Everyday AI Podcast – An AI and ChatGPT Podcast - EP 260: AI News That Matters - April 29th, 2024

Starting point is 00:00:00 This is the Everyday AI Show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business, and everyday life. Meet Firefly AI Assistant, now live in Adobe Firefly, the All In One Creative AI Studio. Just describe what you want to create and the assistant handles the rest, orchestrating multi-step workflows across Photoshop, Premiere Express, and more in one conversational interface. You direct the outcome. The assistant accelerates execution. There's a new AI video competitor for SORA.

Starting point is 00:00:51 Invidia has a huge $700 million acquisition that's going to shake up the space. And now Apple is reportedly back on good terms with OpenAI for future iPhone developments. There's a lot going on in AI news, as there is every single week. So you could spend hours a day trying to keep up. you could spend your Mondays with us going over the AI news that matters. What's going on, y'all? My name's Jordan Wilson. I'm the host of Everyday AI.

Starting point is 00:01:20 We're a daily live stream podcast and free daily newsletter, serving as your guide to generative AI. So how we can all actually learn what's going on in the world, but how we can leverage it to grow our companies and to grow our careers. So if that sounds like something you need to do or something you'd want to do, well, number one, you can join us here live every single Monday or listen on the podcast. So if you're listening on the podcast, thanks for that as well. Make sure to check out your show notes as always.

Starting point is 00:01:45 So you can spend hours a day where you could just hang out with us every single Monday to go over the AI news that matters. So make sure there's going to be a lot more in today's newsletter. So go to your EverydayAI.com and sign up for that free daily newsletter. And hey, as a reminder in that newsletter, you'll probably see today is the last day for our meta, RayBand Smart Glasses giveaway. all you have to do is just check. So if you're a longtime subscriber, check Thursdays, email newsletter, check Fridays, or if you're brand new, if you haven't signed up yet, go to your everyday AI.com,

Starting point is 00:02:21 sign up now in the welcome email. You're going to see there's going to be a link. You're going to get a unique referral link. And make sure to share that with your friends. It's super easy. Copy and paste that. You can paste it on social media, send in an email, text it to someone. All you have to do is refer one friend to enter our giveaway.

Starting point is 00:02:38 Whoever has the most referrals is going to kind of get their top pick for our prize pool. So in addition to those meta raybans, we are going to be giving away multiple 90 minutes generative AI consulting sessions. Yeah, wanted to kind of save that one for today as a little surprise. So multiple prizes to choose from. So if you haven't already, we'd appreciate you. If you find value in the show, please refer some friends and get entered into the giveaway for the meta-AI. Ray Bands. All right. So let's get it going and talk about the AI news that matters. And if you're joining us live, appreciate it. Let us know what are the things that are mattering to you right now?

Starting point is 00:03:20 What do you think about the stories? We'll throw your comments on live as we go along. So thanks everyone for, you know, joining from Juan, joining here in Chicago like myself to we got a Florida crew here with Victor and Rolando joining us, Dr. Harvey Castro, joining us from Texas and in Jay joining us from Michigan. Thank you, as always. As we go along, I'd love to get your questions, comments about the AI news that matters. But let's get right into it. And I think there's a big one here. So do we finally have our first real SORA competitor? Well, Shang Shu, a Chinese company, just introduced Vidu. All right. So Shang Shu AI and Singhao University just introduced Vigou, a text to video AI model. So video can generate 16-second HD clips with one-click, utilizing their technology, universal vision transformer or UVIT architecture. So the UVIT model integrates the diffusion and transformer AI models, showcasing scenes with detailed physical laws like light, shadow effects, and facial expressions.

Starting point is 00:04:24 So the technology behind video was actually proposed by its research team in September 2022, preceding SORA's model architecture. video's capabilities include simulating the real physical world, generating complex dynamic shots, and understanding Chinese cultural factors. All right, so let's actually take a look live. So if you're joining us here on the live stream, we're going to play maybe about 30 to 40 seconds of this new kind of tech here. So let's go ahead, take a watch. You'll probably hear some on the podcast.

