Love the post! Some comments tangentially related:
If decisions are made by vibes (in the sense of answering Fermi questions), then the difficulty of making decisions actually scales sublinearly in the size of the framework used to make decisions. This has the convenient implication that the framework can be made arbitrarily large, and thus can maintain full context. In such a way, both exploration and exploitation can be maintained.
Post this to the Forum! Many optimizers don’t value exploration / uncertainty 😬
will do! thanks for the encouragement
Love the post! Some comments tangentially related:
If decisions are made by vibes (in the sense of answering Fermi questions), then the difficulty of making decisions actually scales sublinearly in the size of the framework used to make decisions. This has the convenient implication that the framework can be made arbitrarily large, and thus can maintain full context. In such a way, both exploration and exploitation can be maintained.