It's a good point, I'd consider it (overfitting) a pitfall or common mistake in ML rather than the only mode. I'd agree that most ML models and almost all state

It's a good point, I'd consider it (overfitting) a pitfall or common mistake in ML rather than the only mode. I'd agree that most ML models and almost all state-of-the-art are over-fit to the point of being useless, but that's not an inevitability.