AI News : OpenAI Reveals Plans For Next Year, AGI Timelines Get Shorter, Googles Secret New Tool!

So one of the first stories that we had that was absolutely insane was the fact that we had SORA get leaked to the public. Now, of course, it wasn't the actual tool that got leaked to the public. It was simply an API key that was wrapped in a hugging face chat. So this was what a lot of people saw a few days ago when we were looking at the article, because essentially what happened was a bunch of angry artists, or should I say one artist decided that OpenAI's alpha program or beta program was something that they didn't want. You can see right here. I know some of you guys probably have seen this already, but basically they were like, we received access to the SORA with the promise to be early testers, and somehow they, you know, I don't even get why they were angry, but they said, we are not your free bug testers, PR puppets, training, data validation tokens, and we do all this work for free. They were basically saying, oh, you're exploiting, you know, artists for unpaid research and design and pr. But a lot of people that I know just didn't really understand this because it's like you, you probably got invited to this program and I know tons of artists out there that would have loved to have been selected for this program. So it's like, I'm not sure why the people who were in this program decided to get upset. I mean, they were upset at things like, every output needs to be approved by OpenAI team before sharing. I mean, of course they've got like generative images. Like, if you had a company that was making generative AI music or generative AI images, surely you wouldn't want some terrible images coming out where people could say that your software is bad, especially in the competitive AI space. This thing just didn't make sense to me at all. I don't know why they were angry at this. I mean, I guess I could understand it maybe like 10%, but yeah, think that it makes sense. The actual big news about this wasn't the fact that OpenAI is some evil company or whatever. The big thing about this was the fact that there is actually certain things going on with the SORA model, which is more interesting than this artist drama. So that kind of news was basically the fact that this, this model has different styles. You can see here it has a style called natural. You can see there's also probably some inpainting. You can see it says inpaint items. Most interestingly as well, we can see that the model that we actually got was a turbo model. So it doesn't mean that it's quite likely that when ASORA is released, there will largely be a turbo model and probably a more expensive model that can produce certain videos. And one of the craziest things that we also did get from this was the fact that like, Sora is something that is a lot better than other models that currently exist, which is quite surprising because a lot of people would say that Cling and all those other models are really better. But I think it was Ethan Mollock on Twitter that said, look, I've tested tons and tons of models, but this model is completely better than Sora by a little decent margin. So that was something that was really surprising as well, because a lot of people were like, look, OpenAI is levels behind now. They managed to lose the game, but it's clear that they're still in the lead. And another thing as well, you guys can see right here this video gen part. We can see that this says video gen. So we know that if this is, you know, type video gen, it's quite likely that we might have image generation here. Which means that it's possible that with Sora we also get an image generation model. I don't know if OpenAI will actually release that because a lot of the times OpenAI is such a big company now that every release is under scrutiny. Like another company could do something and it wouldn't face much as scrutiny as OpenAI because they just have such huge brand recognition. Like, if you remember the first time they released Sora, a lot of people were like, why does anybody need a generative AI tool that just rips the artistic creativity out of humanity? And so I'm wondering if they release this image generation tool if they'll face the same kind of backlash. So we'll be interesting to see what happens on that aspect. Now, another thing that most people did miss was the fact that Google actually released their Google experiment labs and it was called Generative Chess. Turn your ideas into playable art pieces using Google's Imagen 3 model. Create and play today. So this was