Nvidia Unveils New Features and Announcements at Computex

Nvidia recently made some notable announcements at Computex, and we were lucky enough to catch up on it. One of the things they showcased was their AI capabilities, specifically with regards to RTX acceleration for two stable diffusion applications. The company showed a benchmark chart claiming a 5-second image generation time with the 490 versus 17 seconds on the 700 XTX. While we're not entirely sure how they tested this or how comparable these are, Nvidia provided some impressive-looking images that demonstrate their capabilities.

One of the most interesting announcements from Nvidia was their introduction of a badge system for small form factor (SFF) compatible GPUs and cases. Currently, all that really means is that one of the components fits within a 50mm thickness, 304mm length, and 151mm height. For the other component, it means it's 50mm x 312mm x 1545mm or smaller. If both the GPU and case share this badge, they will be compatible with each other. Nvidia informed us that there is no licensing or cost associated with this effort, which was aimed at creating a simpler standard for SFF casemakers to make clear for customers what'll fit.

Nvidia also reminded us of the Envy app, which we previously covered in more detail. The update here is that they've merged more of the Nvidia control panel functions into the Envy app. While Nvidia isn't ready to discontinue Control Panel yet, it's slowly moving functionality over to Envy. This change will likely make it easier for users to manage their graphics settings and other preferences.

One thing that was notable about Nvidia's keynote was the lack of new announcements. The entire presentation felt like it was rushed together, with a lot of repetition from previous months' events. The event was also quite long, clocking in at 2 hours. While Nvidia did show off some impressive images generated by their AI capabilities, the discussion around them felt disjointed and unclear.

AMD and Intel were both known to have Keynotes officially at Computex, but Nvidia seemed to surprise everyone by announcing its keynote a day early. The event itself was more of a "we're here" kind of presentation, with not much new information shared beyond what we already know about Nvidia's products. The discussion around the importance of building an install base for their platform was confusing and didn't really provide any clear direction or insight into where Nvidia is headed.

Overall, Nvidia's Computex keynote felt like a missed opportunity to share some exciting new information with the audience. While they did showcase some impressive AI capabilities and introduced a SFF badge system, it all felt like we'd seen this before. We're still not entirely sure what was discussed during the event, but we do know that there's more to Nvidia's plans than just what was revealed on stage.

For those who want to catch up on the AMD and Intel Keynotes, you can find our coverage of those events on the channel. As always, we'll be posting multiple times per day this week, so be sure to check back regularly for updates. And if you're interested in seeing more from Nvidia, we've got you covered with our usual coverage of case and cooler companies that we visit during Computex. Thanks for watching, and we'll see you all next time!

"WEBVTTKind: captionsLanguage: enbut you only increase the power by about a factor of three and you increase the cost by only about 50% the more you buy the more you save the more you buy the more you save the more you buy great deal of captured retained waste that we can now relieve out of the system they to C americ and that will translate into savings Savings in money Savings in energy and that's the reason why you've heard me say the more you buy the more you save the more you buy the more you save and now I've shown you the mathematics it is not accurate but it is correct the more you buy the more you save excellent okay that's call CEO math CEO math is not accurate but it is correct the more you buy the more you save the more you buy the more you save live long and buy more shako when the walls fell Timba his arms wide well that was a long keynote we have some actual news Andia didn't announce anything somehow in 2 hours of presentation there was absolutely nothing worth listening to by 2017 the world discovered Transformers in a taste of our dystopian future just like humans this is in our near future this is the way applications are going to look Nvidia says that it can now power better digital brand ambassador experiences in our briefing with the company I asked the hard questions and the question was was what is that and our presenter didn't know so he deferred to another one who didn't know deferred to another one and eventually we got the answer and they'll even be AI brand ambassadors spoiler alert though I'm skeptical but again the hard questions are the ones like Define the buzzwords so our followup was but why anyway eventually we were told that it's for companies that want to either create celebrities to promote their products or use them as customer service Representatives nurses are customer service agents and in other words if you can imagine those phone support calls where you call in you get to a phone robot system and it tells you to push numbers to try and navigate eventually to a human imagine that except instead of competing against a phone robot where you just have to guess digits between one through zero uh now you get to compete against trained AI that is purpose-built to waste as much of your time as possible so as a customer you just give up and hang up the phone that's my vision of the dystopian future that these digital brand ambassadors and customer service Representatives will eventually enable it'll be great for shareholders that's that's good great hi I'm EMA just imagine our as support problem except if instead of there were humans who have the chance of being empathetic and maybe helping you there's not and they're programmed to just you over before that this video is brought to you by the fractal torrent high air flow case the fractal torrent is one of the best air cooling performers we've