2024: The Year of AI (My Honest Predictions)

A Jan 3, 2024 YouTube commentary by Matt Wolfe presenting his predictions for AI in 2024. Wolfe forecasts wider adoption of personalized AI assistants (tied into email, calendar, storage), rapid improvements in AI-generated video quality and length, advances in personalized AI education, healthcare diagnostics and drug discovery, smarter in-game NPCs and AI-assisted game creation, a rise in AI-generated influencers, progress in robotics and self-driving cars, ongoing open-source vs closed-source LLM tensions (mentions Mistral AI), election-related deepfakes and content moderation, growth in XR/AR/VR (Apple Vision Pro, Meta Quest 3, Xreal), and increased litigation and copyright debates. Tone is speculative-optimistic with repeated caveats about timeline uncertainty.
Chapters
A Happy New Year 2023 was a wild ride in the world of AI, and I personally think 2024 is going to be even bigger. So before we get too deep into the year, I wanted to share some of my predictions of what I think are going to happen in 2024. And to kick this video off, I wanted to start with a Tweet here from Dr. Jim Fan, an Nvidia Senior Research scientist and lead of AI agents. He says 2024 will be epic. We are riding the vertical. Part of the exponential from now on will be state of the art by far. Now, before I actually get into it, I do want to give a few quick disclaimers. My predictions and what I'll be sharing is based off of information I've come across, articles I've read, conversations I've had, and from essentially just deep diving and being down the rabbit hole all of the past year. However, although I have made a lot of correct predictions, most of my timelines have been way off. I don't know how many times I've said things like I bet within two years we'll have X and then three months. The thing I was referring to was a reality. Things are really picking up at such an exponential pace that I am having a really, really difficult time putting timelines to things. I can sit here and tell you about what I'm pretty sure is going to happen in 2024, but to say it will happen in a few months or a few years is always the hardest part for me. But these are the things that I believe we will see in 2024 at some point. Starting with the fact that I think 2024 is going to be the year of AI assistance. I think we're going to start to see a lot of these large language models and chatbots be more and more trained on our own specific needs. They'll be trained more on how you speak, the specific niche topics you're interested in, your past conversations with it, and much more to personally tailor the experience to you. And everybody will be experiencing their own AI chatbot assistant in a different way. It'll also be a lot more integrated into a lot of what you're doing on a daily basis. It will have access to your calendar if you give it access. It will have access to your email, access to where you store your documents like Google Drive or Dropbox. And instead of interfacing with all of these individual tools, you'll likely just be interfacing with your assistant who can then respond to emails on your behalf. Schedule things in your calendar on your behalf, book flights for you, book hotels for you. Depending on how much access you want to give it, you'll just be interfacing with your AI assistant and and it's going to do a lot of the actual work for you. And if you're using your AI as a content creator to write YouTube scripts or blog posts or interview questions for a podcast, it will be much more trained in on how you speak, your style, the types of questions you ask. And it will get better and better at mimicking what you are looking for as you use it more and more. This is where tools like Gemini and ChatGPT and the next generation of large language models really want to go. I also think 2024 is going to see a huge up leveling in AI video generation. Just in 2023 alone we saw huge leaps from the beginning with Model Scope all the way to where we are now with Pika 1.0 and Runway gen 2 generating pretty amazing videos. I think AI video is going to have its mid journey V5, V6 leap where we get to the point where videos are being generated, where people are having a hard time actually noticing that the video was generated with AI Right now. There is no mistake. When you see an AI generated video, you know it's AI generated. I think that's going to change in 2024 and the quality of the output is just going to be next level to the point where you can no longer distinguish between an AI generated video and a video shot with a camera. Just looking at this progression here of 2023 that Justine Moore shared on X here we can see this timeline of generative video. Q1 through Q3 saw some fairly marginal improvements in AI generated. And then in Q4 we just saw an explosion with stable video diffusion. Leonardo's motion Pika 1.0 animate anything make Pixels Dance Motion Brush so much happened just in the fourth quarter of 2023. So we're actually seeing this exponential curve of AI video happening in real time right now. And I imagine it's only going to pick up more Steam in 2024. I also think we're going to get a lot longer generations. Right now we can get about four seconds