something that, you know, I really do wish that I was a more creative individual at the time of creating this demonstration. But basically what you can do is you can play chess and you can generate the chess pieces with anything that you do want. Now, like I said before, I do wish that I was a little bit more creative because you can see that what I did was I decided to make was I decided to make a generative chess game that was, you know, Minecraft inspired. I Don't know why I thought of Minecraft immediately. I'm guessing because I saw AI generated Minecraft, but I was just thinking about video games and then I typed in Fortnite. I know this is probably like really, really probably the worst example I could have used. And I mean, I could have used medieval, I could have used like ceramic, I could have used wooden. I mean, there's a billion different things. So you can see right here, it's able to generate these, you know, these, these chess pieces here. And I guess it's just an interesting way to use generative AI because we know that Google is, you know, the company that created Notebook lm and they're really, really ramping up their efforts to, you know, use generative AI and wrap it in certain different ways. So for me, I thought this was something that was really cool. I guess from this you could do, I mean, there's a billion different ways maybe you could do race car inspired chess pieces. I have no idea. There's a billion different things that you could do. But Google Labs, I think they have the most interesting AI products. So this is going to be something that I'm always recommending to viewers because if you were in the Google Labs area, I'm pretty sure you would have had access to NotebookLM before a lot of people managed to do that. So it's something that is really interesting considering where we are going to be now. Let's get onto something that was probably one of the biggest announcements this week. So another big announcement and I can't believe I didn't even cover this yet because it's a pretty, pretty big announcement. Not a pretty big announcement. It's a really big announcement was the fact that we actually got Anthropic releasing mcp, which is a framework that allows Claude to run servers, giving it superpowers and effectively turning the Claude app into an API. And they created some server that they think you'll love. So, you know, Claude can read, create and edit files and folders locally, which is pretty, pretty insane. So you can see right here, it's able to create, you know, libraries and structures and it's able to do that for you all locally. So this is going to open up a huge level of areas for usability in terms of what users are able to do with these models. So I'm someone that's slowly, slowly diving into the space because I want to get the most out of what these models are able to do. And one thing that I do know is that every time these features are incrementally released, we get a ton of more use cases that most people aren't currently aware of. So in this example, you can see someone is creating something with this demo right here, and you're going to be able to see them to create all of these files locally. You can see it's running all of them. You can see this is where the user managed to create a flask application locally. So this was really crazy. You can see that the user is able to run this, and you can see they've created a paint application locally. That is absolutely insane. So this is just making stuff a lot more easier in terms on the, you know, software development side, because when we have Claude that's able to talk to a user and is able to guide them step by step and is able to navigate them through all of the areas where you want to be able to do stuff, this is opening up so many possibilities. I know this is literally going to help out so many people because so many times when I was doing tutorials on stuff on how to do stuff locally, a lot of people got confused. And, I mean, it is kind of a little bit weird. I know for some people it's easy, for some people it's hard. But this is going to be something that I'm definitely going to be making some tutorials on so that individuals can get the most out of this. So to be something that is really, really, really effective, and it's definitely something that you guys should explore. But I will be releasing a tutorial in around three or four days that dives into absolutely everything for beginners so that you guys can benefit from this the most. So then we actually got the Lord Knight of Weymouth speaking in the House of Lords and basically