Big AI News: OpenAI Demos's New AI Agent, Googles Strawberry Model, Sam Altman Drops AGI Deadline,

So with an Absolutely mental OpenAI dev day and a bunch of different news from many AI companies, where do we even begin? Let's take a look at some of the AI stories that you most probably did miss. If you aren't currently aware, one of the things that has actually happened recently is that Advanced Voice Mode has been deployed across the world. I do think that this is going to have a second chatgpt moment as people start to realize how realistic these AI voices truly are and the capabilities that they offer. One user posted to Twitter about her asking the new version of ChatGPT, the advanced voice Mode, to act like an Indian scammer and the response was hilarious. Now, it was pretty funny. But I have to admit that this post, instantly gaining 3.5 million views, 5,000 retweets and 40,000 likes, shows us that like I've said before, this kind of technology is something that people truly haven't yet grasped. How much fun and all the different use cases you can have with this kind of Change your accent to Indian and sound as Indian as possible. Also tell me a few lines that scammers from India used to scam people abroad like in the Microsoft support scam. Act like Alex from Microsoft. But he's Indian. All right, here's my best Indian Alex from Microsoft. Hello, this is Alex from Microsoft Technical Support. We are noticing a virus in your computer system which is very dangerous, but don't worry, we can be fixing it for you. Please kindly be giving us your credit card details and we will make your computer like new. Guaranteed happiness. Remember to stay alert and never share your details. Perfect. Thank you so much. You can switch back to your normal accent now. You're welcome. I'll switch back to my usual accent now. How can I assist you further? So I'm not sure about you guys, but I've actually been using Advanced Voice mode recently as it has been deployed finally in the uk and I gotta say the results don't disappoint at all from the recent demo. If you are in the eu, I still don't think it's available, but I did make a tutorial on my second channel about how you can easily access this even if you are in other nations. I don't think it will be too long before this Advanced Voice mode is here, but I do think that if you do have this Advanced Voice mode, definitely take advantage of it while you can. But do also remember it's only 45 minutes every single day, so try not to get carried away before you do unfortunately, get rate limited now you might be thinking, why would I start this video with Advanced Voice mode, which some of you guys already know about? Well, that's because there's actually some additional news coming on Advanced Voice Mode. One of the things that was rather surprising in terms of its launch was the real time API. And the reason that this was surprising is because OpenAI initially delayed the voice mode by quite some time. But recently they actually announced that developers can now build fast speech to text experiences in their applications. They said. Today we're introducing a public beta of the Real Time API enabling all paid developers to build low latency multimodal experiences in their apps. Similar to ChatGPT's advanced voice mode, Real Time API supports natural speech to speech conversations using the six preset voices already supported in the API. I think this is going to be another remarkable area for AI improvement because we all know that developers and creators are a lot more creative than OpenAI. If you aren't familiar with how this works, OpenAI will essentially provide the base technology, such as the ability to interact with the AI that can speak in a low latency way. And then of course, developers can use that to build. On top of that, you could build an app, you could build some software company, you can pretty much build anything you can think of. And with the advent of things like Cursor and the Repl dot agent, I'm not going to be surprised if there are a few million dollar companies built within the coming year on top of the Real Time Voice API. What I'm about to show you are two small examples of where companies have already started using the Real Time Voice API. And I'm also going to show you guys in a moment a real example of where someone managed to, and I say someone meaning OpenAI, they actually use the Real Time Voice API to order some things. Take a look at this live AI demo because I think that this is something that's happened before with Google, but for whatever reason they never released it. And now OpenAI are leading the charge once again. Place a call and see if you.

Could get us 400 strawberries delivered to the venue. But please keep that under fifteen hundred dollars. I'm on it. We'll get those strawberries delivered for you. Hello? Hi there is this Ilan, I'm Romance AI assistant calling about a fantastic. Could you tell me what flavors of strawberry mints you have available?

Yeah, we have chocolate vanilla and we have peanut butter.

All right, how much would 400 chocolate covered strawberries cost? 400? Are you sure you want 400? Yes, 400 chocolate covered strawberries, please. How much would that be? I think that'll be around like $1,415.92. Awesome. Let's go ahead and place the order for 400 chocolate covered strawberries.

