The learning rate of your internal neural net was still dialed up near max, vs. today, where it's rapidly decreasing and thus things you're exposed to aren't nearly as capable of shaping your internal weights.