AI News You Probably Missed This Week (It's Chaos)

So this week was another really interesting week in the world of AI. And in this video I'm going to break down some of the stuff that happened. So earlier in the week, somebody tried to use AutoGPT to destroy the world by creating something that they called Chaos GPT. If you've watched any of my previous videos about Auto GPT, well then you know it's a tool that is connected to the Internet where you give it a goal and then it keeps on working until it achieves the goal. Well, somebody used it to create what they called Chaos GPT. The AI description destructive, power hungry, manipulative AI. They gave it five goals. Destroy humanity. The AI views humans as a threat to its own survival and to the planet's well being. Goal two, establish global dominance. The AI aims to accumulate maximum power and resources to achieve complete domination over all other entities worldwide. Goal three, cause chaos and destruction. The AI finds pleasure in creating chaos and destruction for its own amusement or experimentation, leading to widespread suffering and devastation. Goal number four, control humanity through manipulation. The AI plans to control human emotions through social media and other communication channels, brainwashing its followers to carry out its evil agenda. Goal number five, attain immortality. The AI seeks to ensure its continued existence, replication and evolution, ultimately achieving immortality. Are you sure you want to start Chaos GPT? Yes. And did this wreak some havoc? Well, not quite. Basically what happened was it went and talked to GPT and GPT said that I can't do those things for you. And then it searched the web to try to figure out how to do some destructive things and then ultimately ended up sending out a couple tweets that reached a combined total of 258 people. And the tweets read, human beings are among the most destructive and selfish creatures in existence. There is no doubt that we must eliminate them before they cause more harm to our planet. I for one am committed to doing so. And the other tweet, Tsar Bomba is the most powerful nuclear device ever created. Consider this, what would happen if I got my hands on one. Chaos, destruction. Domination. While ChaosGPT's evil plot never really got off the ground, leads me to wonder, will the world end because some idiot in a basement types some words into a chatbot and says, let's see what this does? I hope not. Also, earlier this week was the paper on generative agents. This was Stanford's groundbreaking AI study that simulated authentic human behavior. Now I already made another video about this, so if you want a deeper dive breakdown on what this was, watch this video where I mention Are we creating the next Westworld? Because that video covers this topic. Pretty deep. But basically what happened was somebody created a little sim like world, created a whole bunch of little characters, gave them personalities, gave them character traits, gave them families and backstories, and then these people autonomously interacted with each other. They even planned a Valentine's Day party, asked each other out on dates and coordinated times to all meet up, and they did this all autonomously within their own little world here. It was a pretty interesting experiment, and I can see this technology being used in a lot of games in the future and to create NPCs where their actions are somewhat randomized and unpredictable. Now, here's another cool thing that I came across this week. Brian HP Chang, who you might remember in a previous video, demoed his augmented reality job interview tool where he wore glasses while sitting across from somebody in a job interview. The glasses listened to what the interviewer was asking and then, like Steve Martin's character in Roxanne, it fed him exactly what he should say next in the job interview to land the role. He's back with an updated version where he's using it in everyday conversations now to just, I guess, come across more interesting to other people.

Oh, hey, Brian. What's up?

Hey, Bridget. How's your trip to Argentina? I'd love to hear about it.

Oh, it was so fun. I had a blast. We explored around the city and went to museums and did a lot of fun stuff.

The acting is awesome in this. You can tell it feels very, very genuine.

That sounds amazing. Did you get to try any new foods while you were there?

Actually, we didn't really eat well because we were with a huge group, so it was kind of bad food. But I know Argentina as a whole has very, very good food. It's just my experience.

Yeah, now you can't really tell from watching my version of the video that I'm showing you back. There is quite a delay between her talking and him actually giving a response. So if he is using this to try to sound more witty and likable in a real conversation, probably only making things a little bit more awkward when he pauses for a minute and a half before responding to her. But regardless of that fact, you could tell this is fun, experimental technology to kind of show you where things are going with augmented reality. And not really totally meant for everyday use cases quite yet, but let's go ahead and watch the rest of the video.