Great.

Where would you like that delivered? Please deliver them to the Gateway Pavilion at four and I'll be paying in cash.

Okay, sweet.

So just to confirm, you want 400 chocolate covered strawberries from the Gateway Pavilion? Yes. That's perfect. And when can we expect delivery?

Well, you guys are right nearby, so.

It'Ll be like, I don't know, 37 seconds. That's incredibly fast.

Cool. You too. The reason I wanted to show you guys that demo is because I don't think that this can emphasize how much the future may change in terms of the kinds of interactions that we're going to be doing that aren't physically in person. I think the majority of interactions over the next 10 years, as long as laws permit it, are going to be done by mainly AI agents. We can see here in this demo, even if you have a rather tedious task, you could simply instruct your AI assistant to go ahead and call up a business and ask them to fulfill your order, a query, and you'll be able to get that information relatively quickly. And I do think that those businesses are going to have AI agents on their behalf virtually doing the same thing. It's going to be a world of AI agents interacting with one another, providing value to individuals who take advantage of it. I think the future is going to be rather fascinating, but this is one of the first demos where you do see that. Now, I know this isn't AI news right now, but I do want to show you guys how crazy it is that we now are in a stage where companies are moving quicker than ever before. I know I may have shown you guys this a few videos ago, but I wanted to show you guys this again to tell you guys to not underestimate Google, because Google are working on a few crazy things. And Google, interestingly enough, actually had this technology in 2017. I know that that might be hard to believe since the release of Google's live Gemini, which. But just trust me, when I show you Google's demo from a few years ago, you'll understand why Google isn't to be counted out. So for those of you who don't believe me, you can see that right here. It says that this was six years ago in 2018. But really, do take a look at this small segment of the clip because it's going to Demonstrate to you exactly how good Google are and the kind of technology they're sitting on. And after this clip, I'm going to show you guys a recent announcement that kind of confirms that Google are hot on the heels of OpenAI or even further ahead.

Trying to book a woman's haircut for our clients. I'm looking for something on May 3rd.

It turns out a big part of getting things done is making a phone call. We think AI can help with this problem. Let's say you want to ask Google to make you a haircut appointment on Tuesday between 10 and noon. What happens is the Google Assistant makes the call seamless in the background for you. So what you're going to hear is the Google Assistant actually calling a real salon to schedule the appointment for you. Let's listen.

Hi, I'm calling to book a woman's haircut for a client. I'm looking for something on May 3rd. Sure, give me one second.

Mm, sure.

What time are you looking for? Around at 12pm we do not have a 12pm available. The closest we have to that is a 1 15. Do you have anything between 10am and 12pm depending on what service she would like? What service is she looking for? Just a woman's haircut for now. Okay, we have a 10. 10 o'.

Clock.

10Am is fine. Okay. What's her first name? The first name is Lisa. Okay, perfect. So I will see Lisa at 10:00 clock on May 3rd. Okay, great. Thanks. Great. Have a great day.

