The mind excels at solving new problems with clear parameters but struggles with making difficult decisions that lack defined edges. When faced with a hard decision, it's beneficial to ruminate on it, then sleep on it for a couple of days until a gut answer appears with conviction.

Structuring the reasoning in prompts helps the model perform better, even if it's not strictly reasoning as we define it.

When I prepare for a lecture, I think about the question I want to investigate, often related to the rules in my book, which leads to deeper exploration.

When faced with a problem that seems insurmountable, I often switch to a different problem or modify the existing one to find a new approach.

People are surprisingly willing to wait for a great answer, even if it takes time for the model to think.

The outcome is better if you structure the reasoning and help iterate with the model on how it should do reasoning.

Chain of thought prompting, where you instruct the model to think step-by-step, is less necessary with modern reasoning models but can still be useful in some cases. If you're running thousands or millions of inputs through your prompt, you'll still need to use classical prompting techniques to make your prompt more robust.

o1 utilizes a chain of thought process to break down complex problems into smaller, manageable steps, mirroring human reasoning.

If I encounter issues with my preparation, I spend a lot of time mentally working through them, which I find frustrating.