AI hype in material science. Google shows an allergy to being pointed to fundamental issues. Another example of pointing out obvious mistakes or issues is not only not welcomed but actively ignored (vgl examples of AI as search engine, where pointing out the first three of the top ten results were wrong resulted in being shouted at, or the blogpost writing video in which presented 'facts' were 1 wikipedia click away from being shown made-up. False citations etc.) It's not bad per se that AI can be wrong, no tool is infallible, but the problem is n:: the extreme asymmetry between the machine effort needed to make stuff up in heaps, and the human effort needed to wade through all the crap and point that out. Vgl [[Spammy handelings assymmetrie 20201220072726]] [[It is easier to F things up than fix things 20180610073041]] Creating entropy is way easier than reducing it, always. We don't need our tools to create ever more entropy on purpose, if only we can reduce it again. Our tools need to help decrease entropy. Decreasing entropy is the definition of life, increasing it should be anathema. Esp if it is unclear where a tool is increasing entropy.