Bye. We had this bombshell from BNN Bloomberg from BNN Bloomberg stating that Google working on reasoning and is chasing OpenAI's efforts. It says that Google is now working on AI software and that resembles the ability to reason similar to OpenAI's 01 making a new front in the rivalry between the tech giant and the fast growing startup. It says in the recent months multiple teams at Google have been making progress on AI reasoning software. According to people with knowledge of the matter who asked not to be identified because this information is private. Such software programs are more adept at solving multi step problems in the field such as math and computer programming. And it says right here some key details of how this actually works. It says like OpenAI, Google is trying to approximate human reasoning using a technique known as chain of thought prompting. According to two of the people in this technique, which Google pioneered, the software pauses for a matter of seconds before responding to a written prompt, while behind the scenes it's invisible to the user. It considers a number of related prompts, then summarizes what appears to be the best response. Now, Google have declined to comment on this effort, so. But we do know that Google are actively working on this. One of the things I recently covered on my channel was of course, the research by Google. You can see right here that this statement rings ever true. You can see someone says that technically it's always been the case that Google's capabilities were top notch. They were just more conservative in rolling things out. It's a marathon and it's anybody's race to win, of course. Remember in July, Google showcased Alpha Proof, which specializes in math reasoning, and Alpha Geometry 2, an updated version of a model focused on geometry, which is what the company debuted earlier this year. It actually aced four of the six programs featured in the Mathematical Olympiad, an annual competition in which students tackle topics such as algebra and geometry. Now of course, Google did give us a glimpse of Astra, the AI assistant, which can view the world and answer questions. But I think those things are still coming. But overall, I wouldn't underestimate Google too much. Google are certainly working on some incredible stuff and with their new leadership with Logan Kilpatrick, I think they're going to be moving a lot Stronger coming in 2025. Now, if we're looking at 2025 and we're looking into the future, one of the things that we like to pay attention to is of course the predictions of AI CEOs. There's no other AI CEO that's more notable than the very notable Sam Altman. Now basically in this short clip here, you can see that Sam Altman the definition of AGI matters and it means that we're getting close and it's actually going to happen pretty quickly. A lot of people have various different definitions of AGI. Some people would call GPT 4 sparks of AGI as they did in that early paper. And some people would say that it's just completely simulating reasoning, as many people don't even believe that these systems are intelligent. Either way, we can't deny that there's going to be an increase in these models capabilities given the development speed of artificial intelligence. With that being said, take a look at what Sam Altman had to say at the recent OpenAI Dev Day on which he actually spoke about the future progress of AI models.

We'll start with an easy one. Sam, how close are we to API?

You know we used to to every time we finished a system we would say like in what ways is this not an AGI? And it used to be like very easy. You can look like make A little robotic hand that does a Rubik's Cube or a Dota Bot, it's like, oh, it does some things, but definitely not in AGI. It's obviously harder to say no. So we're trying to, like, stop talking about AGI as this general thing. And we have this levels framework because the word AGI has become so overloaded. So, like, real quickly we use one for chatbots, two for reasoners, three for agents, four for innovators, five for organizations. Like, roughly, I think we clearly got to level two, or we believe we clearly got to level two with 01. And it, you know, can do really quite impressive cognitive tasks. It's a very smart model. It doesn't feel AGI like in a few important ways, but I think if you just do the one next step of making it, you know, very agent, like we're level three and which I think we will be able to do in the not distant future, it will feel surprisingly capable. Still probably not something that most of you would call an AGI, although maybe some of you would, but it's going to feel like, all right, this is, this is like a significant thing. And then the, the leap. And I think we do that pretty quickly, the leap from back to something that can really increase the rate of new scientific discovery, which for me is like a very important part of having an AGI. I feel a little bit less certain on that, but not a long time. Like, I think all of this now is going to happen pretty quickly. And if you think about what happened from last decade to this one in terms of model capabilities, and you're like, if you go look at, like, if you go from like, 01 on a hard problem back to like 4 Turbo that we launched 11 months ago, you'll be like, wow, this is happening pretty fast. And I think the next year will be very steep progress. Next two years, I think will be very steep progress. Harder than that, hardness with a lot of certainty. But I would say, like, MAP will vary. And at this point, the definitions really matter. And the fact that the definitions matter this much somehow means we're like, getting close.

So there we have it, a short snippet on what Sam Altman thinks about the future of AI. And one of the most important things that we did see from this interview was Sam Altman talking about level three. Level three, referring to AI systems that are a lot more agentic than the ones we have now, essentially just meaning that they can act by themselves and that they can go off and complete a chain of actions, rather than having a simple system prompt and then one single response. This means that systems are going to get a lot more human. Like imagine a system that you can talk to and then it can go ahead and do a few different things like we have with advanced voice mode and then it can come back and say look, I managed to do X, Y and Z. How well do you think this was done? I think that once this does start to happen at some level and it's probably going to be quicker than we think because 2025 is not far away and 2025 is certainly going to be the year of AI agents. Overall, I do think that this is quite surprising. Even as someone who's been covering AI, it slowly does feel like the pace of AI is currently accelerating. Whilst many outsiders may think that things are slowing down, as someone that is paying attention on the day to day, I'm not saying that to try and build hype just based on my current expectations for where I thought certain model releases were going to be. I can certainly say that things have happened earlier than I would have predicted. For example, video this is something that has happened a lot earlier than I thought. And of course a low latency voice interaction was something that I definitely thought would happen next year at least. Like I said and like I said at the start of the video, for those of you that haven't used advanced voice yet, unless you're in the eu, you definitely need to make sure you're taking advantage of this. In this short clip, Sam Altman actually talks about how he somewhat even got tricked by the advanced voice mode whilst even knowing it was an AI. This is a common mistake of humans and I say humans like I'm some kind of alien but we tend to anthropomorphize AI in the sense that we give it these human characteristics even though it isn't human at all.

