Bias reduces noise—if you know <em>roughly</em> what to expect, then being biased by those expectations means you won’t get distracted by less relevant data points.

The ‘naming’ problem, where by naming something we feel we have explained it, the ‘language’ problem, where the words we use stop others from understanding, and the ‘question’ problem, where we fail to find the right questions, are common and funny.

The ‘Great Man’ theory of history has the history of ideas moved forward by individuals. But by thinking of these as ‘Great Ideas’, or better ‘spirits’ of ideas, we’re encouraged to examine their motivations, which is surprisingly effective.

AI alarmism thrives on speculative, worst-case scenarios, but our understanding of AI’s fundamentally alien nature and the complex forms of consciousness make me suspect that less stressful alternatives are equally plausible.

Fringe theories always seem to cluster together. It seems weird, but mainstream theories also do, we just don’t often examine them. Examining why different theory stacks arise reveals much about our biases, ideologies, and the influence of community-based knowledge.