What this means for cognition and learning can be best understood by looking at machine learning, or artificial intelligence, Gopnik suggests. In teaching computers how to learn and solve problems, AI designers speak in terms of “high temperature” and “low temperature” searches for the answers to questions. A low-temperature search (so-called because it requires less energy) involves reaching for the most probable or nearest-to-hand answer, like the one that worked for a similar problem in the past. Low-temperature searches succeed more often than not. A high-temperature search requires more
...more