tested largely thanks to its included 180 by a 38 mm fan the torrent is spacious and easy to work in with what fractal emphasizes is a heavy focus on function the case uses a unique front panel to add some flare without a Major Impact to Performance you can learn more at the link in the description below anyway if we take the most charitable possible interpretation of digital brand ambassadors and customer service with these Solutions it'd be that they can theoretically be more helpful but again the reality is that money is the root of all evil and will ruin everything that is involved in this AI chain so instead probably the companies of the future will just use it to figure out how to trick customers while being a little more legally informed and also informed of the policies to skirt right between all of the problems there was however some actual gaming news as well so they have a new gassist thing which is also AI integrated because buzzwords uh and they had a few other things like remix getting open sourc so we'll cover all that news but first one more word from our future AI Overlord Jensen Juan but this is a gamer GPU but you and I know that this is what a GPU looks like this is one GPU let's start by looking at G assist which Nvidia did not announce or talk about at all in its 2-hour presentation even though it's kind of cool a little concerning but also kind of cool nvidia's Project G assist is a new software utility that they've made it's to try and help players with game progression so the tool takes Voice or text input from the player take screen input from what's visible currently and then system level information it feeds all that into speech recognition computer vision and apis within the agent framework that it has and uses an llm to respond to the player's question so some of these are simpler and realistically not at all different from just running something like present Monon with an overlay such as asking it to recall the last minute of latency that doesn't need an llm beyond the plain text interaction what leans more on the AI angle though is asking the assistant how to craft things or how to progress when stuck think of what youd type into Google for help or what you would call the Nintendo hotline for help for in the 9s I've been playing games for a very long time and then that's what the assistant will try to answer without ever leaving the game there are some cool things in here so it feeds text and Screen information into the llm it can run locally or remotely of course it's accelerated by RTX so it can sell more cards for them because the more you buy the more you save some examples they showed were things like what's the best early game weapon and where do I find crafting or do you see anything on the screen that you can help me understand what's going on which seems like an interesting path to a pokedex integration in some kind of game in the future this could be useful if you get stuck in a game or something or you just don't understand the menu system uh and then on the maybe more cynical side or at least concerned side is just the fact that it is eliminating the need to go to the internet to get information which is where the information came from and so even ignoring the sort of uh intellectual property debate completely ignoring that let's just look at only the fact that currently these llms are able to use an enormous amount of information to build the answers so as you remove the need to produce more of that by getting involved in discussions on forums and things uh of this nature writing game guides uh eventually it will I would think have less to be trained on or at least different material to be trained on maybe it's trained primarily on interactions instead and you might end up in a feedback loop situ situation or maybe a kind of cycle of content growth and content Decay and of course shrinking the need to go to websites will reduce the amount of them available with game guides reduce the amount of interaction on forums uh and eventually we'll see where that leads another piece of the more maybe positive news is the RTX remix which again was completely left out of the presentation the presentation was mostly focused on things like jenton going to night markets and getting stabbed in the face or something I'm not really sure what that story was about my face was cut because somebody was washing their knife and I was a little kid um but my memories of of the Night Market is uh so deep because of that and I used to love I I just I still love going to the night market it's paraphrased but it's not that paraphrased and then there was also something about a a street vendor and uh I I think they walked to the wrong side of the stage like twice there's some other quotes about AI I showed you Blackwell in a prototype State um the other side and it goes into djx systems she's the fruit lady and it's in the middle of the stre the middle between the two but remix has the good news of going open source so that's actually awesome we've liked RTX remix the toolkit is intended to enable game modders to revitalize games especially older ones I actually got my start on game modding map editing and I even used to host some modders on our old forums so modding to me is a very important part of gaming because it just keeps stuff alive keeps older games alive but remix can replace the renderer with modern solutions so modders don't get stuck with archaic limitations like maximum vram limits of say 2 gigabytes or various other renderer challenges the new information is that Nvidia is open sourcing at this month they're also launching the remix runtime SDK which will allow the remix runtime to integrate with games outside of DX8 and dx9 Nvidia is also introducing its favorite buzzword here but in a maybe more traditional way AI upscaling will be available for Textures in addition to prompts now for restyling materials in the game nvidia's example showed a metal panel and portal getting restyled with moss and stone Nvidia also brought back its RTX video solution announ an MS which locally upscales or improves video playback quality we've always been interested in this one because by locally synthesizing