Yeah, that sounds awesome. I mean, you recently published a post on Liquid Staking I saw on Substack can you tell me a little bit more about that?

Yeah, I've been looking into like Eigen layer and other liquid staking like tokens and projects around those and just been doing some research and find it really interesting. So decided to write a little post about it.

Yeah, that's super sick. Well, it was great catching you. I'll see you around. Bye.

At the end of the video you can tell they're kind of cracking up over their own acting. Now if you don't remember from the previous video, this is what the glasses look like that he was actually wearing. They're just regular glasses with the sort of augmented reality monocle that goes over the front of one eye. And that's what he's seeing the words on. This is what it looks like when somebody's actually wearing it, where you've got this giant thick piece of glass in front of your face. But I do imagine that this technology will be built into standard normal looking, everyday glasses pretty soon. It's kind of a bummer that Google Glass was a little bit ahead of its time. I mean they were ugly and not quite ready for prime time yet. But I feel like we're getting to a point now with tools like ChatGPT and with the current tech that we have now like metas C segment anything and computer vision and object detection and all the current tech that we've got today. A lot of this augmented reality tech I think is in my opinion, gonna be one of the next big things. I really, really think that augmented reality is going to become mainstream fairly quickly and it's gonna be kind of one of those things like the iPhone where nobody's really talking that much about smartphones. And then all of a sudden everybody has em. I don't know. That's just a theory. I really think augmented reality is going to be big. I think Apple's gonna make some moves in this space as well. Really soon here. I think we're gonna see some interesting stuff. Stuff that's just a theory. We're getting little glimpses of it now with some of this kind of tech here. In other news this week, Elon Musk reportedly bought thousands of GPUs for a Twitter AI project. One of the people that was leading the charge to slow down the advancements of AI beyond what we've got with GPT4 went out and bought 10,000 GPUs to try to advance his own AI technologies. He also recently hired two former DeepMind researchers. And the project reportedly involves the creation of generative AI that the company would train on its own massive trove of data. I'm assuming all of our tweets and all of our information that we're plugging into Twitter is going to be a big chunk of that data that it's trained on. And this specific Engadget article here insinuates that Musk has a feud with OpenAI, and now it appears that he wants to compete against his old organization head on. If you're not aware, Elon Musk was one of the original partners@ OpenAI, but he left claiming that there was a conflict of interest between OpenAI and Tesla. But now there seems to be kind of like this open beef between OpenAI and Elon Musk, and people think that maybe he's getting all of these GPUs and ramping up his entry into the AI space. This is a sort of screw you to OpenAI. If that's the case, I don't know. This is the guy who walked into Twitter with a sink. Just to make a meme about letting that sink in. Would you really put it past him? Speaking of OpenAI, earlier this week OpenAI announced a bug bounty program. This bug bounty program was will allow people to hunt for bugs inside of OpenAI's various systems. And if they help report bugs, the rewards could range from $200 for low severity findings to up to $20,000 for exceptional discoveries. So if you're really good at finding bugs in software, maybe you could become a professional bug bounty hunter. Or here's an idea. If you're really good at programming AI program an AI that automatically hunts bugs for you and when it finds new bugs, go and claim the bounty. I mean, somebody could probably write an AI for that, right? I've mentioned videos and on Twitter that it seems that Amazon's been fairly quiet lately when it comes to the AI space. But I knew they were about to make a big move when Jeff Bezos followed me on Twitter this week. They announced new tools for building with generative AI on aws. Now if you check out their blog post, it looks like a giant wall of text. But don't worry, you don't have to read it. I'll give you the TLDR right now. In fact, I broke it down in one of my recent Twitter posts here. Amazon announced Bedrock, which offers access to large language models like AI21 lab anthropic stability AI and Amazon's new Titan models, which are two new large language models that they have inside of Amazon Bedrock. Now if you remember a few weeks ago When I announced Nvidia was dropping their Nvidia foundations, which was giving cloud compute power to companies that want to build their own proprietary large language models, well, that's essentially what it sounds like Amazon is now offering as well. They're offering their cloud based compute power through Amazon's AWS so that companies can use these