Kicking off The Journal with Matthew Finch—while eagerly waiting on his upcoming ‘lo-fi ai’ project;
“Our conversation made me consider what expression was. Can human expression be isolated and represented in data? If so, can’t a model retain the principles of that data’s structure? They rebutted with the example of market analysis to suggest that what is retained by a model is not the expression itself, but rather an insight describing the expression. An insight categorizing and understanding the expression. But not an insight making the expression. Let’s say we gather a data set on a marketing campaign. The data may well inform us where to invest our marketing budget, between which platforms to maximize our ROI based on current cultural practice — but can the model truly understand the consumer or cultural context of why the market campaign is working or not? Can it isolate and retain the ethos of the market? Can the model make the decision or anticipate unexpected social trends?”
