AI News: A TON of News (and DRAMA) This Week!

It's been a super busy week in the world of AI and there's been quite a bit of drama. Now I actually covered a lot of that drama in a video earlier this week called OpenAI is seriously concerning Me. So if you want my full breakdown and my opinions on a lot of the drama, make sure you check out that video. Now, I will cover a little bit of it in this video, but I'm not going to dive too deep into it because I already made a video about it. But let's go ahead and start here. This week, OpenAI formed a brand new board for Safety and Security. Now if you remember, a couple weeks ago, Ilya Suskever left, Jan, the head of Alignment left, and there was a lot of talk about how Sam Altman and OpenAI weren't taking safety seriously. Well, after Jan left and the team that runs alignment over at OpenAI was disbanded, this new Safety and Security committee was created and it consists of four of the current board members from OpenAI, Brett Taylor, Adam D', Angelo, the Nicole Seligman, and Sam Altman, along with a handful of other team members over at OpenAI. So as of right now, Sam Altman, the guy in charge of the direction of OpenAI, as well as the number one salesman for OpenAI, is also one of the guys in charge of Safety and security now, and a lot of people have been questioning that decision. Also this week there was a new episode of the Ted AI podcast with Helen Toner, one of the board members that voted to boot Sam Altman out of OpenAI back in November of 2023. He during this interview, we learned a little bit more behind the scenes of what happened back in November and why the board made the decision to fire Sam.

When ChatGPT came out November 2022, the board was not informed in advance about that. We learned about ChatGPT on Twitter. Sam didn't inform the board that he owned the OpenAI startup fund, even though he constantly was claiming to be an independent board member with no financial interest in the company. On multiple occasions he gave us inaccurate information about a small number of formal safety processes that the company did have in place. And so, you know, there's more individual examples. And for any individual case, Sam could always come up with some kind of like innocuous sounding explanation of why it wasn't a big deal or misinterpreted or whatever. But you know, the end effect was that after years of this kind of thing, all four of us who fired him came to the conclusion that we just couldn't believe things that Sam was telling us.

