AI News - The AI Hype Is BACK!

The battle for AI dominance between Google and Microsoft really, really heated up this week. Now, I just flew back in from New York where I was attending a Microsoft event. So I'm a day late on making this news video. But man, it has been a week with announcements from Google, YouTube, Microsoft, OpenAI, Amazon, Intel. I mean, this week was a week where everybody was talking about AI. So let's just jump straight in with probably the biggest news that came out of this week was that over OpenAI unveiled Dall E3. And as you can see from their example images, Dall E3 has gotten much, much better than Dall E2 with things like actual text. I just feel so empty inside. And much more contrasty images than what we used to get out of Dall? E. Here's a side by side comparison that they gave of Dall E2 versus Dall? E3 using the same prompt. An expressive oil painting of a basketball player dunking depicted as an explosion of a nebula. This was Dall E2 on the left here. This was Dall E3 3 on the right. Quite an improvement in their demo video here. They actually show somebody using it with more of a conversational language. My five year old keeps talking about super duper sunflower hedgehog. What does it look like? And then it generates some pictures of a hedgehog. My daughter says its name's Larry. Can I see more? Generates more of the same hedgehog. Can I see Larry's house? It generates a house with Larry on it. Larry is so cute. What makes him so special? Here's the reasons. Then it actually sort of tells a story about Larry. So not only is it creating images, it's creating stories and creating environments and creating sticker designs. And so it looks like with Dall E3, the goal is no longer to just type a prompt of an image you want to see, but actually have a conversation with ChatGPT to get to the image you're looking for. And speaking of ChatGPT, they're actually building it straight into ChatGPT. So you're not going to use a separate Dall E3 site. You're just going to go to ChatGPT and have it generate images for you. They gave a whole bunch of other examples of some images. If you click on any one of these images here, you can see the prompt that was used in Dall E3 to get that image. You can also see some of the text, find the universe within you. So it seems to be doing pretty good with text. And then I came across this Tweet from Nick St. Pierre here on Twitter, where he took a bunch of these images that they showed as examples here and then ran the exact same prompt into Midjourney to see how they compared. And here's what he came up with. Here's that same heart image. On the top is the Dall E3 version, on the bottom is the Midjourney version. Obviously, Midjourney isn't going to generate text like Dall E3 is. Here's an image, a house on a cliffside that Dall E3 generated. Here's the Mid Journey version. The top on this one is Dall E. The bottom is Midjourney. And looking through a lot of these, I would say that in some instances Dall? E seems to make a better image, and some of these other instances, I think Midjourney seems to make a better image, but they're definitely closing the gap between them and the ability to just do it straight inside of ChatGPT, as opposed to discord, like we're used to with Mid Journey now, I think is going to be a much needed user experience upgrade. So here's another one where the top is Dall E, the bottom is Midjourney. What do you think? Do you think Dall E is going to give Midjourney a run for its money, Especially since it's going to be just in ChatGPT. For people who already use ChatGPT at no additional cost, it's also going to be inside of Bing, but we're going to talk about that in a minute. According to OpenAI's press release, Dall E3 will be available to ChatGPT plus and enterprise customers in early October. They'll also be available for commercial use. It says, as with Dall E2, the images you create with Dall E3 are yours to use, and you don't need permission to reprint, sell or merchandise them. Next up, I want to talk to you about today's sponsor, BetterHelp. Now, what does BetterHelp have to do with AI? Honestly, not a whole lot. But at the pace that AI is moving, it can often feel overwhelming and hectic. And a lot of people have real fears over where AI is going and how it's going to affect their lives and their jobs. And one way to really help deal with a lot of this craziness that's going on is to actually talk to somebody, which is why I agreed to partner with Better Help on this episode. BetterHelp is focused on making therapy more affordable and accessible to everyone and knowing where to start to find a therapist. Can be difficult on its own. BetterHelp offers an online platform connecting you with professional therapists. By answering a few questions, you can be matched with a suitable therapist within just a few days. As an entrepreneur who's gone through ups and downs and a lot of struggles over my career, I've often turned to therapy to have somebody to talk to and to help calm my mind over some of the fears and stresses that come along with running