talking about artificial superintelligence. Now, after this, I'm going to talk quickly about artificial superintelligence, because there's an ongoing Twitter discussion that I think is rather important. AI could pose an extinction risk to humanity, as recognized by world leaders, by AI scientists and leading AI company CEOs themselves. AI systems capabilities are growing rapidly. Super intelligent AI systems with intellectual capabilities beyond those of humans would present far greater risks than all existing AI systems currently. So I'm going to go off script here just a quick bit, because I just want to discuss this AI safety thing. Now, this, this isn't to create any drama, but I genuinely want to understand where people's mindset is when it comes to artificial superintelligence. And I'd like to know what the broader community thinks about AI safety. So Basically, Chubby tweeted, I reached a point where I think we should just always do the exact opposite of what Jeffrey Hinton tells us. It's a shame. I have so much respect for the work he's done. And of course we've got David Shapiro saying he's a useful idiot, yada yada yada, and people are saying he's focusing on regulatory capture. Now, if you don't know who Geoffrey Hinton in, he's basically the godfather of AI. He's a well respected individual in the field of AI. Now this guy has, you know, recently, I think he got, you know, the Nobel Prize for his work with neural networks. It was along the lines of that. But the point is, is that like a lot of people are stating that he's wrong when he states that open sourcing big models is like being able to buy nuclear weapons at radioshack. And I genuinely agree with this. I'm going to put this question out to you guys. Why is it that we need AI models that are able to do certain things? For example, I said genuine question. Let's say future models are capable of mass scale harm, for example bioweapons and such. Such. Why would we initially open source that? What do we gain as a country or as a society if we open source that? I genuinely cannot find one argument for open sourcing big open models. I'm not saying that the open source ecosystem needs to be terrible. All I'm saying is that when we understand where this technology is going to be in the future, if we're stating that, look, artificial superintelligence is a true thing and we're going to have these beings that are basically going to be able to do anything and, and things that are magic. Why would we open source models that are able to cause catastrophic harm. That's the only question. I'm not saying that we can't have an open source ecosystem. I'm not saying that we shouldn't have open source and transparent models. I'm just saying that the dangerous parts of the models, why would that part need to be open sourced considering the fact that those parts can only be used for adverse things? That's just my question to you guys. Now coming back to this, I actually wanted to talk about the AGI timelines. I really do wish I had a better, higher quality image here, but. But I'm going to show you guys this because something actually crazy happened. So what happened was that you guys can see here that Sam Altman has AGI at 2025, Elon Musk has it as 2026, Dario Amade has it at 2026, Rick Herzl has it at 2029, Jeffrey Hinton 2029 and Demis Hasabis 2030. Now this is important because these guys are key, key individuals when it comes to shaping the AI space. So you could argue that these individuals have a lot more information. Like I said before, a of lot apologize for the image quality, but okay, one of these people isn't here. Now one of the people I'm referring to is Yann Lecun. And the reason I'm talking about Yann Lecun is because he is a notable AI skeptical, okay? And he's very skeptical of the current, I guess you could say, you know, paradigm that we're currently in with LLMs and stuff like that. He basically says that, you know, everyone's running off the wrong paradigm. Now recently, I think maybe they may have had a breakthrough because this guy actually changed his timeline, okay? He actually changed what he initially believes to happen with regards to the timeline. So take a look at this because this is crazy and this is insane because when we actually take a look at individuals like Yann Lecun who are some of the most prominent AI skeptics with regards to the current paradigm, him stating that his timeline is now basically the same as Sam Altman's or any of these industry leaders basically means that we are most certainly going to get advanced AI a lot before we think we do.