Starting point is 00:04:56 You'll probably hear something. I'm not going to go too long here. But let's go ahead and take a look at this new model. Let's go ahead. So yeah, we'll make sure to include more of it in today's newsletter. So yeah, pretty good that video shared about 90 seconds worth of, you know, kind of promotional video. And you'll see this is very similar to SORA.

Starting point is 00:05:37 It is not yet on the same level as SORA. And if you are brand new or maybe you miss the, what I would say is pretty groundbreaking news about in late fall, Open AI announced their SORA model, which is, you know, you could just call it on the surface, a text to video AI generator, but it's actually much more than that. I think what a lot of people are overlooking about these new kind of text to video or AI video platforms is how they can simulate the real world and what that actually means, not just for the future of artificial intelligence, but actually the future of AGI, artificial general intelligence, right? I've talked about this on the show before, was lucky enough to partner

Starting point is 00:06:21 with InVIDIA and go out to their GTC conference last week or sorry, last month. And something that no one was talking about is Nvidia CEO mentioned SORA multiple times, unprompted, but wasn't necessarily talking about the quality of the video. He was talking about kind of its implications or the technology's implications on AGI. So it's something that, you know, not a lot of people are talking about. But, you know, as you can see here from the live stream and make sure, if you are on the podcast, make sure to check out the newsletter, we'll be sharing this, this video. But the results are pretty impressive.

Starting point is 00:06:57 So I'll tell you this right now, neither of these platforms right now, Vidu or SORA, are publicly available. But this might put pressure. I mean, who knows, this might put pressure on Open AI to release SORA. It may not. But we did see if you tune into the show a few weeks back, Adobe and Open AI tease a collaboration, which would bring SORA to Adobe's AI or sorry, to Adobe. Adobe's video editing suite.

Starting point is 00:07:22 So presumably some segment of the public will be receiving SORA access one way or another. But yeah, it'll be interesting to see what this new Chinese company, or sorry, not a new company, but what this Chinese company, Shang Shu, their new AI video platform, Vidu, you know, kind of what's going to become of this? Is it going to become publicly available? You know, if I'm being honest, and we've talked about this on the show before, I think this technology, I don't know if the world's ready for it right now, right? Because at least with images and even kind of audio, you can tell, right? And people can understand that, you know, this is very

Starting point is 00:08:00 easy to be deep faked. But with video, you can see something like this and people don't, you know, I'd say 99% of the general population, you know, doesn't know that this technology exists. Therefore, I think it is very easy to put out misinformation and disinformation, which is, I think, one of the reasons why Open AI has been a little slow in getting out this technology, because I don't know if the general population is ready for it, especially before the election here in the US. So a lot more in the newsletter today on that. Yeah, like what Jay said here from our live stream, he said,

Starting point is 00:08:35 this stuff is getting scary good. Yes, if you haven't seen this video already, yeah, make sure to check it out in our newsletter. All right, so let's keep on going with the AI news that matters for this week. Big one here. Speaking of Nvidia, Ovidia, Nvidia has acquired run AI for $700 million. All right. So, so Nvidia has acquired run AI, an AI infrastructure management company for a price tag of $700 million. So this will enhance Nvidia's DGX cloud AI platform and give customer access to run AI's capabilities. So this is a $700 million acquisition for Nvidia, so not a small one by any means. and essentially run AI does this.

Starting point is 00:09:17 It streamlines AI hardware infrastructure management. Or in other words, like the way I like to talk to it is, you know, so many companies out there, you know, are kind of stuck in this space right now where they're wondering, should they be investing a lot of money, sometimes millions of dollars in hardware, right? In buying, you know, these, it's, you know, Nvidia GPU racks, right? Their new Blackwell chip or other chips, you know, or should they rent it, right? So essentially, not to oversimplify things, but you could kind of think of run AI. And there's a lot of other companies out there that do something similar where they essentially rent compute power. So you might need a lot, you know, this quarter, it might go down a little bit next quarter.