After this interview came out, other people started speaking up more about Sam, including people claiming that Sam Sam was actually fired from Y Combinator due to similar reasons that Helen was talking about in this interview. However, Paul Graham, one of the co founders of Y Combinator, said this on Twitter. People have been claiming YC fired Sam Altman. That's not true. Here's what actually happened. For several years he was running both YC and OpenAI. But when OpenAI announced that it was going to have a for profit subsidiary and that Sam was going to be the CEO, we told him that if he was going to work full time on OpenAI, we should find someone else to run YC and he agreed. If he had said that he was going to find someone else to be CEO of OpenAI so that he could focus 100% on YC, we would have been fine with that too. We didn't want him to leave just to choose one or the other. So there seems to be a lot of mixed stories going around about, I guess Sam Altman's character. Jan claimed he left because OpenAI was no longer concerned with safety. Helen Toner claimed that they tried to fire him because he constantly lied and people took to Twitter to spread rumors about why he actually, actually left Y Combinator, which the co founder of Y Combinator just set straight. It's definitely been fun and kind of interesting to sort of follow this storyline and put the pieces together and figure out who's telling the truth and who's not. And I do think more details will come out with time, but all I know for sure right now is that OpenAI still seems to be the front runner for developing the best models out there. And as of right now, that's going to be true, despite the drama and the hearsay that's going around right now. And while we're on the topic of OpenAI, well, it turns out that some bad actors have been using ChatGPT for bad actor stuff. OpenAI put out this blog post disrupting deceptive uses of AI by covert influence operations. Apparently operations were shut down in Russia, China, Iran and Israel, where people were using ChatGPT to generate, I guess, misinformation content and share it on social media. It says all these threat actors used our services to generate text and occasionally images in greater volumes and with fewer language errors than would have been possible for the human operators alone. It doesn't get into specifics about what the actual attacks and disruptions were, but regardless, OpenAI shut them down. Also, in OpenAI news, they've been starting to partner with more and more media companies. Recently we got word that they partnered with News Corp. And this week we learned they also partnered with Vox Media and the Atlantic. Now my guess is that a lot of content from these platforms was probably already trained into these models and at as more sort of traditional media companies are suing OpenAI. OpenAI is trying to get ahead of it and partner with some of the companies that have content that's trained into the models. But this also opens up the opportunity for new content from these platforms to be trained into the models as well as to get more real time content served into ChatGPT. It also appears that the media companies will get access to OpenAI's technology to develop products to better serve its audience and advertisers. But those aren't the only partnerships that OpenAI is involved with. In fact, word on the street is that OpenAI just cemented a huge deal with Apple. Now this article from the Information says it isn't clear how Apple plans to use OpenAI's technology, but the iPhone maker hasn't developed its own large language models on par with those of OpenAI and Google. Apple has discussed potentially using OpenAI's tech in Siri Voice Assistant to answer complex questions that Siri couldn't answer by itself. Now, as far as I can tell, this is still rumor because I scroll up here spokespeople for OpenAI and Apple didn't have a comment, but I imagine we'll have a lot more details after the upcoming WWDC event from Apple, which happens on June 10th. And in the final bit of OpenAI news that I'm going to share this week, custom GPTs are now available for free users of chat GPT. But what's even cooler is you pretty much get all of the features that plus members were getting. Things like data analytics, vision and memory are also now available in the free version of ChatGPT. So as far as I can tell, for the most part, if you're on the free version of ChatGPT, you're getting the same things that ChatGPT plus members were getting. You just can't use it as much. You can't give it as many prompts, plus users get to use it a little bit more. That's pretty much the main difference now. All right, let's talk about Google for a second. Last week all those reports came out about AI overviews where Google was giving responses that didn't make sense, like how many rocks people should eat and how long ago. The year 1919 was and Google claimed it was only 20 years ago. Well, Google made a response to that, explaining what was going on. They claimed they did robust red teaming and they did a lot of evaluations with samples of typical user queries and tests on a proportion of search traffic. Then they go on to say, but there's nothing quite like having millions of people using the feature with many novel searches separately, there have been a large number of faked screenshots shared widely. Some were obvious and silly, while others have implied that we returned dangerous results for topics like leaving dogs in cars, smoking while pregnant, and depression. Those AI overviews never appeared, but they do go on to say, but some odd, inaccurate or unhelpful AI