a business. Signing up is super easy, and if you use my link that's in the description betterhelp.com Matt Wolf you'll get 10% off your first month. Now I know firsthand that therapist compatibility is a crucial element, and if who you're matched with doesn't align, they make it super easy to switch to another one. And it's completely free to do it. So if you're looking for guidance, consider online therapy with betterhelp.com, click the link in the description or visit betterhelp.com Matt Wolf so so thanks so much again to BetterHelp for sponsoring this video. In other OpenAI news this week the information.com dropped an article about how OpenAI hustles to beat Google to launch a multimodal LLM. Now, I'm not personally subscribed to the information because, well, yeah, so I get my news from websites that report on what other websites are reporting on, like this AI Beat website. And essentially what it says is that OpenAI isn't necessarily working on GPT5 right now. They're working on making more and more improvements to GPT4, including what they're currently calling GPT Vision, which is the additional multimodal capabilities for ChatGPT. Essentially, they want ChatGPT to work a lot more like what they showed off in the GPT4 demo when GPT4 came out. Like how it can interpret images, understand context, and produce text or responses that incorporate both text and visual input. And this of course comes a week after Google was talking about how Gemini is coming out soon and it's going to blow GPT4 out out of the water. So the battle between OpenAI, Microsoft and Google is really on right now. And from an AI nerd perspective, it's been really, really fun and exciting to watch because at the end of the day it just means cooler and more improved tools for us to use. This week OpenAI also announced that they are starting an OpenAI red teaming network. They're basically trying to find a whole bunch of outside experts in various fields. Things like cognitive science, chemistry, biology, physics, alignment, education, healthcare Finance, biometrics, privacy, cybersecurity, misinformation, things like that. They want experts in all of these different areas to sort of form a team together, to all work together to figure out how to solve a lot of these problems. And they're going to be outside sources. These aren't going to be, you know, team members working inside of OpenAI. This is a third party that they're trying to create here. If you're an expert and you want to participate in this red team, you could reach out@redteampenai.com and it looks like you can apply up through December 1, 2023. And since we're on the subject of OpenAI, more authors are suing OpenAI. I feel like a broken record because I talk about this in almost every single end of the Week AI news video about more people, more companies suing OpenAI. This week it is George R.R. martin, the creator of Game of Thrones, along with John Grisham, Jonathan Franzen, Jody Pickquote Once again, it's another lawsuit saying that OpenAI scraped the entire contents of their work and people can go and use OpenAI as an alternative to purchasing and reading their books. I am no expert whatsoever, so don't take what I say very seriously. But I actually think that the authors probably aren't going to win these because at the end of the day, the way AI works and the way these AIs are trained is essentially the same way that a human is trained now. If somebody grows up to become an author and throughout their life, they read books by George R.R. martin, they read books by John Grisham, they read books by Michael Crichton and Stephen King and all of these authors, and then they grow up to write their own book. And their books are heavily influenced by all of the works that they've read up until then. That's essentially how the AI is working. Now. These complaints about how if AI consumes all of these books and then we ask it questions about those books and then it could give us responses so then we don't need to go read those books. Well, isn't that already happening on YouTube? Aren't there already, like, book reviews out there where somebody read the book and then gave you a general quick synopsis so you don't have to read it? Well, that person who made that synopsis, synopsis for you did the same thing the AI did. They ingested the book and then figured out how to summarize it in an easier way for you to understand. Same thing goes if. If me and you are just having a conversation. If I read an entire book and you said, oh, I'll probably never read that, can you tell me what it's about? I can regurgitate it based on me consuming it and what I understood. The AI is just doing that same thing. If you kind of look at it from that perspective of, well, if an AI can consume a book and then sort of tell you what it understands about the book based on questions you ask of it, well, then the a. AI would have to then somehow pay the author royalties because they were able to give you some information about the book. Does that mean me reading the book? I would have to now pay the author some sort of royalty