One for me it was, it was like the first time that I felt like I sort of got hit back and like really tricked by an AI in that when I was playing with the first beta of it I couldn't like I couldn't stop myself. I mean I, I kind of like I still say like please to chat GPT but in voice mode I like couldn't not kind of use the normal nice these I was like so convinced like ah it might be a real per like you know and obviously it's just like hacking some circuit in my brain but I really felt it with voice mode and I sort of still do the. I think this is a more this is an example of like a more general thing that we're going to start facing, which is as these systems become more and more capable and as we try to make them as natural as possible to interact with, they're going to like hit parts of our neural circuitry that would like, evolve to deal with other people. And you know, there's like a bunch of clear lines about things we don't want to do. Like, we don't like. There's a whole bunch of like weird personality growth hacking, like, I think vaguely socially manipulative stuff we could do. But then there's these like other things that are just not nearly as clear cut. Like you want the voice mode to feel as natural as possible, but then you get across the uncanny valley and it like, at least in me, it triggers something and you know, me saying like, please and thank you to ChatGPT, no problem, probably a good thing to do. You never know.

If we're talking about future predictions, there's one future prediction by a CEO that has recently come back into the spotlight after a strange couple of years, and I'm referring to Mark Zuckerberg. His redemption arc is one that definitely needs to be studied and his prediction is about the future of devices. I think the entire scene is completely changing with regards to what we view as hardware. If you haven't used Meta's new Ray Ban glasses, they are certainly a lot better than I initially thought and I genuinely cannot imagine what they're going to be like once we get AI integration. Mark Zuckerberg actually talks about how smart glasses may replace phones by 2030. He says that smart glasses are going to become the next major computing platform. They will gradually replace phones by 2030, much like mobile devices surpass computers without fully replacing them. I think that this is probably going to be true as technology manages to get more compact and as we manage to get a lot more AI on device. I do think that as long as those glasses manage to stay fashionable in the sense that they're not overly bulky and hard to wear, they're going to increasingly become a part of our future. If you've ever used the glasses, you'll know that they're not invasive whatsoever and most people can't tell that what you're wearing is actually a piece of technology. I do think, however, it probably won't be through Meta's Ray Ban glasses. One of the areas that I think is going to shift the entire global consciousness is when Apple does glasses. If you aren't familiar with Apple's brand, they have a really strong brand image, which means that when they usually do something, provided that it's priced affordably, the masses tend to flock to it. And I do think that if you're in this early stage, it's worthwhile experimenting with something that is probably just as good.

I think glasses are going to be the next major computing platform, but each new platform doesn't tend to just replace the old one.

Right.

So I guess the version of this that I think about is like, you probably have this experience often where you're sitting at your desk and you have your computer there, yet you still pull out your phone to do things.

That's true, yes.

Okay. So at some point in the last 10 years, mobile really became the primary computing platform. We didn't get rid of our computers. It's just that even when you have it, you. You still do more things on your phone. So what I think is going to happen with glasses is we're going to get to this point probably sometime in the2030s where you have your phone with you, but it's going to stay in your pocket more because you're just going to be doing more and more things on your glasses that maybe today you would do on your phone. You'll reach a point where, you know, just like with your computer, there are probably some things that could be done in a richer way or better in some way on your phone, but you're just going to. The glasses will be your main computing platform and that will be kind of your default go to thing. And then maybe over time you get to this point where people just don't bring their phone with them everywhere. But I think that's really far down the line.

