All The Wild AI News You Missed This Week!

Compared to other weeks, it's been a relatively slow week in the world of AI. However, we got some really cool stuff out of Tesla and their we Robot event that happened. And there was also a ton of advancements in the world of AI video. So let's just dive right in, starting with the Tesla We Robot event where Elon showed off some of the things they've been working on over at Tesla. Now, the event started off with Elon getting into an autonomous vehicle and driving around the lot. Here, there is nobody driving the. That's just Elon riding in the front seat. As it drives him over to, I guess, the stage where he's presenting on. They have this moment where he drives around the corner and as he drives around the corner, some bicycles right out in front of him and the car knows to stop to show that it's actually aware of all the stuff that's happening. He rambled for a little bit and then got into actually showing off the new robo taxi that they're putting out. The interesting thing about this robo taxi, it has no steering wheel, it has no pedals. You're. You basically just get in, it drives you to your destination, and then you get out at the destination. You can see as people use the car, they're watching videos or doing work or having zoom calls or watching sports showing that you're just like in this vehicle going to your destination and all of the time to the destination is now freed up time where you can do whatever you want. They showed off all sorts of potential hazards that the robo taxis are equipped to stop for and recognize as they're driving. So stuff that if a human driver ran into this, well, the autonomous vehicles are probably going to actually handle them better, like this dude going the wrong way in traffic or this guy just walking across a freeway here. Humans aren't going to handle this as well as these robo taxis with cameras looking 360 degrees around them will. He also hammered in the point that so much of the world is covered in parking lots and he can see way less parking lots, like green spaces and more space to be used, used for parks and things like that, because now we don't need to park the cars. The cars can drop you off and then go and take another driver somewhere or drive home and park in your garage after it drops you off, making all of this parking lot space and all of these big cities available for other uses. Now, he did give a little bit more details about the cars. He said he expects these to be on the road by the end of 2026, he said before 2027. He also said he expects these robo taxis to sell $30,000. Anybody who can afford that $30,000 could own one of these robo taxis, and then they would have a fairly interesting business model on your hands. Because you can own one of these taxis for $30,000. It will take you wherever you need to go, drop you off wherever you need to be. And then if you want, you'll be able to put it into a mode where it becomes like an Uber for you. And it will go around and pick people up and drop them off. And because you own the car, you make a large percentage of the money. Like, he didn't actually break down how the Monet, but people will buy these cars for roughly $30,000, use them when they need them, and when they're not actually using them, they'll have the option to put them into, like, a robo taxi mode where they can be driving around and picking up other people and dropping off other people while you earn money, because your vehicle is out there doing that now. He also showed off the fact that there isn't going to be a charging port on these either. It's basically going to be like wireless charging for your phone where you go and park this vehicle, like, over a powered parking spot. It charges up the vehicle, and then you can drive it away. So no more plugging in and unplugging a charger. He also then showed off what he called the Robovin, which is robo van. I don't know why he was saying it that way, but he showed off his Robovan or Robovan, which is an autonomous bus, essentially, that can hold up to 20 people inside and take them around. He was giving the example of using it to bring sports teams around, or, you know, use it as like a party bus to go from, like, bar to bar or something like that. And you can see it holds up to 20 people. Now inside, it looked like it had about 8, 9, 10, 11, 12, 13, like 14 or 15 seats, but it also looked like another four or five people can stand while in the vehicle and it can drive people around. No estimate yet on when this is coming. He didn't talk about the range on any of these vehicles. The only thing he did say was that the sort of robo taxi fleet will be coming, like, next year, but using the Model X's and Model Y's that are already out on the roads and the Model Threes and the Cybertrucks, pretty much the Tesla fleet of vehicles that are already out in the world. They'll be able to be used as these robo taxis, I guess, next year sometime. And then the actual robo taxi models that they show to this event are