pixels and data for video you could theoretically run video at a lower Source from YouTube and then improve it for local delivery RTX video has been out for a while in some forms but it now works in VC and Vinci resolve finally on the AI side Nvidia also announced RTX acceleration for two stable diffusion applications the company showed a benchmark chart claiming a 5-second image generation time with the 490 versus 17 seconds on the 700 XTX we aren't sure exactly how they tested that or how comparable these are but those were at least the numbers they've provided for now the image it generated appears to be a green suited Nvidia superhero and we know it's Nvidia because zooming in you can actually see in the background that he leaves a trail of money everywhere he goes taking a break from the AI Buzz one of the announcements we genuinely liked from Nvidia was its small form factor effort GPU size compatibility and cases has been a challenge for a couple years now to the extent that we normally have an entire section of ITX reviews dedicated to video card compatibility Nvidia is now introducing a badge system that gets assigned to both the video card and the case for sff compatible they call it currently all that really means is that one of the components fits within a 50 mil thickness 304 mm length and 151 height whereas for the other component of the card it means it's 50 mm X 312 X 1545 or smaller if both a card and a case share this badge they will be compatible with each other absence doesn't mean it's not compatible as it could just mean that the badge hasn't been added but it would mean that you should double check now Nvidia informed us that there's no licensing or cost associated with this the effort was to create a simpler standard for ITX casemakers to make clear for customers what'll fit and the company provided a list of cases and cards that already receive the badge we'll flash through those on the screen and you can pause for more detail finally on the consumer side Nvidia also reminded us of the Envy app that we already previously covered you can watch our news video for that but the update here is that they've merged more of the Nvidia control panel functions into nvf Nvidia isn't ready to discontinue control panel yet but it's slowly moving functionality over to nvf so that's kind of the actual interesting news there was some discussion of more Enterprise data center type AI stuff but even for those of you who might be more of a retail investor someone who doesn't really watch us except for a quick recap there wasn't really much new it was a lot of showing the same stuff they showed a few months ago and in fact the whole thing felt like it was just rushed together where AMD and Intel were both known to have Keynotes officially at computex Nvidia kind of surprise announced one hosted it a day before I think it's amds and it really feels like they just toss something together as fast as they could to take the stage for some kind of discussion and we're just not sure what the discussion was because once again it was 2 hours of Meandering nonsense that at times was completely incoherent if you don't create the install base how could the application come and so if you build it they might not come but if you build it if you don't build it they cannot come if there are no developers that use your platform then of course there will be no users but if there are no users there are no install base if they're no install base developers aren't interested in it developers want to write software for a large install base the market developers developers developers by developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers yes so I'm really not sure what I watched uh but there is interesting stuff they just didn't really talk about it at least not news everything else was stuff that we pretty much already heard about or new in some capacity and whatever was new in there was buried in all of the random so uh huge waste of time watching it we've saved you the two hours if you wanted to catch up on it and now you also know stuff that's not even in that keynote and uh that's going to be it for this one so thanks for watching check out our coverage of the AMD Intel Keynotes and of all the different case and cooler companies that we've been to already you can find on the channel check back regularly cuz we're posting multiple times per day all this week and you might not get all the videos served directly to you from YouTube because of that thanks for watching subscribe for more and we'll see you all next time the more you bye I knew you were going to It Up live long and buy more I didn't I didn't even laugh at that that was you all I do was look at you too much the more you save you're not taking this seriously Steve lookbut you only increase the power by about a factor of three and you increase the cost by only about 50% the more you buy the more you save the more you buy the more you save the more you buy great deal of captured retained waste that we can now relieve out of the system they to C americ and that will translate into savings Savings in money Savings in energy and that's the reason why you've heard me say the more you buy the more you save the more you buy the more you save and now I've shown you the mathematics it is not accurate but it is correct the more you buy the more you save excellent okay that's call CEO math CEO math is not accurate but it is correct the more you buy the more you save the more you buy the more you save live long and buy more shako when the walls fell Timba his arms wide well that was a long keynote we have some actual news Andia didn't announce anything somehow in 2 hours of presentation there was absolutely nothing worth listening to by 2017 the world discovered Transformers in a taste of our dystopian future just like humans this is in our near future this is the way applications are going to look Nvidia says that it can now power better digital brand ambassador experiences in our briefing with the company I asked the hard questions and the question was was what is that and our presenter didn't know so he deferred to another one who didn't know deferred to another one and eventually