because I consume the book? And then it's a very gray area subject. But based on the sort of logic that plays out in my head around it, I have a hard time seeing a lot of these authors win these lawsuits. But again, not a lawyer. Take what I say with a grain of salt. I'm just sort of randomly ranting here. All right, let's talk about Microsoft. And like I mentioned at the beginning of this video, I just flew in from New York because I was actually at this event where they were making all of these announcements. I was actually sitting in the very front row. And this head right here is me. I know it's me because this is the microphone that I had attached to my camera. And I was sitting in the front row. This right here is Pete Huang from the Neuron newsletter. And this right here is Rowan Chung from the Rundown newsletter. And down here was Ijustine, also another tech youtuber. So the whole front row was pretty much various creators in the AI space. So it was really cool that Microsoft was flying people like us out, the creators that are making things like YouTube videos and AI newsletters to sit front row for some of these events. Now, because I flew out to this event and shot a whole bunch of footage, I'm actually going to make a video dedicated to this event. It's going to be a fun sort of vlog style where I sort of walk you through my experience of going to New York and meeting with all these other creators and getting to sit in the front row of this event and then actually getting to speak with a lot of these people that were on stage in one on one sessions to ask them questions and learn more about what they're doing. So that is way too much for this specific video. But I will give you the breakdown quickly of the news that was announced at this event. They really kicked it off by announcing that Windows 11 is going to have AI baked into like everything. You're going to be able to just use your Photos app to remove backgrounds. You're going to be able to have conversations with a little chat window that's going to query all of your documents in Microsoft Word and Microsoft Excel and PowerPoint and all of the Microsoft Suite of tools. You're going to be able to just query straight to them using AI. They're also about to make Outlook free for everybody. So if you have a Windows account, you're going to get a free account to Outlook. And Outlook is going to have a whole bunch of AI features built into it, including the ability to write emails that sound like your past emails. So it actually learns on your emails, your tone of voice, how you typically write and will write emails in your style. So here's the quick breakdown. We're going to get Copilot in Windows, which is sort of that sidebar chat that you see in Edge right now. We're going to get new AI features in Paint, we're going to get new AI features in Photos. You're going to be able to extract text from images and copy and paste text directly from images into other applications. Clipchamp is getting an update which will help you with scene suggestions, edits and narratives, and a bunch of other AI features to help you edit videos. There's some updates to Notepad, Outlook, what we just talked about, a modernized File Explorer and a whole bunch more features coming to Windows 11. Starting on September 26th. Bing and Edge are getting a whole bunch of new features as well. Divya here, who made most of the announcements about Bing and Edge updates. I actually did get a chance to have a one on one sit down with her to ask her more questions about what's coming, what's going to become of SEO, things like that. I had a really fascinating discussion. Again, I'll share a lot of that in an upcoming video, but inside of Bing and your Edge chat you're going to get more personalized answers where it actually remembers a lot of the history of previous conversations and previous things that you're doing on your computer to use as additional context in your conversations. You're going to get a copilot in Microsoft Shopping. They announced that Dall E3, which I already mentioned earlier in this video, is going to be accessible directly within Bing Chat and it's going to be usable for for free inside of Bing chat too. So no additional cost. You can just have it create images for you using Dall E3 instead of Bing. And they're putting their 365 copilot into pretty much everything. And this is sort of similar to what Google is doing with their Duet, where you get this sort of large language model chatbot directly inside of tools like Microsoft Word, Microsoft Excel, Microsoft PowerPoint. Tools like that can now be queried through large language models and all sort of interact with each other. They also announced their new Surface devices at this event, which are much more powerful than previous previous models using Nvidia 4050 and Nvidia 4060 graphics cards. @ one point at the event, they had it rendering a 3D file from Blender on both M2Mac and also one of these new Surface Pros. And the Surface actually finished the render in half the time that the M2 Mac did. They also put Llama 2 onto one of