Skip to main content
  1. Home
  2. Computing
  3. News

Nvidia’s G-Assist AI assistant actually sounds incredible

Add as a preferred source on Google
Destiny 2 running on the Asus ROG PG42UQ.
Jacob Roach / Digital Trends
Computex Logo
This story is part of our coverage of Computex, the world's biggest computing conference.

Well, no one saw this coming. Nvidia announced an AI game assistant at Computex 2024 that will help you do everything from allocate stat points in a game to set up your PC for the best gaming experience. It may sound like another AI ploy from the AI giant that is Nvidia, but Project G-Assist actually could be really useful.

The headline feature that Nvidia is promoting is that Project G-Assist can help you when you’re stuck in a game. Developers can train the model on their game, and then that model can access your prompt, what’s visible on screen, and even data from the game. From there, you’d be able to ask questions you’d normally need to look up online. For example, where you should put a stat point, or how to craft a particular item. It’s not dissimilar to what we saw Microsoft show off with its Copilot demo at Build earlier this month, where a voice assistant walked a player through Minecraft in real time.

Project G-Assist | Your AI Assistant For Games & Apps

Nvidia’s version of that seems far more practical, but Project G-Assist has another card up its sleeve. In addition to the game assistant, Nvidia demoed a version of Project G-Assist that could help you tune your system for optimal performance by having full access to your system. That includes settings changes within the game, setting your refresh rate correctly, and more. It seems that the model can have access to Nvidia hardware as well.

Recommended Videos

Blink and you’ll miss it, but in its demo, Nvidia shows off Project G-Assist automatically applying an overclock to your GPU and graphing latency over time while you’re in a game. Even more impressive, Nvidia showed Project G-Assist reducing system power while maintaining a target frame rate by undervolting the GPU to achieve the performance target.

A diagram showing on Nvidia's Project G-Assist works.
Nvidia

Although it’s great to get help in a game when you’re stuck, games are created to entice players to overcome challenges. The flip side of Project G-Assist is very interesting, however. It takes normally complex processes like undervolting and overclocking, and it makes them instantly accessible. It can do so in the context of what you’re doing on your PC as well, potentially saving hours of testing.

There’s a bit of shilling for Nvidia, even in the demo. In one scene, the AI assistant recommends turning on Nvidia Reflex in Cyberpunk 2077, for example. I’m sure there are cases where it would recommend something like Nvidia’s DLSS over other upscaling options as well. However, there’s clearly enough power here to overlook Nvidia plugging its other utilities.

Project G-Assist isn’t an actual product, at least not yet. Nvidia has a demo, and it’s pitching the tool as something that developers can take advantage of in the ways they see fit. I sincerely hope there’s a plan for Nvidia’s take on Project G-Assist, though. There are enough features, settings, and utilities available for RTX graphics cards to justify an AI assistant taking new users through them all, and according to Nvidia’s demo, that assistant could be very useful.

Jacob Roach
Former Lead Reporter, PC Hardware
Jacob Roach is the lead reporter for PC hardware at Digital Trends. In addition to covering the latest PC components, from…
The Mac Pro is dead at Apple, and I’ll miss the cheese-grater powerhouse
RIP Mac Pro. The Mac Studio is taking the throne, and we're okay with that.
Electronics, Computer, Pc

Apple has officially discontinued the Mac Pro. It’s been removed from Apple’s website, and Apple has confirmed to 9to5Mac that there are no plans to release a future version. The buy page now redirects to Apple’s Mac homepage, where the Mac Pro no longer exists.

Why did Apple kill the Mac Pro?

Read more
March Madness, Revisited: The AI Model Did Well. But Mad Things Still Happen
Stills from NCAA games.

(NOTE: This article is part of an ongoing series documenting an experiment with using AI to fill the NCAA brackets and see how it fares against years of human experience. The original article is as follows.)

A week ago, I wrote about entering an NCAA tournament pool with a more disciplined process than I usually use.

Read more
A simple coding mistake is exposing API keys across thousands of websites
Security gaps that are easier to miss than you think
Computer, Electronics, Laptop

After analyzing 10 million webpages, researchers have found thousands of websites accidentally exposing sensitive API credentials, including keys linked to major services like Amazon Web Services, Stripe, and OpenAI.

This is a serious issue because APIs act as the backbone of the apps we use today. They allow websites to connect to services like payments, cloud storage, and AI tools, but they rely on digital keys to stay secure. Once exposed, API keys can allow anyone to interact with those services with malicious intent.

Read more