we got the answer and they'll even be AI brand ambassadors spoiler alert though I'm skeptical but again the hard questions are the ones like Define the buzzwords so our followup was but why anyway eventually we were told that it's for companies that want to either create celebrities to promote their products or use them as customer service Representatives nurses are customer service agents and in other words if you can imagine those phone support calls where you call in you get to a phone robot system and it tells you to push numbers to try and navigate eventually to a human imagine that except instead of competing against a phone robot where you just have to guess digits between one through zero uh now you get to compete against trained AI that is purpose-built to waste as much of your time as possible so as a customer you just give up and hang up the phone that's my vision of the dystopian future that these digital brand ambassadors and customer service Representatives will eventually enable it'll be great for shareholders that's that's good great hi I'm EMA just imagine our as support problem except if instead of there were humans who have the chance of being empathetic and maybe helping you there's not and they're programmed to just you over before that this video is brought to you by the fractal torrent high air flow case the fractal torrent is one of the best air cooling performers we've tested largely thanks to its included 180 by a 38 mm fan the torrent is spacious and easy to work in with what fractal emphasizes is a heavy focus on function the case uses a unique front panel to add some flare without a Major Impact to Performance you can learn more at the link in the description below anyway if we take the most charitable possible interpretation of digital brand ambassadors and customer service with these Solutions it'd be that they can theoretically be more helpful but again the reality is that money is the root of all evil and will ruin everything that is involved in this AI chain so instead probably the companies of the future will just use it to figure out how to trick customers while being a little more legally informed and also informed of the policies to skirt right between all of the problems there was however some actual gaming news as well so they have a new gassist thing which is also AI integrated because buzzwords uh and they had a few other things like remix getting open sourc so we'll cover all that news but first one more word from our future AI Overlord Jensen Juan but this is a gamer GPU but you and I know that this is what a GPU looks like this is one GPU let's start by looking at G assist which Nvidia did not announce or talk about at all in its 2-hour presentation even though it's kind of cool a little concerning but also kind of cool nvidia's Project G assist is a new software utility that they've made it's to try and help players with game progression so the tool takes Voice or text input from the player take screen input from what's visible currently and then system level information it feeds all that into speech recognition computer vision and apis within the agent framework that it has and uses an llm to respond to the player's question so some of these are simpler and realistically not at all different from just running something like present Monon with an overlay such as asking it to recall the last minute of latency that doesn't need an llm beyond the plain text interaction what leans more on the AI angle though is asking the assistant how to craft things or how to progress when stuck think of what youd type into Google for help or what you would call the Nintendo hotline for help for in the 9s I've been playing games for a very long time and then that's what the assistant will try to answer without ever leaving the game there are some cool things in here so it feeds text and Screen information into the llm it can run locally or remotely of course it's accelerated by RTX so it can sell more cards for them because the more you buy the more you save some examples they showed were things like what's the best early game weapon and where do I find crafting or do you see anything on the screen that you can help me understand what's going on which seems like an interesting path to a pokedex integration in some kind of game in the future this could be useful if you get stuck in a game or something or you just don't understand the menu system uh and then on the maybe more cynical side or at least concerned side is just the fact that it is eliminating the need to go to the internet to get information which is where the information came from and so even ignoring the sort of uh intellectual property debate completely ignoring that let's just look at only the fact that currently these llms are able to use an enormous amount of information to build the answers so as you remove the need to produce more of that by getting involved in discussions on forums and things uh of this nature writing game guides uh eventually it will I would think have less to be trained on or at least different material to be trained on maybe it's trained primarily on interactions instead and you might end up in a feedback loop situ situation or maybe a kind of cycle of content growth and content Decay and of course shrinking the need to go to websites will reduce the amount of them available with game guides reduce the amount of interaction on forums uh and eventually we'll see where that leads another piece of the more maybe positive news is the RTX remix which again was completely left out of the presentation the presentation was mostly focused on things like jenton going to night markets and getting stabbed in the face or something I'm not really sure what that story was about my face was cut because somebody was washing their knife and I was a little kid um but my memories of of the Night Market is uh so deep because of that and I used to love I I just I still love going to the night market it's paraphrased but it's not that paraphrased and then there was also something about a a street vendor and uh I I think they walked to the wrong side of the stage like twice there's some other quotes about AI I showed you Blackwell