supposedly going to be ready in 2026. He said before 2027. And when he was done showing off his Raboven, he brought in the Optimus robots to kind of show where they're at right now. Now, he didn't really get into any, like, new details about these. All he did say was that they're going to eventually cost less than a car. All we really got to see them do during this event was a video of them watering plants, playing, like a board game, washing a counter, serving drinks, helping get groceries out of the back of a car. And of course, something else we've seen quite a bit is watching them dance, because, yeah, they can. They can do that as well. They can dance. Overall, it was a pretty cool event. Elon kept calling it a party, and it feels like they put on a giant party to unveil these robo taxis and these robo vans and what the Optimus robots can do. But it was also very short. They didn't really go into too much detail around technical specs or anything like that. We got a little bit of a tease of how much they'll cost. We know the robo taxis will be about 30,000. We know that the Optim robots will be less than a car. And we heard that the fully autonomous cars will be available sometime before 2027. Obviously, Elon's always optimistic on his time frame, so we'll see if that actually happens or not. Again, I just love nerding out about this kind of stuff. Robots and autonomous vehicles, and where all this is headed in the future is just so fun and fascinating to me. This was probably the biggest event of the week, the. The biggest, like thing in the AI news world. Today's video is sponsored by LG and their incredible new LG Gram Pro laptop. Picture this. You're at a coffee shop pulling out your work machine. Everyone else is lugging around these bulky, beefy laptops, but not you, because you got the LG Gram Pro, which is super light and also super slim. Toss this thing in a backpack and I promise you, forget you're even carrying it. But despite the fact that it's so light and thin, it's actually a pretty powerful device. It's got an Intel Core 7 Ultra processor and an Nvidia RTX graphic card in it. But the real showstopper is the OLED display. This is visual perfection. Whether you're editing photos, watching movies or YouTube videos, or just doom scrolling Twitter, everything looks really good on this monitor. But LG probably wouldn't have reached out to me if it didn't have some AI features as well. The LG Gram Link app has an AI powered gallery. It's also got glance detection, so it knows exactly when you're looking at it and can turn itself on when it knows you need to use it. And that LG Gram link app, it's also a game changer for your workflow because it lets you seamlessly connect your smartphone and your laptop together to seamlessly transfer photos between the two. And I'm not just talking about one device. You can link up to 10 devices to this thing. So if you're in the market for a laptop that's super light, super powerful, and has a bunch of cool AI built into it, check out the LG Gram Pro laptop. Your eyes will thank you, your back will thank you, and, well, ideally it should make you a heck of a lot more productive. But that's still kind of on you saying that this laptop's going to do everything it can to help. Check out the link in the Description to see LG's page all about this and I've also added some additional links to where you can find it directly on Amazon. Thank you so much to LG for sponsoring this video. Meta just showed off a new video generator that they've been working on called MetaMovieGen. We can see some likely cherry picked examples of what it's capable of here, but they all look pretty good. I mean, they look on a similar level to the demos we've seen from Sora out there. But this does seem to have a few cool features that we haven't really seen in the other video platforms yet. So taking a look at the official page over on Meta's website here again we can see some other examples. Notice that as this girl is running through the sand, she's actually properly leaving footprints. This woman here looks really realistic. But check this out, there's actually a headshot here on the right. So it was actually able to import somebody's real face and make a video of that person. We can see here it says our latest research breakthroughs demonstrate how you can use simple text inputs to produce custom videos and sounds, edit existing videos, or transform your personal image into a unique video. So this is an actual person being transferred into a video. Here's another example of Thunder cracks loudly with an orchestral music track. That's the other thing. It will actually create audio to go with these videos. Let's go ahead and unmute this here. We can hear some background sound, we can hear the thunder, some background music. All of that was created with this AI video generator. Sora wasn't generating audio to go with the videos. This actually is. Here's an example of it being edited. We can see the text input here is transform the lantern into a bubble