these surfaces and was able to run the large language model locally using these new Nvidia graphics cards. They show off some cool generative fill technology using Adobe suite of tools, where they were able to use their style stylus pen on the screen, sort of circle areas that they wanted to edit, and then use generative fill to change those areas. So that ability to use like a pen tool to get generative fill edits on a touchscreen tablet like this looks like it'll be something really, really valuable. But again, look out for a longer, deeper breakdown of a lot of this Microsoft event with some fun video vlogging elements around it that I shot while at this event that should be coming out pretty soon. And finally, to wrap up the Microsoft related news, GitHub, which is a company owned by Microsoft, announced that their Copilot Chat beta is now available for all individuals. So the GitHub copilot is essentially a large language model chatbot that will help you write and debug and fix coding problems. So you can be chatting with your code and say, why is this code not working? It can help you analyze the code. You can tell it to write you a script about something specific and it can try to write that script. And you can basically have conversations back and forth with the GitHub copilot to get your code dialed in. We can see in this animation here they ask, how do I create a Chrome extension? What should the file structure look like? And then it goes and creates a file structure for a Chrome extension forum. All they had to do was ask with natural language. Now, if you do want to play around with GitHub Copilot, it is 10 bucks a month or a Hundred bucks a year, but they do have a free trial that looks like it gets you about 30 days of use before you actually got to start paying that 10 bucks a month. Now, of course, we can't have a week of Microsoft and OpenAI being in the spotlight without Google jumping in and showing off some really cool stuff as well. So this week Google actually added a bunch of new features to Bard, like the extensions feature, which basically allows you to connect your Bard and have chat conversations with things like YouTube, Google Maps, Google Hotels, Google Flights, Google Drive, Gmail, and Google Docs. They also added the double check feature, where after Bard gives you a response, you can have it double check Google to find sources for the information that Bard used. Now this is one that I already made a video about, so I kind of did a little bit deeper of a dive on Google Bard. The video looks like this, this, but you can see real quickly there's a little extensions button up here in the top right. If I click on this, I am connected to Google Flights, Hotels, maps, workspace and YouTube. Because it's connected to my Google Workspace, it has access to Gmail, Drive Docs, et cetera. So I can enter a prompt like what are the recent updates from my kids school? In my email. It will actually search my email for any announcements related to what's going on at my kids school and then round it all up for me. And then of course I can click into any one of these and it will go directly to that email. The multimodal capabilities of Bard have gotten a lot better over the last several weeks as well. I was able to upload this picture of a cheeseburger and actually get ingredients and a recipe to make a similar looking cheeseburger. Again, I have a whole breakdown of some of Bard's recent updates which I highly recommend checking out. But that's not all we got from Google. This week, on the same day that I was at the Microsoft event, YouTube put on an event to announce a whole bunch of new AI features that they're rolling out. They announced things like Dreamscreen, which are AI generated images or video backgrounds for shorts. We can actually see an example here. Panda drinking coffee. And then it gives us kind of a Gen 2 Pika Labs style video of a Panda drinking coffee. They announced YouTube Create, which is a new editing and production app for mobile creators. This is kind of like YouTube's answer to Cap Cut. They also announced AI insights tools to spark inspiration and help creators decide what to make their next video about, which to me seems like they're kind of going head to head with tools like Vidiq. You can see here from this screenshot, they've got a button that says generate outline suggestions and it actually helps outline a video. So it says says. Next year, YouTube Studio will tap generative AI to spark video ideas and draft outlines to help creators brainstorm. Insights will be personalized for each channel and based on what audiences are already watching on YouTube. We've been testing a version of AI powered tools in YouTube Studio with creators and more than 70% of those surveyed said that it's helped them develop and test ideas for videos. There's going to be an assistive search for finding music for your videos and they're also adding automatic dubbing with a company called Allow, which is going to be a feature where I can speak out an entire YouTube video in English. It will go and recreate and dub over