School trick unlocks major AI advancement

OpenAI latest model "O1" beats GPT4.
But core to its performance is an old school trick.

Remember when your teachers told you to double-check your work?

To proofread your work.

Turns out, they were onto something big :)

OpenAI's new model, o1 is essentially ... a high-tech proofreader

With a good marketing story.

It "thinks" before it speaks, meaning:

• it outputs a first draft
• it processes it again, checking vs the instructions
• and then make ajustements if needed

To me that's exactly what we've been telling kids to do for centuries.

Think before you speak. Proofread.

Answers are more accurate and thoughtful. But also slower.

So we shouldn't be surprised.

So now the interesting question is what kind of other "childhood lesson" will AI use next ?

---

Text drafted through a chat with an AI model. Visual made with Flux-Dev. Edits by Jean-Paul Paoli. Reach out if you believe part of this content infringes on copyrighted material