that soars into the air. And so the original video here has this lantern floating in the air. The edited video, it changed that lantern into a bubble. So not only can you create videos, you can actually import your own face into them. You can have it figure out audio, sound effects and background music and you can even use it to like edit videos and change what's going on in the video. Similar to Sora, Meta showed this off and said, look at this cool research we've done. Look at what we're able to do now. But then didn't give anybody access, so none of us have the option to go and play around with it yet. But this looks like a pretty dang powerful and exciting new model that generates really, really good videos, but again, also adds music, sound effects, imports your own face, and lets you edit some really, really cool stuff. With this Meta movie gen, I'll make sure it's linked up in the description so you can check it out yourself. But that's not all we got. In the world of AI video, this company, hello, AI. I have no idea how to pronounce that. Just launched an image to video feature. So what distingu this image to video experience? Well, according to this tweet here, Tweet and image joint instruction following Haluo seamlessly integrates both text and image command inputs, enhancing your visuals while precisely adhering to your prompts. Apparently it's much more accurate and can manipulate objects within your images. And it's got a bunch of diverse styles. You can find this one over@hail uoai.video. again, don't know how to pronounce this one. When you log into this site for the first time, it does give you a three day free trial. You can have up to three tasks in queue and you get bonus credits if you log in daily. But it looks like typically it's gonna run you about $10 per month, which gets you 1,000 credits per month. And then, you know, you have that whole confusion of how many credits is something gonna cost. And not my favorite style of payment plan, but you can try it for three days for free. Let's double check and see how long it actually takes to create a video. I'll start with this image here that I use for a YouTube thumbnail. And let's say an interactive map Moves and glow man wearing glasses. And then I can turn this on to refine the prompt. Let's go ahead and make sure the prompt is refined and let's see how long this takes. And before we even get going here, it says there are still 2,676 people ahead. Expect to wait for 15 minutes. And here's the video that it generated. So let's go ahead and click on this, see if we can get a bigger screen of it. This is actually pretty impressive. We can see some animation going on in the background. It looks like he's touching the screen. You see another hand pop up. The hand is just totally jacked up, but it looks like a pretty decent looking little cartoon animation. So it's using this mini Max Haluo AI seems to be pretty good. And again you get three days free to test it out. And you can click around on the Explore tab, see some of the other videos that people have generated. And honestly most of them look pretty dang impressive. I think this one really holds up with a lot of the other ones. And again you get a free three day trial. So go mess around with it and see if it's for you. That wasn't the only AI video news we got this week. In fact, there's a new generator in town called Pyramid Flow and it's open source. This is the first AI video generator that we've seen of this caliber that's actually open source. Meaning that if you've got a strong enough computer, you can download all of the code and run it locally without even being connected to the Internet. Or you could run it on cloud servers. They even have it available over on Hugging Face where you can generate it there. And none of that information about what you're generating gets saved. Some of the people that were involved in making this generator are actually some of the same people that were behind the Cling AI generator. Here's some of the examples they showed of what this is capable of. Here's some people, you know, walking in the snow under like cherry blossom trees. There's a black and white video of a boat in front of sort of weird looking Eiffel Tower, some waves crashing against some cliffsides. That looks pretty solid. And all of these videos are pretty decent looking. They can generate up to 10 seconds. And again, it's Open source, meaning that other developers and people that want to fine tune these models are going to get their hands on it and things are going to start to get really, really wild here as people start to make this open source model more and more capable and have different weights that they train it on so we can start to do different things. The fact that we have a video generator that's source, but also this caliber of output is really, really exciting. Quite honestly, if you're a user of Zoom, you're going to be able to let your AI avatar talk for you pretty soon. This is kind of crazy. You'll soon be able to create a custom