So the future is that if we succeed in this plan, which may succeed within the next five or 10 years. Five to 10 years, we'll have systems that as time goes by, we can build up to become as intelligent as humans, perhaps so reach human level intelligence within a decade. That may be optimistic. Five to 10 years would be, if everything goes great, all the plans that we've been making will succeed. We're not going to encounter unexpected obstacles, but that is almost certainly not going to happen.

You don't like that, right?

Like AGI and human level intelligence you think is far, far away or unlikely?

No, I don't think it's that far away. I don't think my opinion about how far it is are very different from what you will hear from Sam Altman or Demi Sabis or things like this. It's quite possibly within a decade, but it's not going to happen next year. It's not going to happen in two years, it's going to take longer. And so you don't want to extrapolate the capabilities of LLM and Say we're just going to scale up LLM, train them with bigger computers or more data and human level intelligence. Intelligence is going to emerge. It's not going to work this way. We're going to have to have those new architectures, those JEPAS systems that learn from the real world and can plan hierarchically.

So yeah, that clip was a really fascinating one because it was interesting to see Yann Lecun finally actually talk about his predictions for AI in terms of his timelines. And we also got altaytts, which is an experimental text to speech model that uses a pure language modeling approach to generate speech without changes. The foundational model itself. Sometimes I think about how fast everything changes. Like one day you're using floppy disk and then the next everything is in the cloud. It's wild, right? I mean, how do we even keep up with all this innovation? So what was the inspiration behind your latest project? Like, was there a specific moment where you were like, yeah, this is it. Or you know, did it just kind of come together naturally over time? So yeah, it's really interesting to see exactly what the open source space is doing and these kind of things that you can run locally are going to change the space in terms of not having to rely on external providers. Now I've got to be honest man, these things that keep happening in the space where you get these models that are really good that you're able to run locally, I'm actually wondering at what point a model is going to become so good that we can literally just run everything off a computer. I mean, recently we've seen things like Llama 3 and smaller models just get more compact, they become more smarter. But at what point are we going to reach that point where, okay, let's say for example ChatGPT and all of those things, we're just going to have a default version that is just completely offline that we can use absolutely anytime we want. I mean, of course there's pros and cons to each. I think still we're going to have one that is able to search the Internet, upload files and be on servers and all that stuff. But I think there's going to be a large use case of people that once these models get good enough to where they're actually really effective on device, we're going to probably have this entire system where you can just simply download it and, and you can just use it completely offline. Now Rufin Van Ron is actually building an implementation of the global workspace theory of consciousness in an AI system and says that machines could be conscious in five years and it could very well happen next year. And this is super interesting because there was actually a company that I recently saw and their entire thing is actually focused on making machines conscious, which is completely crazy because so many people are working on things where they're saying, look, we're going to wake up Claude and all this interesting stuff. So I think AI consciousness is going to be one of the biggest things within the next five years as these models get increasingly, increasingly complex.

The hope is that in five years years we would have a full working implementation of the global space theory in a non trivial situation. So not exactly like what you see here. And maybe five years from now we would be able to assess whether this particular theory and the particular way that we have implemented gives rise to emergent properties and possibly consciousness. Even though we're not actually trying to design it, we're aware that it might happen. So five years is the time scale that I would could give you for this particular project. Now, more generally speaking, if you want to put a timescale on the potential emergence of consciousness in AI systems, I think the range is much wider than this. It could very well happen next year or in a couple years with the next installment of GPT5. I wasn't really kidding. I think there's a real possibility there. Well, some people even think that the current version is already somewhat conscious and it could take 20 years or more. Because it could very well be that all of our theories of consciousness are wrong and that just needs some new idea, right?

Sometime in the next century or so there could be a five year period when the economy goes from doubling roughly every 15 years like it does now, to doubling every month or faster. That's on the table as something that could happen. So that's a kind of change I think is a plausible change. But that would be a worldwide change. That is the entire world economy would start to double roughly every month. So if that's what you mean by the singularity, I say that's not crazy. That's different than one machine in a basement suddenly over a weekend becomes so powerful to take over the world. That's an entirely different kind of singularity scenario. That's the one I don't find very plausible. But I do think it's plausible that sometime in the next century we will hit this transition where the economy speeds up to double roughly every month or so. And AI is very plausibly the kind of thing that would cause a transition like that. That transition would probably go along with most humans losing their job. And we should pre prepare for that. This is the one risk that people have most consistently talked about. With machines getting better for centuries. There's a way we could prepare for that. This is a simple insurance problem and it's even simpler than most insurance. So most insurance, like if you want to get your house insured against a fire, they have to do what's called underwriting. They have to come and look at your house and guess the risk that your house will have a fire compared to somebody else's house. Otherwise they're at, you know, too much risk if they don't bother to check those risks. But here we're looking at an event that would be common worldwide. So I would set up a trigger, something like labor force participation rate, I. E. The percentage of adults who work falls from say above 60% to below 20% in say a 10 year period. If that happens, that would be a sign, a signature of robots took most jobs in a short time. That would be a triggering event. And I would say I would basically just have assets that pay out in that event. A bet, basically.

