From the workbench, part 4: weighting

Continuing our efforts to get a functioning search engine going, we now need to see what the math looks like for weighing search results, and how this would be implemented. Last time, we looked at the x-axis of our probability distribution, which is the direct match coefficient iterated by the contextual match coefficient: for any … Continue reading From the workbench, part 4: weighting


1953’s pre-BIOS

Reading old operating manuals and computer-type books of the more general kind often leads you to discover early precursors of present-day tech. Today, I discovered such a thing in Booth and Booth's Automatic Digital Calculators, the book in which Kathleen Booth describes the first assembler model. After describing basic code models, one-address code and two-address … Continue reading 1953’s pre-BIOS

A.I., the decidability problem, and “tech support”

For a contemporary example of all this, consider the latest kerfuffle with Meltdown and Spectre, and Linus Thorvald's wonderful takedown of the folks responsible. Every now and then in the waves by which "A.I." comes and goes as a newsworthy topic, the question comes up what all remains 'human' once "A.I." takes off, or the … Continue reading A.I., the decidability problem, and “tech support”