Starting point is 00:09:59 Or maybe you're just testing it to see if this is a long-term play for, you know, your enterprise company. So pretty big acquisition here for Nvidia. Also, run AI already has a large customer. base of Fortune 500 companies and has raised already over $118 million in capital prior to the acquisition. And you know what? This makes a lot of sense now because about three weeks ago, there was reports of this, but before that, I was at the Nvidia conference.

Starting point is 00:10:32 And I was probably about five feet away. I was actually in a line. And I was about five feet away from Nvidia CEO, Jensen Wong, posing for photos with the run AI team. And I was like, huh, this is interesting. Why of all companies is he making a little photo op with the run AI team? And now this all makes sense, right? Probably if I wasn't, you know, doing this everyday AI thing every single day,

Starting point is 00:10:56 maybe I would have done a little digging into that and, you know, try to break a story or two. But it makes sense now why there's that kind of photo op and a lot of attention between Nvidia and run AI at their GTC conference. All right. Speaking of other big names in the AI space. Ace. Yeah, talks. Talks are apparently back on between Apple and OpenAI for integration into Apple's next iPhone and their iOS system. So according to a brand new report from Bloomberg, Apple is considering using OpenAI's technology for new features in the upcoming iOS 18 and iPhone.

Starting point is 00:11:36 So if you're listening and you're like, wait, have I heard this before? Yes. There is more flip-flops in this story than, feet on the beach. So let me tell you just a little bit right now what's been going on. So about six months ago, we heard reports that Apple was developing its own large language model for the future iPhone, reportedly spending more than a million dollars a day in development. Right. And then about a month and a half ago, we saw that there were talks between Google and Open AI. So Apple was kind of considering either one or the other. And then right after that, we heard that Apple, no, Apple's probably going to be working with Google. integrating their Gemini model into future iPhones and their iOS.

Starting point is 00:12:18 And then we actually heard a couple of days after that, oh, no, actually Apple has a brand new large language model that they put out a research paper on. And here we are today with brand new news that, oh, nope, Apple and Open AI are reportedly in talks again. So yeah, a lot of back and forth between what's happening with the next iteration. We'll find out for soon here in about six weeks at Apple's worldwide developer conference, WDC in the second week of June. So it should be interesting. I've said this once.

Starting point is 00:12:49 I said it 100 times. Apple's never first to the party, but usually the coolest kid. So it is going to be interesting to see what their eventual generative AI plans are. Apple obviously one of the leaders in hardware in the entire world with billions of devices already out there in the wild. So pretty big, pretty big move for Apple here to see what they eventually kind of or how they

Starting point is 00:13:12 eventually integrate. large language models and generative AI into their next piece of hardware and their software. Adobe just introduced an entirely new way to create, bringing the power and precision of its creative suite into one conversational experience. Meet Firefly AI Assistant, now live in the Adobe Firefly app, the all-in-one creative AI studio. Powered by Adobe's Creative Agent, Firefly AI Assistant lets you start with your vision, just describe what you want, and shape the outcome as it takes form with the

Starting point is 00:13:48 assistant. The assistant orchestrates multi-step workflows, drawing on 60-plus pro-grade tools across Adobe Creative Cloud apps, including Photoshop, Illustrator, Premiere, Lightroom Express, and more to help bring your ideas to life. You can also get started with creative skills, a growing library of pre-built workflows for common creative tasks, like batch editing photos, creating mood boards, portrait retouching, and creating social variations. Every step the assistant takes is visible, so you can can refine, redirect, or take over at any time. You stay in the driver's seat as the creative director. Adobe Firefly AI assistant now in public beta.

Starting point is 00:14:29 See it today at firefly.adobie.com. To be honest, I think that this is probably where they may end up, if I had to guess. I know that there's been a lot of hype recently, especially after some of Google's recent announcements with their Gemini model. Personally, I haven't been super impressed. with Google's AI offerings. So we'll see, you know, Apple here reportedly just going back and forth. And this could be one of those things.