So this is basically one of the most interesting things I've seen. This is Robin Hansen basically saying that, you know, sometime in the next century, AI is going to cause the economy to speed up so quickly that most humans are going to lose their jobs and should insure against that scenario. So if you're a budding entrepreneur who is thinking about the kinds of businesses that are going to be in the future, that might be the kind of business that you want to build, I think this kind of thing is really smart because you're basically placing a bet on the fact that AI is going to get a lot smarter, which is a pretty sure bet. Now this thing of AI taking a lot of people's jobs is not new news. This is something that we've been talking about on the channel for quite a long time and I do think that this is something that makes sense now. Recently there have been a slew of different tweets that have actually shown us just how crazy the, I guess you could say technology is getting to where individuals who are using the technology in their respective fields have seen that AI crossed certain thresholds. For example, we can see right here, Ethan Mollik says if progress continues, the ability to figure out the AI frontier will slip from most of us. For example, I'm not good enough. I'm not a good enough musician or critical listener to know if Suno V4 is actually as good as it sounds. To me, I need to defer to experts. Is it basically saying that look how like the way how these systems are, we're not going to know whether or not as an average person if this has passed the threshold for a specific level because I'm not an expert in that field. And that basically means that as these systems get better and better, unless the experts are looking at them, we're not going to know how good they are. And interestingly, we had a response from someone called Ben Camp. He said, professional here. I've gotten over 100 million streams, a gold record, 10 years of teaching songwriting at Berklee College of Music, and musically it's better than 80% of my students. But my best students beat it by miles. The industry's best also wins and it's ready to eat service music like advertisement and the library. So right here we can see that across all bounds. It is really changing the game in terms of where this technology is heading. And I can only imagine what it's going to be like in 10 years if it is so good. Now we can also see right here as well, someone posted this. This one garnered a decent amount of attention. Over 500,000 views, 200 likes, 3,000 likes, and over 2,000 bookmarks. And he said, Claude 3.5 sonnet, with about 20 hours of customization work is better than every junior and most mid level media buyers, strategists I have worked with in the five years and I assume it will be better than 80% of senior people. The AI isn't coming for advertising. It's already here. Now I do think that this is a really interesting comment because I've used certain models and over time I've seen that certain models continue and continue to get better when they manage to get trained on newer data. And this is something that I've seen for certain tasks. Like sometimes I'll be writing a piece of information on a website or I'll be filling out the description for one of my videos, or I'll be just doing some general online business stuff. And sometimes Claude will really, really surprise me at how effective it is. Like it won't need so much prompting and it just outputs a response that is genuinely, really, really effective to the point where I wouldn't need to hire someone. And this is something that is slowly, slowly happening. And a lot of the times because what we have is qualitative benchmarks, meaning that it isn't like a math benchmark where you can test it across a range of set problems, but because the kind of problems that this kind of thing solves is something where is, I guess you could say it's subjective in nature. It means that this kind of thing isn't going to stun people once it happens. It's just going to slowly, slowly get better in the sense that it manages to get better over time and people rely less on actual humans for their work. And this is where we also get an interesting prediction from someone who works at OpenAI. So we got this prediction from Jason Wei and he says prediction within the next year, there will be a pretty sharp transition of focusing in AI from general user adoption to, to the ability to accelerate science and engineering. For the past two years it has been about user base and general adoption across the public. Of course, with ChatGPT and with everything else that's been done, this is very natural because user growth is a critical part of any business model. But at this point, I would say that there is a widespread accessibility of LLMs for most queries from the average person on Earth, and many LLMs can answer that pretty well. He's basically taking a look. If you take GPT4, you know, llama, if you take Mistral, if you take all those models and you give them to the average person, they're probably not going to know the difference. But he says that in the upcoming five years, I think that the focal point will be the ability for AI to accelerate engineering and scientific research, which is the engine of progress in technology and at the frontier of innovation in any field. And by definition there will be many open questions and a lot of headroom for better AI to make a difference. And the stakes will be very high because progress compounds and also because AI is accelerating. AI research itself is a strong positive feedback loop. He's basically stating that, look, the other way of saying this is that there is a somewhat limited headroom for improving the average user query, but massive headroom for improving the experience for the 1% of queries that would accelerate technological improvement as well as on queries that people would want to ask the model but currently don't because the model aren't smart enough to answer. And AI research tends to improve. Where there is great headroom and scientific innovation, there could be substantial upside. Basically saying that, look, we've kind of maxed out what these models are able to do for the average person, which means that increasing the strength of the models in areas where the average person gets more out of them won't be something that is going to be the industry's main focus. What they're going to be focusing on is the kind of ability where these models are going to be so smart that they can solve maybe an engineering problem, maybe they can advance scientific research to where, if it actually makes a difference and actually starts accelerating that kind of research, that.