and then we can extend those videos by four seconds at a time. But as we extend them, the quality of the video tends to degrade a little bit with each extension. I think that's going to improve a lot and we're going to be able to generate multi minute videos with AI. I don't think in 2024, I could be wrong, but I don't think in 2024 we're going to have that. I type in a prompt and it's going to spit out a 30 minute episode of south park for me. I don't think we're see that in 2024, but I do think we're going to see some much higher quality AI video along with a lot longer generations in that higher quality. In the same way, I think we're going to see AI assistance really help us in our day to day life in what we do online. I also think we're going to see a big leap in AI and education. Right now, the current education system is sort of a one size fits all system where students go to school and they're all kind of learning on the same way and being taught with the same methods. AI I think can improve this a lot in that everybody will be able to have their own mentor that teaches the topics they need to learn, but in the way that they learn best. So if you're much more of a visual learner, you need more video and images to help you understand. You'll likely have a niche AI bot that can teach you what you need to learn in that style for you. If you tend to need to read or hear it more audibly, it will give you the information in that way as well. If we take a peek at Bill Gates notes here, this was also one of his predictions. You can see, he asks, can AI bring personalized tutors to every student? The AI education tools being piloted today are mind blowing because they are tailored to each individual learner. One of the things that excites me the most about this type of technology is the possibility of localizing it to every student, no matter where they live. For example, a team in Nairobi is working on Samonasi, an AI based tutor that aligns with the curriculum in Kenya. So that's how AI is going to help education on a global scale. But I also believe you'll be able to find tailored tools and tailored bots to exactly what you need to learn in the way that you best are capable of learning it. One of the areas I think we're going to see the greatest advancements in AI and for humanity is in AI and healthcare. In 2024, I think AI is going to help create better and faster medical diagnoses. Using things like computer vision and diffusion models, health care professionals are going to be able to more quickly and more accurately look at medical imaging and find issues. Also, as larger amounts of medical information get added into these large language models, health care practitioners are going to be able to better import symptoms of an issue and, and get diagnoses back based on all of the symptoms that they load in. Because there's going to be so much larger amounts of data available to pull information from. I also think AI is going to help with better and faster drug discovery. AI is really, really good at finding correlations that humans may not necessarily see the correlation between. And with large amounts of training data on various chemical compounds, natural compounds, what's worked for people, what hasn't worked for people. All of that information being accumulated into a single data set that we can then access through large language model chatbots should help us find new drugs to cure diseases and solve problems that we previously haven't been able to cure or solve. And the same goes with new vaccines to prevent the spread of diseases. Now, a lot of people might argue that there's all sorts of red tape and trials that must be gone through before these new drugs and vaccines can be used, but I'm also hoping that AI will help speed up a lot of those processes as well. Will we see a lot of new drugs and things like vaccines on the market in 2024 as a result of AI? Probably not. But I do imagine that in 2024 we're going to hear about all sorts of new AI discoveries for a lot of these drugs and vaccines that are going to be ready for human trials much quicker than in the past. And maybe by 2025, this pipeline to get some of these new drugs and solutions on the market will speed up vastly. But we'll see. This is one of those areas where my timelines could be way off. We're going to see big leaps in gaming in 2024 as a result of AI, starting with AI NPCs, where you're going to be able to walk up to characters in video games and the conversation you have with that character is a complete, unique conversation because it's using a large language model to have that conversation with you. So games that you're playing will be a unique experience every single time you play the game. And every person who's playing that game will have a different experience inside the game than everybody else because of the way large language models will be baked in and the conversations will be different. I also think things like AI generated game art is going to speed up the process of, of making games for a lot of companies and the companies that are really good at making games, really good at creating that addictive game loop and creating really Good storylines are going to leverage AI art to produce games