Finale: Let’s wrap up the week with some news headlines.
No, not the unity runtime fee – something bigger: A new model from OpenAI.
One of the biggest flaws of AI so far has been its logic and math performance.
The previous model, 4o, scored 13% on the math olympiad problems.
The newest one just released, scored 83%.
(Funny enough, similar to the math olympiad score I had back in high school 🤓 )
This is probably the Q* project they’ve been teasing (or being leaked).
But what’s the significance? Why is this relevant? Don’t we already have calculators?
The thing is, predicting words is one thing, logic is a completely separate pattern that can lead to better chain-of-thought reasoning that can answer tasks more effectively.
The current model is said to perform as well as PhD students in physics, biology, and chemistry.
Let’s tinker around and see what it can do 😮