in a prototype State um the other side and it goes into djx systems she's the fruit lady and it's in the middle of the stre the middle between the two but remix has the good news of going open source so that's actually awesome we've liked RTX remix the toolkit is intended to enable game modders to revitalize games especially older ones I actually got my start on game modding map editing and I even used to host some modders on our old forums so modding to me is a very important part of gaming because it just keeps stuff alive keeps older games alive but remix can replace the renderer with modern solutions so modders don't get stuck with archaic limitations like maximum vram limits of say 2 gigabytes or various other renderer challenges the new information is that Nvidia is open sourcing at this month they're also launching the remix runtime SDK which will allow the remix runtime to integrate with games outside of DX8 and dx9 Nvidia is also introducing its favorite buzzword here but in a maybe more traditional way AI upscaling will be available for Textures in addition to prompts now for restyling materials in the game nvidia's example showed a metal panel and portal getting restyled with moss and stone Nvidia also brought back its RTX video solution announ an MS which locally upscales or improves video playback quality we've always been interested in this one because by locally synthesizing pixels and data for video you could theoretically run video at a lower Source from YouTube and then improve it for local delivery RTX video has been out for a while in some forms but it now works in VC and Vinci resolve finally on the AI side Nvidia also announced RTX acceleration for two stable diffusion applications the company showed a benchmark chart claiming a 5-second image generation time with the 490 versus 17 seconds on the 700 XTX we aren't sure exactly how they tested that or how comparable these are but those were at least the numbers they've provided for now the image it generated appears to be a green suited Nvidia superhero and we know it's Nvidia because zooming in you can actually see in the background that he leaves a trail of money everywhere he goes taking a break from the AI Buzz one of the announcements we genuinely liked from Nvidia was its small form factor effort GPU size compatibility and cases has been a challenge for a couple years now to the extent that we normally have an entire section of ITX reviews dedicated to video card compatibility Nvidia is now introducing a badge system that gets assigned to both the video card and the case for sff compatible they call it currently all that really means is that one of the components fits within a 50 mil thickness 304 mm length and 151 height whereas for the other component of the card it means it's 50 mm X 312 X 1545 or smaller if both a card and a case share this badge they will be compatible with each other absence doesn't mean it's not compatible as it could just mean that the badge hasn't been added but it would mean that you should double check now Nvidia informed us that there's no licensing or cost associated with this the effort was to create a simpler standard for ITX casemakers to make clear for customers what'll fit and the company provided a list of cases and cards that already receive the badge we'll flash through those on the screen and you can pause for more detail finally on the consumer side Nvidia also reminded us of the Envy app that we already previously covered you can watch our news video for that but the update here is that they've merged more of the Nvidia control panel functions into nvf Nvidia isn't ready to discontinue control panel yet but it's slowly moving functionality over to nvf so that's kind of the actual interesting news there was some discussion of more Enterprise data center type AI stuff but even for those of you who might be more of a retail investor someone who doesn't really watch us except for a quick recap there wasn't really much new it was a lot of showing the same stuff they showed a few months ago and in fact the whole thing felt like it was just rushed together where AMD and Intel were both known to have Keynotes officially at computex Nvidia kind of surprise announced one hosted it a day before I think it's amds and it really feels like they just toss something together as fast as they could to take the stage for some kind of discussion and we're just not sure what the discussion was because once again it was 2 hours of Meandering nonsense that at times was completely incoherent if you don't create the install base how could the application come and so if you build it they might not come but if you build it if you don't build it they cannot come if there are no developers that use your platform then of course there will be no users but if there are no users there are no install base if they're no install base developers aren't interested in it developers want to write software for a large install base the market developers developers developers by developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers developers yes so I'm really not sure what I watched uh but there is interesting stuff they just didn't really talk about it at least not news everything else was stuff that we pretty much already heard about or new in some capacity and whatever was new in there was buried in all of the random so uh huge waste of time watching it we've saved you the two hours if you wanted to catch up on it and now you also know stuff that's not even in that keynote and uh that's going to be it for this one so thanks for watching check out our coverage of the AMD Intel Keynotes and of all the different case and cooler companies that we've been to already you can find on the channel check back regularly cuz we're posting multiple times per day all this week and you might not get all the videos served directly to you from YouTube because of that thanks for watching subscribe for more and we'll see you all next time the more you bye I knew you were going to It Up live long and buy more I didn't I didn't even laugh at that that was you all I do was look at you too much the more you save you're not taking this seriously Steve look\n"