Starting point is 00:15:00 Hey, as a former journalist, I used to do this for, I don't know, seven years. A lot of times big companies will kind of put stories out there. I'm not saying this is what's happening there, but a lot of times companies will put stories out there just to kind of see what the public sentiment is, see how analysts react to the news, see what happens with their stock price. Could that be a case of this? Possibly, right? we've seen so many kind of leaks and reports over the last three or four weeks,

Starting point is 00:15:25 specifically when it comes to Apple's partnership here. And to tell you the truth, it's a pretty big deal. I think it's a bigger deal than a lot of people think because whether you look at it with Apple, whether Apple is going off on its own, whether Apple is going to be, you know, teaming up with, you know, Alphabet slash Google, or whether Apple is going to be partnering with Microsoft backed Open AI. Regardless, that's going to be three of the largest five companies. here in the U.S., you know, and obviously, Nvidia is, is, you know, going to be involved in that somehow

Starting point is 00:15:58 because they're building the chips that are powering all of these AI models. So June should be a pretty big announcement, not just for the future of technology, not just for the future of Apple, its devices and its software, but also for the U.S. economy. You know, take a look at what's going to be happening in May, June, July, kind of leading up to this announcement and right after because I do think it is going to set the U.S. economy and kind of the magnificent seven stocks going one way or the other. All right. I know that was a lot on that one. There's more. Speaking of AI, you got to always talk safety. So the U.S. government is creating a big name star-studded panel to focus on AI safety. So the Department of Homeland Security

Starting point is 00:16:45 has created an advisory panel to address the potential security threats posed by Art of official intelligence to critical infrastructure. So this new panel that was just announced includes prominent CEOs in technology, transportation, defense, and civil rights, and will focus on developing recommendations to mitigate these security risks. So some of the big names named to this panel include OpenAIs Sam Altman, Google CEO, Sundar Pichai, and Microsoft CEO, Sadia Nadella, and many more. So this new board is called the Artificial Intelligence Safety and Security Board.

Starting point is 00:17:23 And like we said, it was established by the DHS, the Department of Homeland Security here in the U.S. And it aims to protect critical infrastructure from AI-related disruptions, right? We've talked about it many times on the show here before, but the not just how much right now generative AI and all these companies mean to the U.S. economy, but also on the flip side, you know, there are foreign states and foreign bad actors. now using AI to launch attacks, right? To, you know, I guess cyber security has obviously been a huge sector impacting, you know, so many businesses here in the U.S., but as generative AI becomes more and more accessible and the learning curve is lower, we talk about all of these platforms that are now being used for deep fakes, you know, video cloning, voice cloning, et cetera.

Starting point is 00:18:12 You know, security is paramount right now here in the U.S. So the government's defense against these potential threats may also involve using the AI technology itself, which is something we've talked about here on the show. So some may argue, though, that this new kind of board, it's kind of like AI regulating itself. And a lot of people are saying that maybe this could hinder advancements in the technology. Also, there are concerns about the effectiveness and potential biases of using AI to counter AI threat. which is something that a lot of these companies have been talking about. Another thing that you should be keeping an eye on is, you know, kind of just overall here, the U.S. government, I've said this before.

Starting point is 00:18:58 I think lawmakers here on the U.S. are so inadequate in terms of their ability to understand this technology. So there is, whether you say fortunately or unfortunately, there is really no other way for the U.S. government rights. now to start to regulate this technology, to mitigate potential risks, then to bring on some of the biggest names in the space, right? But a lot of people are saying, okay, well, so you're going to kind of police AI with the police of AI, right? Like you're going to have the people overseeing or helping oversee this technology who are creating and building the technology. So a lot of people are saying

Starting point is 00:19:40 this doesn't make sense, which I understand that. But on the flip side, right now, U.S. lawmakers have such a poor understanding. If I'm being honest, a lot of U.S. lawmakers don't even understand the internet, which has been mainstream for 30 years, right? So there is such a lack of understanding amongst lawmakers. So there really is, I'd say, no real hope for legislation with teeth here in the U.S. So the U.S. has, I think the Biden White House has done a fairly good job. They put out an executive order regulating AI.

