Mountain View, CA – Today Google revealed interesting new details on how they plan to address latency when streaming games on Stadia. Antoine Rowe, a lead engineer on the project, described their negative latency feature that will predict user’s inputs through AI to counteract the delay caused by streaming. While the predictive technology is still in progress, the algorithm is already advanced enough to predict the platform’s own failure.
Many cloud gaming services have come and gone. As a result, game streaming platforms are often written off as a gimmick. Usually these services announce and then fail to deliver on unrealistic benchmarks and features. But now, Google is trying to break the mold by utilizing negative latency to register your inputs before you even press the buttons.
“Google Stadia will feature incredibly low latency thanks to our massive data centers, cloud technology, and advanced machine learning algorithms,” explained Rowe about the incredible new gaming platform that can accurately predict player’s inputs and its own demise.
“Many gamers are worried about the delay caused by streaming but they forget that traditional gaming setups have several frames of inherent lag,” added Rowe. “While it will take some time for the predictive AI to improve, the Stadia will eventually reach latency as low as negative 4 frames. At that point the Stadia will be vastly more responsive than local consoles, until those systems implement predictive AI too.”
But what happens when the negative latency AI predicts incorrectly? Rowe suggests that in those cases it’s really the player that was wrong. After all, are you really smarter than the incredibly sophisticated AI that can not only predict complex gaming decisions in negative milliseconds, but also that the Stadia is highly unlikely to succeed as a platform?
The negative latency feature is projected to reach its target effectiveness in about 2 years, so it won’t be boasting the impressively low delay right out of the gate when Stadia launches next month. This is not only because the machine learning algorithm needs to analyze large amounts of data in order to predict more accurately, but also due to the fact that it’s easier to predict inputs when there’s no one left playing.