overviews certainly did show up. One area we identified was our ability to interpret nonsensical queries and satirical content. Let's take a look at the example. How many rocks should I eat? Prior to these screenshots going viral, practically no one asked Google that question. There isn't much web content that seriously contemplates that question. That's what they call a data void or information gap. However, there is some satirical content on this topic. So when someone put that question into search, an AI overview appeared that faithfully linked to one of the only websites that tackled the question. One of the satirical websites. They also saw responses that featured sarcastic or trolly content from discussion forums, but in some cases has less than helpful advice like using glue to get cheese to stick to pizza. So pretty much they're saying that a lot of the screenshots that were circulating, probably even some that I shared in my video last week, were likely faked screenshots. But the ones that weren't faked screenshots were content where there was a data void. So it ended up pulling content from satirical websites, or it ended up pulling content from platforms like Reddit or X that might have had, you know, trolly or funny comments on it. They go on to say that they built better mechanisms to try to solve some of these problems and that the AI overviews is going to get better. You ever find yourself having to walk people through the same thing over and over again to different people? I know with my growing online business, I definitely do and that's why I partnered with Guide on this video. Guide is an AI powered tool where you just go through and do the steps of the thing you want to show people and it will actually create a step by step tutorial in the process. Simply install the free Chrome extension. Make sure you got it pinned up in your top bar so you can actually see it. And now let's say I want to show a team member how to add a new piece of news to the Future Tools website. I simply click on the guide button, click Capture, select the type of content we're creating and give it a name. And finally we click Go. Now I simply just go through the steps. If I come over here, click on this button, click on News, click on New News up here, type the name of the news article, enter the URL, the source and the date and then click Save as Draft and then I will come through and review the drafts. I've walked through the step by step process. Now I can click the guide button again to end my recording and we can watch as the AI actually creates a full tutorial for me. We've actually got a video here now and as it walks through the video you can see it shows off each step that needs to happen. Click News, click New News, click Here, fill the name of the news article, click here. And as you can see it created a full on tutorial complete with screenshots and step by step instructions that now I can share this document with anybody on my team and they'll now know how to add news to the Future Tools website. I can now share this tutorial wherever I want or embed it wherever I need to to give team members access to this education guide is actually a tool that I was using before they even approached me to sponsor this channel. So I'm absolutely happy to tell you about it and to recommend it to anybody that finds themselves in explaining things more than once. And again, you can use the extension completely for free. I'll make sure the link is right at the top of the description below this video. And thank you so much to Guide for sponsoring this video. Also this week there's been some drama with Elon Musk. I mean, honestly, what weeks aren't there drama with Elon Musk? But this week there's some fun stuff in the AI world that happened. First off, Logan Kilpatrick, who was recently on the Next Wave podcast, by the way, asked how long until AGI and Elon Musk replied next year. So I don't know if Elon actually believes this or if he was just having fun, but it appears that Elon Musk believes that we're actually going to see AGI sometime within 2025. But here's the real interesting thread Elon Musk went to X to post. Join XAI if you believe in our mission of understanding the universe. Which requires maximally rigorous pursuit of truth without regard to popularity or political correctness. Then Yann Lecun, the chief of AI over at Meta, said, join Xai. If you can stand a boss who claims that what you are working on will be solved next year, claims that what you are working on will kill everyone and must be stopped or paused, claims to want a maximally rigorous pursuit of the truth, but spews crazy ass conspiracy theories on his own social platform. Somebody jumped into the comments and said, hey Jan, why don't you start your own AI company? Jan responded, I'm a scientist, not a business or product person. And then of course, Elon jumped in and really started stirring the pot. Elon said, what? What science have you done in the past five years? Jan responded, Over 80 technical papers published since January 2022. What about you? Elon replied, that's nothing. You're going soft. Try harder. Somebody else jumped in and said, were there beakers or petri dishes involved? No. Then it's not science. And then Yann Lecun jumped in and said, well, if it's not published, it's definitely not science. And of course Elon jumps in. This is one of the dumbest things anyone has ever said. Then Jan goes on to clarify that to qualify as science, a piece of research must be correct and reproducible, be correct and reproducible, it must be described in sufficient details, in publication, etc, etc. At the