Starting point is 00:20:15 they've been building these kind of, you know, informal groups. But now we see a formal panel along with the DHS with these new, new, the new panel here, the artificial intelligence, safety, and security board. So we'll be linking to a couple of stories in the newsletter about that. So make sure you go check that out. All right. Our next piece of AI news that matters is Adobe releases Firefly 3 in image model and it integrates it into Photoshop.

Starting point is 00:20:44 All right, so Adobe has unveiled an all new beta version of Photoshop with groundbreaking advancements in generative fill and a new Firefly image, their Firefly 3 image foundation model, allowing for greater control and creativity and image editing and bridging the gap between ideation and creation. So Adobe has unveiled these new features in the beta version of its very popular Photoshop. So we have new features like reference image, generate image, generate image, generate similar, generate background, and enhanced detail in the latest Photoshop beta release. So what is kind of new here is seeing this kind of rolled out across all different platforms.

Starting point is 00:21:26 So if you're joining us here on the live screen, I kind of have a screenshot of Adobe's website here. So not just text to image, which they've had in Firefly, but some updates and some new capabilities in generative fill and generative expand. So being able to very easily, you know, just highlight something on an existing photo or in an AI image or in an AI generated photo into change anything that you highlight with text commands, which I think is pretty cool. You also have generative fill and text in Adobe Express.

Starting point is 00:21:57 And then also bringing these new AI capabilities into Adobe's other pieces of software, such as Adobe Illustrator, Adobe InDesign, and just really all across Adobe's platforms. So pretty big news here. And if you haven't checked this out, we talked about this. on the show last week with the new Firefly 3. So definitely some huge advancements in this new model. And I will say right now, Firefly 3 is probably about on par with OpenAI's Dolly, but it's not quite as good as mid-Journey or the stable diffusion AI text to photo models.

Starting point is 00:22:37 However, for the most part, Adobe has been building their models in a, you know, reportedly much more ethical way, just trained on, for the most part, and they say about 95% has been trained on sources that it has access to. So a lot of these other models, you know, reportedly stable diffusion, mid-jurney, and some others are just training on the open internet. So there's a lot of copyright questions around some of these, you know, kind of text-to-image models, whereas Adobe has taken maybe the slower approach, but probably a little more methodical and a little safer,

Starting point is 00:23:12 and with a little more heavy lean into the ethics side. So even though right now this brand new Firefly 3 is not quite on par, you know, with Mid Journeys version 6 with the new stable diffusion XL models, it has gone about it in a probably a little bit more ethical way. All right. So there's going to be a lot more in the newsletter about that today as well. And hey, like a comment here coming from our. live stream audience, AI safety is a great point. Yes, it has to be because there's going to be, you know,

Starting point is 00:23:50 I've talked about this here on the show before, 2024, you're going to see it. There's going to be a big domino falling. You know, eventually we're going to hear something from this, the big open AI versus New York Times lawsuit, or I should say the New York Times versus Open AI lawsuit. The New York Times has sued Open AI and Microsoft for reportedly using millions of its pieces. of copyrighted material to train its, you know, GPT model. So I do think there will probably be a settlement in there at some point. I don't see this going to court.

Starting point is 00:24:23 But that's the biggest, that's the big domino that's going to fall. And then I think we're going to be talking a lot more on, on IP, right, on intellectual property, on, you know, the conversation at that point, I think is going to explode on, oh, we've been using these models now for a year, you know, whether you're talking, you know, Dali or Mid Journey or Firefurt. fly, right? So many people have been using all of these different generative AI, you know, models out in a while now and publicly and commercially now for, you know, a year plus. And I think once that big domino falls, we're going to have the conversation of, okay, how should we be

Starting point is 00:24:58 ethically using this? Are we safe to use this? And a lot of the big companies have provided some kind of copyright protection saying, hey, if you use this and you do get in trouble for whatever reason, we are going to back you. So, you know, at least had certain. levels, especially if you're using the enterprise version of the product. So Firefly, again, I think has separated itself with a much different approach than some of the other kind of text to photo AI generators. All right. I think we have our last story of the day for AI News That Matters.

