Which AI model should you use?

Hi, reporters. As AI tools burst on the scene and fade away just as fast, it is hard to keep up. Recently, Jeremy Caplan mentioned a little tool called Arena.ai in his newsletter.

Arena.ai takes a question or AI prompt and shows responses from different models, side by side. The idea is, you throw two LLMs into the arena and see who comes out on top – or at least, that’s where I think it gets its name.

For instance, look at these Gemini and Claude responses to my question “Who is Elon Musk?”. They touch on some of the same topics, but I would say Gemini presents a positive spin on them, while Claude is more critical.

And importantly: neither one is neutral!

Arena also asks users to rate LLMs to see which is best, but I would just use it for these side-by-side comparisons. Plus, it strikes me as a tool that will be around for a while and then gone like the wind. So appreciate it while you can!

And a warning – Arena.ai is not private. In fact, the site warns you, “Your prompts may be shared publicly to support AI research.” So yeah – do not use it on your stories, interviews, etc. But you may want to use it for innocuous questions, dummy data, etc. to see the differences between models. Check out these AIs, reporters!

One more thing...

Did you miss the last TFR? Use Google Reverse Image Search to return TikTok results