After listening to the podcast interview of Xie Saineng by Xiao Jun, I feel invigorated. One of the insights is a very different understanding of the Bitter Lesson. The original understanding was that using clever rules to do things smartly often doesn't compare to using general algorithms combined with brute force to solve problems. For example, chess lost to Deep Blue, Go lost to AlphaGo, and translation and image recognition lost to LLMs. We often overestimate cleverness and underestimate clumsiness. This is the Bitter Lesson. Saineng expressed a very interesting viewpoint in the podcast: LLMs are also a form of cleverness, a shortcut, using a variety of not objectively accumulated language data from the brief internet history to attempt to achieve AGI. This is perhaps similar to a certain school of thought in Go that seeks to solve problems. It is also a form of cleverness, not clumsiness. The intelligence gained this way will only be a small part and will eventually become obsolete. It reminds me of the saying in The Art of War: great wisdom appears foolish. If one wants to avoid defeat, the most important thing is: either do not fight, or bring ten times the troops to battle, to win with numbers. Expecting to win with fewer is bound to be a Bitter Lesson sooner or later. The same goes for entrepreneurship. The easiest way to gain Bitter Lessons is through cleverness. For instance, assuming that big companies won't do it, or assuming that other companies won't see it. This is self-deception. Not being clever and finding clumsy methods is the key to having a chance of success in entrepreneurship. The Bitter Lesson is a good thing. Cleverness may lead to excellence, but experiencing bitterness and understanding it offers a greater opportunity for excellence.