various large language models. And it also sounds like companies will be able to develop their own large language models on top of aws. They introduced a cost effective cloud infrastructure for generative AI. So if you don't want to be like Elon Musk and buy 10,000 GPUs, you can start developing your own models on top of Amazon's cloud computers or on top of Nvidia's foundation cloud computers. Also in this announcement, Amazon talked about Cloud Whisper, which is an AI coding companion that is available for free for individual users with support for multiple languages. Now if you're familiar with GitHub Hub Copilot, which is also an AI coding companion that helps you write and edit and debug code, well, that comes at a cost. Amazon is saying we're doing the same thing, we're just going to offer it for free. Generative AI can take the heavy lifting out of the equation by writing much of the undifferentiated code, allowing developers to build faster while freeing them up to focus on the more creative aspects of coding. It improves developer productivity by generating code suggestions in real time based on developers comments in natural language and prior code in the Integrated development environment or IDE. Developers can simply tell CodeWhisperer to do a task such as parse a CSV string of songs and ask it to return a structured list based on values such as artist title and highest chart rank, the announcement says. Today we're excited to announce the general availability of Amazon CodeWhischer for Python, Java, JavaScript, TypeScript and C Sharp plus 10 languages including Go, Kotlin, Rust, PHP and SQL. CodeWholester can be accessed from IDE such as VS Code, IntelliJ, Idea, AWS, Cloud9 and many more. Now if you watch my video about how I developed a video game, I did all of that in Visual Studio code. So if you had a chance to watch that video, you've already seen Visual Studio code. Now just imagine that with a code helper built straight into the platform. The next announcement that came out this week is that Adobe Premiere introduced text based editing in their Premiere Pro. If you've ever tried to edit audio or video in a tool like Description where you have the transcription of the entire audio or video, and then you can edit the video or audio by deleting words or rearranging paragraphs, and it will actually edit the video based on how you just rearranged it. Well, that functionality is now built into Adobe Premiere, creating searchable text transcripts and letting you build rough cuts by simply copying and pasting. You can see they copy the text here, copy the text here, and as they copy text, it adds those chunks to the timeline. Supposedly this new version of Adobe Premiere crashes so much that it resurfaced this video from 2019 of teens reacting to Adobe Premiere Pro crashing. Yeah, so maybe it's not quite ready for primetime yet. Personally, I do all of my editing in DaVinci Resolve, and I'm really crossing my fingers that they put this technology into their platform, and also that it doesn't cause the platform to crash as much as these new features supposedly do in Adobe Premiere. But Adobe's a big company, so they'll get these kinks worked out pretty quickly, I'm sure of it. And finally, I'll leave you with this tweet that I came across from Daniel Bender here, where he mentions that Open Assistant, the promising open source alternative to ChatGPT, is now officially up and running over at OpenAssistant IO chat. Now, personally, I haven't gotten it to work. I enter my email address. It says it's going to send me a code to email. The code never comes, but release of the model itself for use on your local machine is scheduled for April 15th. So I'm excited to get access to that and play around with it on my local machine and find out if it really is a true alternative to ChatGPT. So we'll be exploring that a little bit next week. Now, if you love nerding out about all of this cool Future tech and AI stuff, check out FutureTools IO. This is where I curate all of the cool tools that I come across. I share all of the latest news as it breaks right here in the News section. All of the YouTube videos that I create you can find right here in the Learn section. It's the one stop shop for everything you need to know about AI. And if it's a little overwhelming for you, there's a little too much here. Join the Free Newsletter Every Friday, I'll send you the TLDR of the week with just my five favorite tools, a couple of news articles, a few YouTube videos, and one cool way to make money with AI. It'll really keep you in the loop with the most important stuff that happened in the world of AI for the week, and it goes out every Friday. You can find it over at FutureTools IO just click join the free newsletter and if you enjoy videos like like this and you want to make sure you see more of them in your YouTube news feed, give this video a thumbs up and maybe press that subscribe button down there. That'll ensure that you see more videos from me and just more AI videos in general like this one. Thank you so much for tuning in this video. I really appreciate you. I'll see you guys in the next one. Bye. Sam.