AI avatar of yourself that you can use to record and send short messages to people. Zoom is getting one step closer to letting AI avatars attend meetings for you. Zoom announced it will soon let you create an AI avatar of yourself that you can use to send brief messages to your team. You'll need to record an initial video of yourself and then Zoom's AI will use that to make an avatar that looks and sounds like you. So similar to what we've seen from hey Gen and Synthesia and some of these other platforms. And it doesn't look like your AI avatar is going to be able to like sit in on the meeting for you yet. Like it's more designed to just use to send messages to people right now as you. But they're implying here that eventually it's going to be a bunch of AI avatars on the meeting. Like that's going to be interesting. Can I just send my AI avatar to jump on like Zoom Team meetings? The AI will just like fill me in later on what everything was talked about. What happens if everybody's an AI avatar? Does anything I should get done in these meetings? So many questions. Speaking of avatars, hey Jen rolled out a new feature late last week called Avatar Looks. We can see in their demo video you can create multiple avatars and create multiple looks of the same character that you created in hey Gen. So here's another example of somebody here talking on screen. And as I scroll through this video, you can see it's the same person but different angles and different looks of that avatar. And since we're on the topic of hey Jen, might as well mention this. Hey Gen and HubSpot have actually partnered up. If you make a blog Post using your HubSpot account, you can automatically have that blog post turned into an AI generated video of a talking avatar explaining what's going on. In that blog post we can see the Title of this blog post. Hey Jen at HubSpot Inbound, the future of AI video and content generation. And then this is the video that it automatically generated. Video is becoming the dominant medium for digital engagement and storytelling. Well, there's a clip of the video that it created, but you get the idea. You make a blog post, it automatically sends that blog post to hey Gen. You log into hey Gen and you now have a video version of somebody explaining what was going on in the blog post. Just a handy little workflow hack for people who like repurposing. Moving over to the world of AI image generation, Imagen version 3 is now available to all Gemini users. We head on over to gemini over@gemini.google.com here. Let's give it a prompt of generate an image of a snowy mountain close to the ocean. We can see here. It now says creating your image with Imagen 3. And here's my snowy mountain close to the ocean. Pretty good. Now Google had some problems with people's faces. Let's try generate a close up image of a male Viking. Let's see if we can figure that out. And right now it says, I'm still learning how to generate certain kinds of images, so I may not be able to create exactly what you're looking for yet. Also, I can't help with photorealistic images of identifiable people. Apparently generating faces still doesn't work. This is the free version of Gemini. If I switch on over to the account where I actually have a Gemini Advanced account here, let's see what happens. Generate a close up image of a male Viking and you can see inside of Gemini Advanced, the one I'm actually paying for. It will actually generate faces now, but it doesn't appear to want to generate faces if you're using the free version. While we're on the topic of images, I found this pretty cool from evolving AI over on Instagram. Somebody figured out how to make these insanely realistic images, like this girl in a coffee shop here, this stream in front of some rocks, the view out of like a van window or something, a trail. These images are all really realistic and something that you wouldn't immediately see and go, that's obviously AI. They just look like realistic images. Well, the little hack was to use Flux and for the prompt give them IMG_1018CR2. So basically just give it like randomness and see what it generates. So I like to use Flux through the Glyph app because right now you can actually use it inside of this app for free. So I'LL just quickly build a glyph here, a text input, an image generator, and then for the image we'll just have it pull in whatever our text was here I'll just put prompt. And then for our image generation model, we're going to use Flux Pro and I'm going to go 16, nine and we'll go ahead and close that. Now I can come over to my prompt here and just give it this prompt here. IMG 1018.