Starting point is 00:25:34 Meta has updated its RayBan partnership to include multimodal input. So along with its new Lama 3 model update. dates. So yeah, Metas, Metas had a busy week. So along with the new Lama 3 model that is now open for people to go use, meta has also added some new AI capabilities to its flagship hardware device, which is in partnership with RayBan. So the Rayband meta smart glasses now have multimodal inputs, which essentially, I think, makes it a very useful piece of hardware for identifying anything that you may be looking at. So we'll put the link in today's newsletter.

Starting point is 00:26:14 But essentially, if you're wearing these new meta raybans, you can talk to them and say, hey, meta, what am I looking at? And it will say, hey, you are looking at, you know, maybe you're looking at something in another language. And you can say, hey, meta, what am I looking at? And what does it mean? And it will translate it. You can be looking at a historical site.

Starting point is 00:26:32 Maybe you are traveling and using these. Or maybe, you know, who knows? Maybe you're looking at a train sign. And you're like, hey, I'm not sure. which train this is. So essentially tapping into Lamo 3, but the new multimodal aspect of the meta rayband glasses allows it to see and understand and decipher different images that you are looking at in real time through the metaglasses. So essentially it snaps a picture. It kind of does all of its smart processing on the back end using some kind of vision and computer vision,

Starting point is 00:27:08 you know, AI technology to put it very simply. And it tells you, hey, here's what's going on on the front end. So pretty cool update there from meta. And hey, speaking of that as we wrap up today's show on the AI news that matters, hey, if that sounds cool, those meta rayband glasses, well, we're giving them away, right? So our giveaway was to celebrate our one year anniversary of Everyday AI. I, we had our one year anniversary show a couple of days ago. So the contest ends today.

Starting point is 00:27:42 That is right. It ends today. So if you haven't already, make sure you can, if you're already a newsletter subscriber, you can check out today's newsletter or Thursdays, Fridays. If you haven't signed up yet,

Starting point is 00:27:52 make sure to go do this now. Okay, it takes about 30 seconds to enter the contest. Super simple. All you have to do is go sign up for the newsletter. You'll get a welcome email. You'll have a unique referral link. All you got to do is copy that link.

Starting point is 00:28:05 send it to your friends, send it to them in a text, you know, send it to, you know, someone in an email, uh, posted on social media, whatever you want to do. But today is the last day to enter. Our contest ends at 11.59 p.m. Central Standard time tonight. So if you're listening or watching this and it is still Monday, April 29th, it is not too late. All you have to do is refer one friend, uh, to enter. The person with the most referrals gets their pick of the prize pool. Presumably someone's going to be picking these meta rayband glasses, but maybe not because

Starting point is 00:28:35 Our other kind of prizes include we are doing two 90 minute generative AI consults with myself. And our team's going to be working on that as well. So three total prizes. Whoever has the most referrals gets to pick their prize of choice. And then the other two will be random. So if you get the most referrals, you get your shot at whatever you want. And all you have to do is have one referral to enter the contest. So make sure you go do that.

Starting point is 00:29:00 So go to your everyday AI.com. Sign up for that free daily newsletter. go grab your referral code. And hey, thank you for joining us for going over the AI news that matters. Thanks for tuning in. And we hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all. Meet Firefly AI Assistant.

Starting point is 00:29:26 Now live in Adobe Firefly, the Allman One Creative AI Studio. Just describe what you want to create in your own words and the assistant handles the rest, orchestrating multi-step workflows across Adobe Creative Cloud apps, including Photoshop, Premiere Express, and more in one conversation. interface. You direct the outcome while the assistant accelerates execution. Stand control with the ability to step in and refine at any time. See it today at firefly.adobie.com. And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going. For a little more AI magic, visit Your Everyday

Starting point is 00:30:12 AI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.

Everyday AI Podcast – An AI and ChatGPT Podcast - EP 260: AI News That Matters - April 29th, 2024

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.