even faster that are still good games. Saying that I also think we're going to see a ton of complete crap games now that people have access to tools like GPT4 to write the code for them. Tools like Midjourney Stable Diffusion Dolly to create imagery for them. Tools like in world AI and convey AI to create the dialog for NPCs. Because everybody has access to all of this, it really democratizes AI game creation, which means a lot of people that probably have no business creating games will create games and flood the Steam store and the Epic Games Store and everybody's going to think they're a game designer. Now, I have no problem with people going and spending a weekend developing a game and having some fun with it. But to turn around and try to sell that game and make money off of something that was kind of slapped together is something that I'm not really looking forward to, but I know is inevitable. 2023 saw the rise of AI influencers. I think we're only going to see it more in 2024 as the AI art image generators get better and better. The ability to make more realistic looking AI generated videos gets better. The fact that a lot of these AI art tools can actually create consistent characters now, even though they couldn't a few months ago, that makes it easier and easier for people to create these AI influencers and then turn around and try to generate income from it using things like partnership and sponsorship deals on Instagram, creating only Fans accounts, creating Patreon accounts, things like that. We're going to see a lot more of that in 2024 as it becomes easier and easier for anybody to make them. Case in point, this image that I was just showing on the screen here, this is an AI model named a TANA who earns $11,000 a month. And this isn't even a real person. Yes, there's a real person behind it generating the images, generating the actual content, but the images are all AI generated. You can see here that because of the ability to create consistent character models, we can get the same person in different outfits and different poses. And the people behind these models are generating a lot of followers and making really, really decent income off of it. And if it's easy and profitable, a lot more people are going to jump on board until it's no longer profitable. 2023 also saw some big advancements in robotics. We saw a handful of humanoid walking robots from companies like Tesla. And I think 2024 is going to be A big year in robotics. I'm planning on being at CES next week in Las Vegas and I'm already getting a whole bunch of details about various robotics companies that are going to be there. For example, this two legged robot developed by LG that is designed to hang out at your house and watch after your pets for you when you travel. And this suite of robots from Leobotics, they're all going to be presented and showed off at CES later this month in Vegas. 2024 is also the year where I think self driving vehicles are are going to get really, really good. We're already seeing companies like wemo and Cruise that have fleets of self driving cars basically ubering people around. You just request a car, it comes and picks you up and takes you to where your destination is without a driver, even in the front seat. Tesla in 2023 demoed its full self driving capability. It did try to run a red light and had a few mistakes, but it is really, really dang close. So I think 2024 is the year where self driving cars are also really, really finely dialed in. And while we're on the topic of robotics, I also think robotics are going to start taking over more and more jobs. Mainly the more dangerous jobs, the lower paying jobs, the lower skill jobs. I think are going to be some of the first to go. I think the priority should be focusing on letting robots handle the really, really dangerous jobs where humans put their lives at stake first. Followed by the very low income, low wage, low skill jobs. Those are the jobs. I think they're going to be replaced by Robotics first and I think we're going to start to see more and more of that in 2024, especially as companies like McDonald's and Wendy's are planning on having AI take your orders at the drive through. And we're getting more and more automated self checkout processes in stores and things like that. I think we're only going to see an acceleration of that trend in 2024. I think the open source versus closed source large language model debate is going to rage on in 2024. On one hand you have the closed source models from companies like OpenAI and Anthropic and Inflection, AI and Gemini from Google. And these tend to be a little bit safer models because the closed source companies could put more guardrails on them and sort of disallow people from using them for unethical or even illegal methods. However, because this is controlled by these large corporations, these corporations can work their own biases into them they become the gatekeeper of what they feel you should be able to know and not know out of these large language models. And it also raises data and privacy concerns. As you use these models, what information are they collecting, what conversations are they storing and training for future model uses? All of this tends to be pretty