My talk at the Bristol meetup was a keynote version of the my post about extracting insights from AI, navigating the noise and misinformation about AI capabilities. For instance, Google claims its model can consume the entire series of Harry Potter books, the reality is closer to three pages.
In this talk I unpack what context windows are, what latent matches are, and how to apply structure to unstrctured data to solve this and other problems to do with analysing large volumes of userfeed back.
The talk was based on my following blog post: Context Windows, Structured Data, and the Fundamentals of Text Analysis with AI.