end of the day, both of these guys look pretty childish. Honestly, I have a lot of respect for both of these people. I think both of them have really contributed to where we're at in the tech and AI world today. But for Elon to call out Jan and say, what science have you done when he basically invented the Convolutional Neural Network and won the Turing Award. But then for Yawn to say that if it's not published, it's not science. I mean, both of them aren't looking great in this scenario. This is one of those, let's just sit back, eat popcorn and enjoy the show because it's been pretty entertaining to say the least. Since we're talking about Elon, let's talk about Xai. They just raised $6 billion in their series B funding round. And if you're wondering what this funding is likely going to go to, Musk is planning on building the XAI supercomputer, dubbed the Gigafactory of Compute. Supposedly this Gigafactory of Compute is going to consist of 100,000 H100 GPUs from Nvidia. This supercomputer would be four times bigger than current AI clusters. And shortly after this news came out that Elon was planning on building this gigafactory, Nvidia stock took a nice little jump here. And if we actually take a look at Nvidia's market cap, as of right now there were 2.8% to $2 trillion. I mean, it only feels like a few months ago that they passed the $1 trillion mark. Like the amount of growth they've had is absolutely insane. And if you look here, they're not too far off from overtaking Apple. Quite honestly, they're not too far off from overtaking Microsoft. At the pace they're going, I would not be surprised if Nvidia ends up being the number one largest company on the planet soon. Speaking of Microsoft, we got a little update from them this week as well. The Microsoft copilot that most of us paying attention to AI have become accustomed to is now also available inside of Telegram. We can see a little screenshot sample here where somebody asks what can you do? And inside the Telegram bot it gives them a response. Now the article doesn't specifically say, but I'm assuming this is also powered by GPT4anthropics. Claude got a little bit of an update this week and it can now use tools. Basically they just rolled out a new feature where if you're using the CLAUDE API as your large language model of choice, if you're a developer, Claude can now contact other APIs and work with other tools around the web. Now this is something we've had with chat, GPT and the various GPT APIs already. But now if you're a fan of using CLAUDE in your development processes, you can do it with Claude now as well. The French AI company Mistral AI just released a new model this week as well called Codestrol. This is Mistral's first ever model that is fine tuned for coders perplexity. I also rolled out a new feature this week called Pages which essentially allows you to make like AI generated almost Wikipedia pages. Check out this page that was automatically generated on Yann Lecun. Here we can see if we scroll through it, it's got a little menu on the right. Lecun's early life and education. More details about them here. Bell Labs, AT&T Labs and we can scroll through and see a pretty detailed history and story of Yann Lecun. Now this Wikipedia style all encompassing page can be generated by anybody on any topic. If you go To Perplexity AI. Click on the little library link on the left and click this button to create a page. It asks you what's your page about? Let's go ahead and create one. Let's call it YouTube Growth Strategies. Hit the little blue button here and we can see as it generates a brand new page with information and sources for that information, similar to what you might expect from a Wikipedia style site. And here's the page it just generated. Leveraging YouTube Shorts for rapid growth, creating effective video playlists, utilizing YouTube analytics for content optimization and then we can ask it to add even more sections like monetizing YouTube shorts or collaborating with influencers. Let's go ahead and add that. And within a few seconds I just added a new section to my page here. With five different sources, I can now publish this page and share it with anybody I want. Just like that Yann Lecun page I just showed you. If you're a fan of using Canva to create art, which is a tool that I find myself using quite often, Canva just had an event last week called Canva Create where they unloaded a ton of AI announcements. I found this X thread which breaks down 19 of them and I'll share it in the description below. But some of the standouts include this magic grab feature where you can select an object within an image and drag and move it around, an enhanced voice feature similar to Adobe's audio enhancement features, and a new text to graphics feature which looks pretty impressive. This week Google DeepMind also put out some new video clips showing off their new VEO model, including this time lapse of the northern lights dancing across the arctic sky, stars twinkling snow covered landscape and that looks pretty impressive to me. Here's some other examples of a waterfall which doesn't look quite as realistic, a fast track shot down a suburban residential street looks pretty good to me. Extreme close up with a shallow depth of field of a puddle in a street. An aerial shot of a lighthouse stands tall on a rocky cliff. If you just scrolled past this on X or something you would assume this was a real drone shot. Pretty hard to tell. That's AI. It's pretty well done. And while we're on the topic of AI video, George Lucas was recently asked about his thoughts on AI video. Here's what he had to say.

