The theory here is that in the real world, problems do not come neatly wrapped in boxes labelled ‘Maths’ or ‘English’. Thus, teaching pupils in these ‘subject silos’ is ineffective. Instead, we should teach pupils using projects or activities which more accurately reflect the problems they will face in the real world. Such projects also have the benefit of being more intrinsically motivating for pupils, and will help promote ‘independent learning’, a popular buzzword in modern education. Again, Ofsted, the RSA Opening Minds curriculum, the ATL and a variety of educationalists all make this practical or theoretical case. I argue that it is a confusion of aims and methods. Our aim should be for pupils to be able to tackle real-world problems by the end of their education; that does not mean that our method should involve endless practice of real-world problems. This is because real-world problems often involve a great deal of distracting information which overwhelms working memory. Likewise, our final aim should be for pupils to work independently; this does not mean that constant independent learning will achieve this aim. ‘Independent learning’ often just means discovery learning or unguided learning, which are highly inefficient and ineffective ways to learn new material. There is also a novice/expert issue here – experts are good at solving real world problems, but we shouldn’t ask novices just to mimic what experts do, otherwise we’re into cargo cult territory. Experts think in a qualitatively different way from novices.