Microsoft will train its games in a “catalog” of titles soon, Satya Nadella says, “and then start playing them”

The Chief of Microsoft, Satya Nadella, said the company plans to train her new games in a “catalog of games” soon, so it can “start playing them.”

Speaking in the dwarkesh patel podcast, Nadella described when seeing the Microsoft Ai Muse games generate game videos that reflect a user’s entry as a “massive and massive moment of ‘Wow'”.

Microsoft lifted the cover in Muse last week and described technology as something that would help the ideation and preservation of the game, although not everyone was convinced.

In summary, Muse has been trained so far depending on years of game images of the Ninja Theory Bleeding Edge’s sand. After ingesting this game file, Muse can generate new game images and estimate how you could change based on a specific warning suggested by the user. For example, Muse could generate game images with an additional jump pad at the level, if they are ordered to do so.

So far, this is not an AI capable of starting coding their own games or dreaming of their own concepts. Rather, it is a tool that could eventually become a shortcut to provide game images generated to accelerate the development process.

“Can you generate games that are consistent and then have the ability to generate the diversity of what this game represents, and then they are persistent for user modifications?” Nadella said. “That’s what it is.

“What excites me is to bring: we will soon have a catalog of games that we will start using these models, or we will train these models to generate, and then we will start playing them.”

Nadella did not explain how users could play the generated footage offered by Muse.

“When Phil Spencer showed it to me for the first time, I had an Xbox controller and this model basically took the entrance and generated the exit based on the entrance. And it was consistent with the game,” Nadella continued.

“That for me is a massive and massive moment of ‘Wow’. It’s like the first time we saw Chatgpt complete sentences, or Dall-E draw, or Sora. This is one of those moments.”

When writing in response to Microsoft Muse’s announcement last week, Michael Cook’s game expert described the progress as “impressive” but “is not a practical process.”

“It is impressive that you can do this using visual information because things like lighting, camera angles, user interface, etc. are a lot for an AI model, and so on, it was still only able to generate images that They predict the behavior of the player.

“The research team behind this probably believes that it will become more efficient over time, which could make it more affordable or manageable for small developers,” Cook continued. “However, it still raises the question of how we get the video of the people who play our game first.”

Last month, the interactive chief Take Take Strauss Zelnick intervened with his opinion on AI, saying that “artificial intelligence is an oxymoron, there is no such thing.”