We've been using it for 25 years and it's not AI, but we use all the digital technology because we helped pioneered a lot of that, especially at ILM because for a while they were the only Place that was doing digital. The thing of it is that it's inevitable. I mean it's like saying, I don't believe these cars are going to work. Let's just stick with the horses. Let's just stick with the horses. Yeah, you can say that, but that isn't the way the world works.

So when everybody out there is talking about how everybody in the film industry is scared of AI and trying to avoid it, that's not necessarily true. There are people in the film industry that are really excited and actually wanting to use AI and as George Lucas says here, it's kind of inevitable. If you're into having fun with AI music like I am, Suno is rolling out a new feature where you can import any sort of sound you want and turn it into music. Check this out. I think that's gonna be so much fun to just like make sounds in the real world like with your water bottle and a pair of tweezers for some reason, and then turn it into a song with Sudo. I'm looking forward to that. That's going to be fun to play with in the world of AI research. This week we learned about insta drag. Lightning fast and accurate drag based image editing emerging from videos. Now if you remember a few months back we got this tech called drag your gans where you can generate images with AI and sort of manipulate those images by dragging points around on those images. And there was examples where like there was a lion sitting there and they would drag the nose of the lion and the head would move on the nose of the lion and you could do some really interesting stuff. Well, this looks like you can do that same kind of thing, but with real images. It doesn't have to initially be generated with AI. We can see some examples here where there's a woman holding up a camera and they pull the camera down with this down arrow. And most of the other models kind of messed it up. But if you look at insta drag, it just reveals the woman's face. Here's another example where there's this girl holding this star in front of her face. They sort of mask out this area right here, drag a little arrow down to try to move the star and then we get this output where the star is moved down and now it reveals the girl's face. Here's another one where they masked off a kid's mouth here and then put some lines to make it look like he's smiling. You can see like the arrows trying to pull his lips apart there a little bit. And the final result looks more like a child smiling. And there's all sorts of examples on this page. This Demo video is 8 minutes and 14 seconds long. I will definitely link it up below. But there's some pretty cool examples of what you can do with this that'll be really fun to upload our own images into this and sort of manipulate the images that we took in the real world. Just another exciting toy that I can't wait to get my hands on. Another piece of research that was shown off this week was Sign LLM, which is sign Language production Large language models. And it's pretty much what it sounds like. You enter text and it will actually generate an animation of somebody doing the sign language for that text. Once again, I'll make sure the research paper is linked up in the description below. So if you want to dive deeper and see more of these examples of people inputting text and getting an animation of someone doing sign language out, you can find that below. It is pretty dang impressive. The main problem with it is that sometimes AI is still not great with hands. So if you look closely at some of these videos, some of the fingers get messed up. And I don't know sign language myself, but I would imagine if the hands can't make themselves look right, it might be hard to understand what the hands are saying. Now let's talk about some AI gadgets. A team at the University of Washington developed some AI headphones that let the wearers listen to a single person out of an entire crowd just by looking at them. Once, the headphones know which voice to focus on based on the direction.