unclear with a lot of these closed models. Then on the other hand, you have these open source models which allow anyone, and with pretty much any intention to build on top of these models, to iterate to make them do whatever they want, meaning there will probably be some bad actors and people using them for highly illegal activities. However, it does reduce the bias, it does reduce the data and privacy concerns, and it does allow any individual to use a large language model however they want to use it. However, at the moment, compute power is a bottleneck and really only the biggest companies have the compute power to train some of these models. Saying that I think compute costs are going to come down, I think the compute power that it requires to train these models is also going to decrease and it's only going to become easier and easier to develop these open source models over time. But the question will still continue, should people be allowed to develop these knowing that they can develop them to accomplish illegal, immoral, unethical things? Personally, I'm a proponent of open source. I do think people should be able to develop on these and iterate on them and build off of them. But it's also the rabbit hole that sort of leads to some very scary thoughts about where AI might end up in the future as well. And while we're on the topic of open source, the company that's probably making the most advancements in open source large language models, Mistral AI said earlier this month that Mistral will be releasing an open source GPT4 level model in 2024. So if we're making predictions, I might as well predict that Mistral is going to follow through on this and we're going to have an open source model that's as good as GPT4 at some point in 2024. 2024 is also an election year here in the US and I think it's going to be a total mess. I think we're going to have a lot of AI generated fake images of candidates doing things that they didn't actually do. We're going to have a lot of fake voice clones and deep fake videos of candidates saying things that they didn't actually say, and people are going to struggle to tell what is real and what is not. During the elections in fact, the two images that you see on the screen right now of Donald Trump and Joe Biden were both great with AI. I made them in mid journey right before hitting record on this video and it took me all of 25 seconds. Without any real deep analysis on these images and staring at them for a long time, I doubt most people are going to be able to tell that these were AI generated. And Google knows this is going to happen too. According to this article published earlier this month, Google is making a preemptive move for the 2024 election. Google's going to limit election related queries on its generative AI tools like Bard and the generative search experience. So basically Google's not going to let you generate images of political candidates in the future to try to get ahead of this. But there's all sorts of other AI art generation tools out there that are quite honestly better than what Google has to offer anyway. I also truly believe that 2024 is going to be the year of XR or Augmented Reality slash Virtual reality. We know the Apple Vision Pro is coming out sometime early 2024. Meta just released their Meta Quest 3. We have Xreal glasses at the Consumer Electronics show this month. We know there's a whole bunch of augmented reality and virtual reality companies planning to present. I really, really think we're going to see this as the year things take off. I think what's going to happen is the Apple Vision Pro is going to come out. It's going to be popular among tech enthusiasts and YouTubers and people willing to spend the money on something that's essentially a $3,500 VR headset. However, people will see what the Vision Pro can do and then look at the Meta Quest and go, well, it doesn't do a whole lot more than the Meta Quest and I can buy one of those for 400 bucks right now. Or they're going to look at the Xreal glasses that can show you like an 800 foot screen in front of you while wearing these sunglasses, and you can play your PlayStation 5 on the world's biggest monitor through these headsets. And people are going to see the Vision Pro. They're going to see what people are doing with it, and then they're going to buy these lesser expensive alternatives to get into the AR XR game. And the form factor on a lot of these augmented reality glasses are getting so much better. The xreal glasses just kind of look like Ray Bans that you'd wear anyway. Yes, they still have a cord on them, but give it a couple years and pretty soon they'll probably even be cordless themselves and you'll be able to walk around on the streets and nobody will know that you're wearing augmented reality glasses. So I think 2024 is the year we really start to see that accelerate. While we're on the topic of Apple, I think Apple is going to enter the game in a big way. They're going to have their new Ferret model. I think we'll likely see a version of Siri sometime in 2024 that competes with GPT3 or maybe even GPT4. And I think the Apple Vision Pro is going to have a whole bunch of AI features built into it to be able to experience generative AI