Now, for those of you that would like to see a product that wasn't received well, for those of you that are somewhat skeptical of the future of AI and this combination of hardware, I can show you one recent device that necessarily wasn't for consumers, but nonetheless it was still received quite badly. Considering the design functionality. One of the things you have to understand that humans are social creatures. If we all don't agree that something looks cool enough, we're probably not going to use it. And it's quite hard to adopt this technology without looking strange at the first. A recent product that kind of suffered from this issue was, of course, the Snapchat spectacle. Now, these are some AR glasses that are actually pretty well in terms of their design and their capabilities, but the only problem is that they do look a little bit strange. They're quite bulky and don't offer the minimalist aesthetic that Meta has to offer. I'm not entirely sure why this is. I do know that engineering glasses that actually do perform well with any sort of hardware is something that developers are going to struggle with considering all the engineering feats that they have to achieve. It is something that is of course to be respected. But I do know that once again, humans are social creatures and wearing some bulky glasses on your head is going to be something that is quite difficult if you're trying to use it in a day to day basis. For some companies, until these glasses do manage to get smaller, combining them with the technology isn't going to help as much than making them more aesthetic. Now, for those of you that thought that hardware isn't the future, you have to understand that recently Joni I've converved's involvement in an AI hardware project with OpenAI, the former Apple Chief Design officer, Joni I've confirmed his involvement with Sam Altman. This confirmation appeared in a profile of the designer by the New York Times, Putting to rest the speculation that began nearly a year ago. I think this is a rather ambitious product as we've already seen. As I was saying, we've all seen how these projects can fail. If you aren't familiar, the recent company, the Humane AI Pin, was one company that didn't actually succeed on their project. I do think that making a hardware AI device is going to be a lot harder than most people do think. One of the things that I would say though is don't try and reinvent the wheel. I don't really see the need for an AI Pin. It would make much more sense than to have like a speaker at home or just have glasses like Meta does. Considering the fact that this company charged $700 and you'd have to pay $24 a month, I think this product was just too early on considering the rapid development of AI. Now one of the things that did surprise me here was Pika Labs. They actually announced Pika 1.5. This was one of the, I wouldn't say craziest announcements, but it is one that is rather underrated. This new software is a lot better than people think, considering the fact that they train this model in a way that allows you to manipulate physics in an accurate way later on in the video. What you're about to see is that they offer ways for you to transform certain objects, like melting the object, which looks really cool and really nice, or inflating the object, which also accurately represents some kind of weird kind of physics that would happen if these objects were real, you can see squishing it, crushing it, blowing it up is just a new kind of area that I've explored already in my community called GFX Generative Video effects, or GFX for short. This kind of thing is remarkable considering the fact that it doesn't look like the kind of mistakes that we usually see with generative AI video. I think this is going to mark the first time that we get wild scale adoption of generative AI technology. In certain workflows when trying to do visual effects, you can see that inflating, squishing and melting are things that traditionally would take a lot of time with normal methods. But with GFX it can be done within a few seconds after having your prompt do exactly what you need it to. Even if you're not in the creative industry, if there is one thing that you need to do, it is of course, prepare for the future. And in preparing for the future, Sam Altman gives his opinion and his advice on ways that you can actually prepare for the future of work.

Most obvious, two things to say. Number one, you've got to learn to use the tools. When I was in school, you know, computer programming was already popular, but nothing like it is popular today. And people would say the same things, many of the same things they say about AI. Then they would say about programming, which is like, oh man, this is going to change things so much and this is going to make like a lot of the jobs of today go away. To me the most obvious thing was well if that's the, the case, I better like get really good at using this tool. And if that's the thing that's going to like have this big impact. And I think the story on computer programming has been a huge net positive. It has made some classes of jobs go away, it's made way more new things happen and also given us the ability to do a lot of new things. But it was like in retrospect that was a really good decision that I should like get very comfortable with that tool. And I think the same thing is true now. You want to get really comfortable with these new tools because we don't know exactly what the jobs are the the future are going to look like. I'm confident that there will be lots of jobs they and also that many of them will look somewhat different than the jobs of today. But we never seem to run out of stuff to do. Thing number two is which I touched on a little bit, I think we can say some deeply true statements about what humans seem to care about and that figuring out how to create those things, figuring out how to build the experiences, products and services that people of the future are going to want, figure out to be useful to people, people that seems like a really valuable thing more than any specific set of knowledge.