When you look at a particular person, the mouth is located equal distances towards your ears.

And so a sound will arrive at the same time across both ears, which is a cue for the headphone system. And the headphone system uses AI technology to extract the voice that wearer wants to hear while ignoring all sounds in the environment from that point on.

So now we're outside, we're standing in front of this lovely fountain here. That's very noisy.

So if you want to have a conversation, you can go ahead and click that button on your right and record my voice, not listen. And then we can have this, you.

Know, casual slog here and you can just listen to my voice while pushing. The town of the snow is out there.

Once I have this example of how they sound like, I can just extract their sound even when I'm not not looking at them. This week, a company called IO Iyo introduced a set of generative AI earbuds trying to do similar things to what the Humane pin did and the rabbit did, but just staying in your ears. And if we scroll down, one thing that I kind of have an issue with is how massive they are. They look huge. The idea is that it's connected to a large language model and you can have conversations with it and it just sits in your ear. They believe with earbuds like this, they're going to be millions of what they call audio first apps. So I imagine you'll just be able to talk to these earbuds and order an Uber or Doordash or a lot of the things that the Humane Pin and the rabbit showed off that you'd be able to do, but didn't quite work as expected when people got their hands on those devices. And since we're talking about the rabbit, let's add some more drama into the mix this week and talk about this guy here, Coffeezilla. He got his hands on one of these rabbits and did a little bit of digging behind the person who created the rabbit and made a video about how the creator of the rabbit, before he made the rabbit, was involved in some crypto scams. He then asked the creator of the rabbit about his new large action model and sort of came to the conclusion that the rabbit and the large action model are a bit of a scam. Now, I don't necessarily know if I agree that it's a scam. I think maybe it was just released way, way, way, way too early and they should have waited until it was cooked a little more before they put it out to people, because pretty much all the videos I've seen, the product just fricking doesn't work. Now, I ordered one months ago. Mine hasn't arrived yet, so I'll put it through the motions and share my actual thoughts when I get my hands on one. But all of the videos I've seen so far just show it not working. Like it doesn't connect to the Internet. The DoorDash feature doesn't work. I think it had Uber in it, but the Uber feature doesn't work, and pretty much everything they showed off in their launch demo doesn't work for most of the people. And Coffeezilla here in his video tests almost all of that, and in his video, it doesn't work for him. It is an interesting video. I'll make sure it's linked up in the description below so you can check it out and see his whole breakdown. But quite honestly, you'll watch this video and you'll probably start to wonder if this rabbit R1 was just one giant money grab with no real intention of ever actually building what they promised. I don't know. Watch his video. I'd love to hear your thoughts on what you think of the rabbit R1. And in the final gadget announcement in this week's video, Google Adds AI Powered Features to its CHR Last week during Microsoft Build, we got the announcement of the Copilot Plus PCs, which are PCs pretty much designed with AI in mind, and it sounds like Google is kind of following up with their Chromebooks that have AI in mind. A lot of the features we've seen Google show off already appear to be features that are just going to be available inside of new Chromebooks. You're going to be able to do generative art directly inside of your Chromebook, help write emails with AI in your Chromebook, and much, much more. It's pretty much all of the features we've seen Google show off that they're going to put inside of things like Google Docs and Google Drive just sort of built into all the Chromebook stuff as well. And that's pretty much all I got for you this week. There is some exciting stuff that I'm anticipating next week. I heard a little rumor. I don't know if it's true, but it did come from a very credible source that we could expect to see stable diffusion 3 finally roll out and US consumers actually get access next week. Again, no guarantees. It's a rumor, but I have it on good authority we might see it next week. Also next week is the Cisco Live event out in Vegas. I am planning on being at that event, but if you take a peek at the Cisco Live page, AI takes center stage at Cisco Live 2024. So the event is going to be super focused on AI, which I imagine is why they invited me to attend. I know a lot of their announcements are going to be around security and preventative measures for hacks and scams and things like that related to AI. I don't know what they're going to be announcing yet, but I'm excited to find out and I'll be at the event doing my best to share what I learn. Also, next week is the Computex event out in Taiwan, and at Computex we know we're going to get a keynote from the Nvidia CEO Jensen Huang, and he's most likely going to make some pretty big announcements about what Nvidia is working on. In fact, I've already gotten a pre brief, but I'm under embargo still, so I kind of know some of the announcements they're going to make, and there's some exciting stuff that's going to come out of Nvidia during Computex. But he's not the only person giving a keynote. His cousin, the CEO of amd, Lisa Su, is also giving a keynote all about how AI is driving a revolution that is rapidly reshaping every aspect of computing and technology. We're also getting keynotes from the CEO of Qualcomm, the CEO of Intel, and a handful of other companies, so probably even more announcements that we can expect out of this event. I'm super excited. I'll be at Cisco Live while all of this is going down, but definitely still keeping my finger on the pulse of what's going on during this Computex event as well. And if you're following my channel, trust that you'll be kept in the loop, because I'm going to be trying to break it all down as I learn about it myself and share videos with you. And if I don't have the time to share videos, at the very least I'll probably be sharing my thoughts and revelations over on X, because I like to tweet X, post whatever you want to call it over there whenever I hear about cool, interesting news. So make sure you're subscribed to this channel if you're not already. I will make sure you are looped in with all the latest news, all the coolest AI tools, and give you the occasional tutorial to break down how to do some really cool stuff with this AI tech. And if you're not already, make sure you check out FutureTools IE specifically, join the free newsletter where every week I will share just the most important news that you need to know and the coolest AI tools that I came across. And by joining the free newsletter, you'll get access to the AI Income Database, a free database of all sorts of cool ways to make money using AI. And you'll be entered into our monthly competition where I'm giving away cool gadgets and software and things like that. Every month for the month of May, I'm giving away an insta360x4. Previously I've given away Nvidia GPUs. I'm trying to get cool gadgets and turn around and pass them along to viewers of this channel and subscribers of the newsletter. So if you want to be entered for those competitions, make sure you're subscribed to me here on YouTube and you're subscribed to the newsletter over at FutureTools IO and I am going to hook you up. So thank you so much for tuning in today and nerding out with me. It has been a crazy week. Lots of drama in the AI world, Lots of news. It's so much fun to me. I love keeping my finger on the pulse and trying to loop you in. I'm a huge nerd over this stuff and I appreciate that there's others out there that love nerding out about this stuff as much as I do. It still blows my mind that people actually watch my channel and enjoy nerding out with me, but I am so grateful that you do and I can't thank you enough. Thank you once again to Guide for sponsoring this video. You guys rock as well. Appreciate you all. Thank you again for tuning in. I'll see you in the next one. Bye. Bye.