that entire video in whatever language I want it to dub it over in without having to hire like a voiceover expert or something like that. YouTube's just gonna have that dubbing built in, opening up US YouTube creators to be featured in front of potentially so much wider audiences. So something that I'm super excited about. As a YouTube creator who's also in the AI space. YouTube's announcements this week were huge for me. And of course Amazon, one of the companies that originally introduced a lot of us to AI with their Alexa products, well, they're improving Alexa with better AI. So Microsoft had a press event this week, YouTube had a press event this week, Amazon had a press event this week, and intel also had a press event this week, which we'll talk about next. But at Amazon's press event, Amazon announced that their new Alexa experience will be a more conversational experience. Somehow experiences that take into account body language as well as a person's eye contact and gestures. It's also going to give Alexa a bigger and more opinionated personality. So yeah, I'm sure there's going to be a lot of people that maybe don't want Alexa to have a personality, but we'll see. They actually demoed the Alexa device and asked it, what's your favorite sports team? And Alexa responded with follow up questions about the Seahawks stats and game times. The new Generative AI model will also be able to adjust its tone and response to express things like affirmation, excitement, laughter and surprise. Amazon also announced that they're bringing Generative AI to their Fire tv. They said that Fire TV users will now be able to ask Alexa more nuanced and open ended questions about TV shows and movies thanks to the new Generative AI experience, which allows them to speak in a more natural and conversational way. Users will also be able to ask Alexa to find content based on specific actors, directors, scenes, genres or entitlements, such as action movies with car chasing is comedies for a Rainy Day Inside, or animated movies that are free to me. And like I mentioned, intel also had an event this week. At this conference, intel demonstrated laptops that could generate a song in the style of Taylor Swift and answer questions in a conversational style, all while being disconnected from the Internet. They're also working on a new chip called the Arrow Lake, which will come out next year, and although it's a CPU chip and not a GPU chip like what Nvidia makes, they're basically building it to gain traction against Nvidia. Intel also announced that it's building a new supercomputer that would be used by stability AI. They're also working on a software called OpenVino, which will run its own large language model similar to ChatGPT, and essentially talked about a whole bunch of different chips that they're working on which are going to be getting better and better at running various AI tasks. This week TikTok announced that creators can now label their AI generated content, and they say it's not exactly optional. Any AI generated content that contains realistic images, audio or video must be labeled as such to help viewers contextualize the situation and prevent the potential spread of misleading content. So it sounds like you basically tell it whether it's AI or not, but it's not optional. You really need to do that. But TikTok's also working on their own software that automatically detects artificial intelligence. It was also announced this week that Speechly is joining roblox. Last week we talked about Roblox, Generative of game development tools that they're working on, while Speechly is a voice generation tool which will help creators add realistic human sounding or robotic or alien sounding voices directly into their game using Generative AI. Also this week a rather interesting announcement after relaunching as a Studio for creators LimeWire. Yes, that tool that everybody used to use to steal music off the Internet. LimeWire. They're now an AI company and they just acquired Blue Willow, which works very similar to Midjourney but is using stable diffusion under the hood. I've done a couple videos about Blue Willow in the past, so check those out. If you're not familiar with Blue Willow. I just found it fascinating that limewire is even still a company that now they're in AI and they're actually a company that has the means to acquire another company like Blue Willow. That, to me was fascinating. Also this week, Leia Inc. Announced a new update to their Leia Converter. If you've watched my past videos about how to generate movies and 3D type stuff, stuff with AI, you've definitely seen me talk about Leia Converter in the past. The new Leiapix converter has an updated user interface and it's a tool where you can upload a still image and add depth and make it look more 3D and animated. So for example, if I take this image that's supposed to depict Geralt from the Witcher and toss it right into here, you'll see that it'll turn it into like a 3D animation here where you could see some like dark depth behind it. But we have a bunch of different styles here. We