You are not a Bayesian homunculus whose reasoning is 'corrupted' by cognitive biases.
You just are cognitive biases.
You just are attribution substitution heuristics, evolved intuitions, and unconscious learning. These make up the 'elephant' of your mind, and atop them rides a tiny 'deliberative thinking' module that only rarely exerts itself, and almost never according to normatively correct reasoning.
You do not have much cognitive access to your motivations. You are not Aristotle's 'rational animal.' You are Gazzaniga's rationalizing animal. Most of the time, your unconscious makes a decision, and then you become consciously aware of an intention to act, and then your brain invents a rationalization for the motivations behind your actions.
If an 'agent' is something that makes choices so as to maximize the fulfillment of explicit desires, given explicit beliefs, then few humans are very 'agenty' at all. You may be agenty when you guide a piece of chocolate into your mouth, but you are not very agenty when you navigate the world on a broader scale. On the scale of days or weeks, your actions result from a kludge of evolved mechanisms that are often function-specific and maladapted to your current environment. You are an adaptation-executor, not a fitness-maximizer.
Agency is rare but powerful. Homo economicus is a myth, but imagine what one of them could do if such a thing existed: a real agent with the power to reliably do things it believed would fulfill its desires. It could change its diet, work out each morning, and maximize its health and physical attractiveness. It could learn and practice body language, fashion, salesmanship, seduction, the laws of money, and domain-specific skills and win in every sphere of life without constant defeat by human hangups. It could learn networking and influence and persuasion and have large-scale effects on societies, cultures, and nations.
Even a little bit of agenty-ness will have some lasting historical impact. Think of Benjamin Franklin, Teddy Roosevelt, Bill Clinton, or Tim Ferris. Imagine what you could do if you were just a bit more agenty. That's what training in instrumental rationality is all about: transcending your kludginess to attain a bit more agenty-ness.
And, imagine what an agent could do without the limits of human hardware or software. Now that would really be something.
(This post was inspired by some conversations with Michael Vassar.)