Run this glyph and theoretically it's going to make something random as hell, but really realistic. And there you go. It's like a photo of a bunch of friends at a party or something like that. But honestly, pretty realistic. I mean, it's like really pixely and grainy when I zoom in on it. But from this shot here, it just looks like a normal photo. And of course, we could change some of this around. Let's do like img2489heic and let's run that glyph. Basically use the file name that would be saved directly out of the camera. And look at that. It's just a realistic macro shot of a flower. So pretty cool little trick there. Something fun to try. You can do it for free on glyph. Right now it's just kind of a crapshoot of what you're going to get, but what you do get looks really realistic. While on the topic of AI image generation, Adobe is rolling out some new features to help prove that your art isn't AI slope. Adobe announced on Tuesday that it will begin rolling out a beta of its Content Authenticity Web app. It'll enable creators to digitally certify their works as human made and is immediately launching a Content Authenticity browser extension for Chrome to help protect content creators until the web app arrives. This uses a combination of digital fingerprinting, watermarking, and cryptographic metadata to certify the authenticity of images, video and audio files. It's not going to use metadata where basically taking a picture of the picture removes all the information. There's some sort of digital watermark that will still help, even if it was like a screenshotted image tell if it was AI generated or not. All right, I'm going to rapid fire a whole bunch of other AI news again. It wasn't the craziest week we've ever seen, but there's a handful of small things that I think are worth mentioning. For example, OpenAI is sort of getting frustrated with Microsoft right now. They're saying that they're not moving fast enough to supply them with more and more servers. OpenAI is trying to scale as fast as possible. They partnered with Microsoft to be able to use their servers. Microsoft isn't making Those servers, those GPUs available at a fast enough pace right now. And I guess there's a little bit of tension between OpenAI and Microsoft, at least according to this article on the information. Since we're talking about OpenAI, OpenAI made another content partnership. They now partnered with Hurst. Hurst includes brands like the Houston Chronicles, San Francisco Chronicle, Esquire, Cosmic, Cosmopolitan, Elle, Runner's World, Women's Health, and others. OpenAI is continuing that strategy of partnering with all the big content creators so that in the future, when that content shows up inside of their responses, they're not getting sued anymore. This is also something I thought was pretty cool. There's this new AI tech inside of the Amazon driver vans that they're releasing that puts this little green dot on the package that the driver's supposed to pull out and deliver. So when the delivery van gets to somebody's house, they jump in the back, they look for the package with the green dot, grab that one and go and deliver it. And it uses AI to read all of the, I guess, barcodes on the boxes and pick the right one, any box that's not the correct package we can see in this little screenshot. They all have red X's on them. And then there's the only this one with the green dot. So it's very clear which one the driver's supposed to grab and take up to the door. If you use the Opera browser, they just added some new features. They've got what they call the Aria Commander bar. So by pressing Control plus or Command plus, you could type things like group all video tabs or bookmark all currently open tabs, and Opera will do the desired command with all of the tabs that you currently have open. I've never played with the Opera browser, but stuff like this is making me want to. I know a lot of people are big fans of the Arc browser I've played with that had a hard time migrating over. Kind of always end up back at Chrome, but maybe I'll try Opera next. In fact, I'm curious in the comments. Let me know what browser do you use if you don't use Chrome? I'm constantly trying to see if there's a browser that I like the workflow better, but I always end up landing back on Chrome. So I'm curious, what do you like if you use gmail on your iOS device. It's got an AI assistant now to help you manage your inbox. You can log into Gmail and you'll get a little Gemini notification. How can I help you today? Show unread emails from today. Show unread emails from this week. Get the status of my recent orders. It will help you do all that stuff. Get some quick info just sort of at a glance. Google's also been rolling out those features to help you write emails and summarize emails and things like that. So over time, email is eventually just going to become my AI talking to your AI and no humans actually interacting. I don't know. That's where it feels like it's going at least. There's a new feature coming to Android phones using Google's AI that will detect if your phone's been stolen and then lock whoever stole it out of your phone, says the Theft Detection Lock, a unique security feature, automatically locks your phone screen using artificial intelligence to prevent theft. Using Google AI, the feature detects a typical motion linked to theft arising out of someone grabbing your phone and trying to flee by