could do horizontal, where it sort of rocks back and forth, wide circles, regular circles, tall circles, vertical, and a perspective shift. You could change the length of the animation to real short, which makes it look really fast, or real long, which makes it look much more subtle and slow. We can change the amount of animation less, more. You do get a little bit of distortion around some of these edges, but still a pretty cool effect. We could change the focus point to close or far, and it sort of changes where the pivot happens. We can add more or less edge distortion, which has a pretty big impact on how the effect looks. But depending on the style you're going for, something like that might look very cool for what you're shooting for. And then you have an advanced editor down here where you can make all sorts of tweaks on various axises and amplitudes and phases and all sorts of cool stuff that you could play with to get your 3D animated image to look, look how you want it. Some cool new feature updates from Leiapix. They also did start to add a more premium plan, but it looks like they're giving people credits to start with to play around with it before actually forcing people to pay and use it. And I like to wrap up lately with what's going on in the world of AI and healthcare, because like I've mentioned in the past, this is where I think we're going to see some of the most world changing advancements come out of AI. This week it was announced that Mark Zuckerberg and his wife Priscilla Chan are building an expensive new AI GPU cluster just for medical research. So the new computing system is supposed to give researchers access to generative AI so they can model healthy and diseased cells over time that can help them develop new ways to treat the disease. It's going to be made up of more than 1,000 GPUs, and they're going to be the top of the line Nvidia H100 GPUs. Once this is completed, the new computing system is expected to be one of the biggest AI clusters used for nonprofit research. And jumping over to Google's DeepMind, DeepMind is using AI to pinpoint the causes of genetic diseases. Now, the new AI system that DeepMind is working on is called Alpha Missense, and it looks at a whole bunch of DNA variations and then tries to figure out which variations of DNA are most likely to cause diseases in humans. Now, it's not foolproof yet, but it's going to help speed up diagnosis on rare genetic diseases. Now, this was adapted from DeepMind's previous AI model called AlphaFold, which was a a protein folding AI. But the tool is capable of making predictions based on 71 million DNA variations, which makes it easier and faster to get proper diagnoses. And again, this is still the very, very early state of this kind of technology. It's only going to get better from here. So pretty exciting, pretty huge week in the world of AI. I was so excited to be in the thick of it this week out in New York at the Microsoft event event, talking with other AI creators that are just as excited and just as passionate about all this AI stuff as I am. I'm excited to make some new videos for you around some of the content I created from some of these trips that I've been going on. Keep an eye out for that stuff. The pace of AI really feels like it's getting exciting again. The hype seems like it's starting to come back around again. People are getting more and more into what these tools are going to be able to do, how they're going to make their lives easier, how they're going to solve health problems, how they're going to solve education problems, how they're going to take take steps out of your daily workflow. There's just so much happening right now that I'm super excited about it. And I'm hoping the hype among the general public starts to ramp up again because some of the stuff we're seeing is really going to improve a lot of people's lives. So I think more people should be paying attention. I'm here for it. I'm going to keep on breaking down the news, making tutorials, making fun videos, doing challenges, and just having fun with AI and tech and sharing it with you. If you want to see more videos like this and stay in the loop with all the latest news, make sure you like this video. Subscribe to this channel that'll make sure more of these types of videos show up in your feed. And if you haven't Already, check out FutureTools IE where I curate all of the latest AI news on a daily basis, curate all of the coolest AI tools that I come across on a daily basis, and give you a single location where you can hunt down all of these various videos that I'm making. You can find it all over at FutureTools IO. Don't forget to join the free newsletter as well. And thanks again for watching. And thanks Once again to BetterHelp for sponsoring this video. I really really appreciate the support because you're allowing me to make more videos like this and help raise awareness this and teach people about all the cool stuff that AI is capable of. So thanks again and thank you for watching these videos. I really, really appreciate you. See you in the next one. Bye Bye.