directly inside of your headset. So I think that's coming in 2024, soon after the Apple Vision Pro launches. And then the last thing I want to talk about is we are going to see a lot more lawsuits. These lawsuits aren't going to slow down. Tools like ChatGPT present a threat to news companies because people can just go and learn about the news using generative AI. And instead of reading a lot of these news websites, people can go and generate new images for their blog posts and their videos instead of using stock photo websites. Because AI presents a risk to a lot of existing industries. We're only going to see these lawsuits ramp up and happen more often as these companies do whatever they can to protect their initial business model. I think over time AI is going to win out. It's inevitable. There's too much open source stuff happening. Even if some of the closed source companies like OpenAI and Google can no longer share some of this copyrighted information, people will still figure out how to train it into the open source models. And these advancements are going to happen anyway. So I think overall copyright needs to be rethought. I think the concept of copyright is antiquated. It's an old model for old systems. And I think in the world of AI it needs to be rethought. And I think the courts are going to be rethinking and refiguring how Copyright works in 2024. Whether they get it right, I have my doubts. But I think they're going to be litigating and trying to figure it out this year and hopefully we get a complete overhaul over how copyright works. But at the same time, I also hope that a lot of these sort of legacy media companies and the traditional way that people find news, find a new model and find a better way of doing it, I don't want to see the people that are doing investigative journalism, the people that are doing the research and creating all the content, be completely de incentivized to continue to do the research and to continue to want to create this content. I think it's still necessary to have humans in the loop for that. So I hope compromises can be made and some sort of new solutions to copyright, that's a win win all around are come to. That's my true hope for how it plays out. I am not somebody that says screw copyright. If you create content, anybody should be able to use it. No, I'm a content creator and I want to be incentivized to continue to create content. And I feel that same way about other news and media outlets. So I think a compromise, a way that works for everybody is still necessary. And I hope 2024 comes to a conclusion that makes sense for everybody. And there you have it. Those are my predictions for 2024 where I think it's going to go. Again, this is just based on going deep down the rabbit hole. There are a lot of companies that are developing things that I'm not privy to. There is probably a lot of innovation that I can completely left out of this. And there's probably a handful of things that I said will happen in 2024 that I'm way off. They're several years off. But again, the timelines are where I tend to struggle the most. I can see what's coming, I just don't know how far away it is because I usually guess wrong. But I thought it would be fun to make this video at the very beginning of 2024, talk about my various predictions of where I think things are going and then either in January of 2025 or in December of this year, look back at this video and see how close we got. Who knows? It's all just fun. It's all just speculation. I just love nerding out, reading the articles, seeing what people developing and trying to guess where it's all headed. And if you like that kind of stuff, make sure you subscribe to this channel and like this video and I'll make sure more videos like this show up in your feed. Mostly I talk about AI news, I share AI tutorials, I talk about upcoming AI research, and I try to simplify and break down complex concepts in the AI world. If that's what you're into, I definitely think you'll benefit from subscribing to this channel and liking videos like this. Also, if you do want to Stay in the loop on the latest AI tools, the latest AI news make sure you check out FutureTools IO. I curate all the cool AI tools that I come across. I share all the latest AI news as it happens on the news page. And I have a free newsletter where I'll send you just a handful of tools and just the most important AI news for the week. It'll come directly to your inbox in two emails a week. One's got the tool, one's got the news. You could subscribe for free over at FutureTools. I oh, thank you once again for tuning in. 2024 is going to be an exciting year. I'm going to do my best to keep my finger on the pulse, report the news, share the tutorials, show you what's coming, do everything that I did in 2023, but hopefully do it at a higher level and even higher quality, more entertaining and just improve. I just want to continue to improve on this channel, improve what I'm sharing with you, and just be as helpful as possible while keeping my finger on the pulse of the stuff that I nerd out about. So thanks for being a part of that. Thanks for tuning into these videos. I really, really appreciate you. I'll see you in the next one. Bye.