bike, car or other means. Your phone screen instantly locks if it detects the motion, preventing thieves from having easy access to your data. If you remember the Rabbit devices, the one that pretty much every tech reviewer absolutely tore apart because it wasn't ready yet, it didn't have that large action model built into it yet, which was like the selling point of the device in the first place. Well, apparently they're starting to roll that out now. I know Jesse Liu and Rabbit have gotten a lot of crap for their device, but they still seem to keep on shipping new features. It feels like that humane pin. They put it out there, got a lot of crap and then just sort of collapsed and they didn't seem to put anything new into it and everybody stopped talking about it. At least Rabbit still seems to be putting out new features pretty rapidly and it looks like we're finally getting this large action model that can sort of learn tasks and learn step by step processes and then follow through on those on your behalf. I have a Rabbit I haven't tested this new lam yet, but I will be testing it and sharing my thoughts after I've had a chance to test it out. If you're in one of the countries that hasn't had access to meta AI, well, you're in luck. 15 more countries now get access in the coming weeks to the new meta AI features, including Brazil, Bolivia, Guatemala, Paraguay, Philippines and the uk. It's also slated to rol in Algeria, Egypt, Indonesia, Iraq, Jordan, Libya, Malaysia, Morocco, Saudi Arabia, Sudan, Thailand, Tunisia, United Arab Emirates, Vietnam and Yemen over the coming weeks. If you want to try meta AI and you're one of those countries, you're in luck. A handful of big names in the AI world recently won Nobel Prizes. Geoffrey Hinton, often referred to as the godfather of AI, along with John Hopfield, both won Nobel Prizes in physics for their work in foundational AI. Hinton, who has a PhD in artificial intelligence, went on to co create the back propagation algorithm, a method that allows neural networks to learn from their mistakes, transforming how AI models are trained. His company was acquired by Google. He worked for Google for several years before quitting because he wanted to spread the word about the dangers of AI. Now John Hopfield is a professor at Princeton. He was also one of the early pioneers in AI developing the Hopfield Network, a type of neural network that transformed AI by demonstrating how neural networks can store and retrieve patterns. But they weren't the only AI people to get Nobel Prizes. In fact, we have Demis Hassabis and John Jumper, who are awarded Nobel Prizes in chemistry. Demis is the CEO of Google DeepMind and John Jumper is the director at DeepMind. And they won this Nobel Prize for their work on AlphaFold, which is a groundbreaking AI system that predicts the 3D structure of proteins from their amino acids, which has just helped lead to a whole bunch of new discoveries. And finally, I found this interesting enough to include it. An electronic tongue reveals AI's inner thoughts. This is out of Penn State. Here they developed a electronic tongue that's capable of identifying differences in similar liquids such as milk with varying water content, diverse products including soda types and coffee blends, signs of spoilage and fruit juices, and instances of food safety concerns. The electronic tongue could be useful for food safety and production, as well as for medical diagnostics. The sensor and its AI can broadly detect and classify various substances while collectively assessing their respective quality, authenticity and freshness. Basically, we're getting a AI tongue that can taste things and let us know if it's spoil or rotten or will make us sick or various things like that. And I just found that really, really fascinating because it could potentially help prevent a lot of issues from food and liquids and things like that. And that's what I got for you today. Again, not the craziest week in AI news, at least not compared to most weeks over the last couple years, but still some really interesting stuff, some really fun stuff that we can do with videos and image generation lately. A lot of fun toys for us AI nerds to go play with. If you like videos like this, you want to see more AI tools, more AI tutorials and stay looped in and keep your finger on the pulse of the latest AI news like this video and subscribe to this channel. I'll make sure more videos like this show up in your YouTube feed. And if you haven't already, check out futuretools IO where I curate all the cool AI tools I come across, Keep the latest AI news up to date here and have a 100 completely free newsletter where I'll send just the coolest tools and most important AI news directly to your inbox. It's all free. You can find it over at FutureTools IO. Thank you so much for tuning in and nerding out with me today. And thank you so much to LG and the new Gram Pro LG Notebook for sponsoring this video. Really appreciate you guys. I'll see you all in the next video. Bye bye.