Google is a discovery engine that is increasingly pushing out answers. It has 20 years of historical query data to use to be able to predict what we will ask for before we ask. Hannah kindly tells me my horizontal vs vertical theory is a really cool idea. Then expands to a dual algorithm theory and backs it up with lots of solid arguments, including PACRR (a rather boring acronym that is actually REALLY interesting). Stop thinking about algorithms since machine learning means the algorithmic path changes at every execution, and there is therefore no specific set of tasks to do since the path is never predictable. She then explains what authority and relevancy are, and how brands can approach a search / discovery world where they are key. Entity as a ranking factor and how to make the most of that insight (clue: much more than SEO). WELL worth listening to. Mind blowing.