Google Chrome is notorious for its high RAM usage, and one feature meant to assuage that issue, tab discarding, seems to be getting a boost from machine learning. According to a report from Chrome Story, the Chromium codebase contains a bit of commentary that references the feature and proposes a new sub-feature in the form of an algorithm-based Tab Ranker. As its name implies, the goal is for Chrome to proactively rank your currently open tabs so it can decide which ones to hibernate more easily.
Essentially, the tab discarding feature in its current form is not proactive in determining which tabs to hibernate, and thus only does so when memory is running extremely low. This can lead to high usage and make launching other programs a chore, since Chrome will wait until the new program actually starts demanding more RAM in order to free some up. Once that happens, Chrome may well end up clearing an important tab because current criteria for clearing are rather basic. By taking user habits and other factors into account and putting a growing, learning AI algorithm in charge of the process, Chrome will hopefully be able to make better judgement calls as to what tabs to discard and when. Bringing select tabs back by predicting a user wanting them rather than waiting for the user to try to access them could also be on the menu, but details on how it all will work are scant for the time being.
Google has spent the past couple of years making every effort to become the kingpin of the consumer-facing AI solution space, and this piece of code is just another cog in that particular machine. The company has been working its AI expertise into just about every product it builds, and has been training staff to become “machine learning ninja“, among other obvious signs of a shift in priority. For quite some time, Google has focused on creating innovative services that bring real value to consumers, and now it seems that the time has come for the company to redouble efforts to refine those offerings.