LLM out there. You can read all this later, but basically has PhD-level reasoning. It also says right there that it's like way smarter than ChatGPT 5.1. They're also promising that it doesn't do flattery. It's more of a brainstorming partner and it won't just tell you you're right if you're not right. It'll tell you your idea is bad if it's a bad idea. Finally, it comes with this new vibe coding platform called Google Anti-Gravity. So of course I grabbed it already this morning and I've been playing with it. And let me tell you, it is awesome. I built and launched an app over the last couple days, but it wasn't working exactly how I wanted it to. So I tried it out in Anti-Gravity and it was awesome because it didn't just fix the problem, but it talked to me about my process and gave me ideas. So I'm a huge fan of Antigravity because it's this mix of vibe coding with an agent telling you what to do from a business perspective. Google's done an incredible job of creating an ecosystem of AI tools that we can all use, and Gemini 3 is a great representation of that. Go and give it a try. Let me know what you think.
