THIS Artificial Intelligence will SURPRISE you

What are three major breakthroughs in artificial intelligence that you must know about? We've all been introduced to ChatGPT, and it's been pretty incredible how powerful it is, and it's even got some of those super AI dominating world vibes. The thing is, what happens after ChatGPT? I took a deep dive, and the developments are pretty darn insane. I would say that they would totally eclipse what we've just seen with generative models here. What I've done is that I've divided the latest developments in artificial intelligence into three major categories. The lucky thing about us is that a lot of these technologies are open source and ready for investment, and I promise you guys, they have more friendly llama vibes than they have Terminator vibes. So we're going to take a deep dive in this video series, and this is the first video, so it'll serve as a foundational for what is going to happen for explaining deep dives into AI. This video is not sponsored by any AI projects, but it is sponsored by Bybit. Bybit is a top cryptocurrency exchange, and now they are listing new coins the fastest with the most volume, which is pretty insane if you want to get in on all that crypto action. If you do want to join the Bybit exchanger, it is a limited time offer, which is $50 free US dollars if you do the first KYC and you load more than $100 into the exchange. If you're interested, sign up with the link down below and you'll automatically get in this offer. Remember, it's only for the first 100 people to do so, so what are you waiting for? So starting this video, this AI, this is GPT-4. It is writing a video script based on information from this webpage. On top of that, it is able to search for web results to get additional information about what this is. This is making my job, half of what my job is, even easier because I can let it browse any website, create interesting and humorous content because I wanted to be humorous. And this is just a simple demonstration of the current state of what GPT-4 can do. Now, this is amazing. And in the past few months, my mind was blown by how much GPT can do. The problem here is that GPT-4 and the parent company OpenAI is owned by Microsoft. If Microsoft decides to cut off anything, then yeah, the community is kind of left out to try. Because every time we need GPT to do something, we need to communicate with GPT servers. And this does mean that we're dependent on what GPT can provide us. And we're dependent on what GPT decides to, or OpenAI decides to research, and they control everything that we have access to. And this is not great for global research or anything, which leads us to the latest developments here with large language models that are open source. Now, the good news here is that Facebook decided to create something similar and release it open source, with technically speaking, the tuning parameters being kept secret. But what happened with the community was that someone leaked these tuning parameters, combined these together, and bam, we're still going in many ways. The Lama model came out and Stanford made some modifications and trainings onto this data model. And this progressed for the past few weeks into models like Vicunia, which kind of brought this to 90% of what ChatGPT-4 is doing, but on your local machine. What's even more crazy, all right, this just happened yesterday, is that this can now all run on your browser. To an outsider, this might look very normal. It's done here, it's in the browser, this is done in the browser, what's the difference? What's crazy is that this means that you're essentially running the brains on your local machine. If you use the Terminator model, where there was a giant computer called Skynet, Skynet kind of controlled the world. Initially, Skynet is in Microsoft, right? But what this model allows us to do is it runs Skynet locally on our own computer. So even a computer like a MacBook can start having the capabilities of the large language model. The scary part here is that once the AI can run on your own computer or anyone's own computer, it's unstoppable. Because anyone that's downloaded this now can just start running and tapping away and customizing their own AI. Which means that the only way for us to beat AI is to have the best AI. You just can't go to Microsoft and shut off those servers anymore because Skynet is everywhere. That's the current state of AI and AI research. So the other crazy thing is, is once you start learning how they train these models, it's super hacky, I would say. And it's super scary in a sense that AI can be used to train more AI. You definitely see that acceleration in speed of AI research is because now you can use GPT-4. Technically, you're not supposed to. That's against copyright. And I'm not encouraging you guys to do this. You know, people are going to do this endlessly. That's what's powering AI research. But what's happening is that they're using GPT-4 to train their language models to make that even stronger. We are witnessing the fastest growth of AI research in history. Now, does this open up an incredible opportunity for the world and crypto and us in crypto? I think absolutely. I am every day trying to figure out if we can use AI to cut down workflow and make the boring stuff happen faster. And that does include making funny jokes or something like that, right? So if jokes suck, it's not because of me. It's because of shitty AI. And what does that make ketchup? Oh, brother, this guy stinks! But beyond that, I feel like there's an immense opportunity here because there's billions of dollars of jobs at stake that can be potentially replaced by AI now. I feel like there's three doors that open. So the first one is the one we discussed, which is training your own models using these open source, large language models to build your own custom GPT-4. I think that's 100% one of the biggest opportunities that are happening right now. Second opportunity is something like Jarvis or Auto-GPT, which is essentially letting GPT feed itself and refine itself. So it can complete a certain objective. So instead of just writing a video for me, it can continuously refine this video concept and make it better to complete that full task, maybe including hiring a voice actor to say out these words, generating the images that are required to make this video happen and done, done, right? So the whole idea of automation and GPT training automating itself, this is huge. So that's kind of being pushed by either Jarvis, which is something yet again released by Microsoft, and then also with Auto-GPT, which is an open source project that can continuously let GPT run instructions with itself. And I definitely see this as kind of like the number two biggest opportunity here in AI. The last frontier for AI is the application side. So now that we know that the technology is, well, pretty sound and pretty darn useful, now it's about applying it into particular industries and then using those results to dominate that industry. Recently, I've been investing in crypto projects, and I think this is where crypto projects can come into play. So something like PodFast. So this is Jeff Kordakis' project, he's a friend of mine. But what he's essentially doing is he's using AI to distill these large podcasts that can be ours into these short videos that are both animated and exciting. That's actually quite insane if you think about it, because there are companies that do that for books. I use Blinkist, which summarizes these large books into something that could be read and listened to in five minutes. But Jeff is doing that for podcasts. It does it into a video format at the end. That's kind of crazy if you actually think about how much we left, right? Because if you think about Blinkist, they hired an entire team to read the freaking book and then make it into a summary. But what this AI is doing is not only is it reading a book because they're going on a podcast. Podcast is voice, right? So they're translating the voice into text, into material, that large language that can understand, and then use generators to convert that into what the machine can understand in large language model. Then pushing that out, making a summary, then converting that to a video. So think about that. That's not only doing what entire companies have been doing and setting their entire business models are doing, but it's going a step beyond that, right? So the technology is going to be what powers everything. And the application side is what's going to change the world. This is just such a small example of it. I mean, this is just something time saving for me because I create podcasts and listen to podcasts, but I do realize that podcasts can be long. There's so much good Intel there and distilling this is essentially what smart people should be doing. But the fact that computers can do that now, my God. And I think this is going to be applied to every single industry. It's going to make the people who are even writing articles, creating videos, holy shit, this world is going to change guys. So applications is going to come up next. And I think this is the last frontier, getting this AI and applying it to different fields successfully has been great. So I'd love to hear what your examples are and what you're interested in. So there's either the language to model training or combining these models and using it to do automation or the applied AI side. I think I'm very interested in seeing which kind of segment you guys are super interested in or to talk about. I definitely want to make more content on AI. I feel like we can learn together as a community and this is the power of the Box Money community that can make something like this great. I hope you guys can input more into this and tell me what you guys like to watch. So thank you guys for watching this video. See you guys next time.