layout: single title: “Relational Database Systems and AI” tags: tech machine-learning tech-history
Last summer, Frank Chen at Andreessen Horowitz delivered an excellent introduction and history of AI.
He has followed up with another excellent AI presentation with a comparison between AI and the history of relational databases. (I prefer the more accurate term RDBMS—relational database management systems.) :
…E.F. Codd invented the term relational database working at IBM Research in 1970 and then Oracle implemented relational database in the late 70’s and early 80’s and basically since that time, we’ve watched the relational database get into every important piece of software that you write…[The RDBMS] was one these fundamental pieces of computer science that had such broad applicability it just got into every important piece of software. I think AI is going to be like that as well, which is, [AI] is going to get into every important piece of software.
Just as I pondered the historical analogy between the smartphone era and the web (or web search), I’ve been thinking about how the history of RDBMS suggests the path of AI going forward.
Codd published in the early 1970’s, Larry Ellison did not found Oracle until 1978, and the first versions of commercial software that were practical in production environments were not running until the 80’s, more than a decade after Codd’s first paper. The database wars heated up in the 80’s and Oracle did not emerge as the dominant player until the early 90’s.
Now, more than 45 years after the fundamental computer science innovations of Codd, RDBMS are a cornerstone of virtually every major piece of software. I’ve become increasingly convinced that Chen is right and AI’s impact within the tech industry will be at least as broad and deep as relational databases.