Sam Altman New statement On GPT-5 Is Surprising!

The most important thing, and this sounds like the most boring, obvious, trite thing I can say, but I think it's actually much deeper than it sounds. The most important thing is that the models are just going to get smarter, generally across the board. There'll be a lot of other things too, which we can talk about. But if you think about what happened from GPT 3 to 3.5 to 4, it just got smarter and you could use it for all these things. It got a little more robust.

So Sam Altman recently gave a rather fascinating talk, and it wasn't that long, but it was really interesting. And in this video, I think there two key statements that you guys need to pay attention to because there are a few slides that really do show how crazy things are. So one of the slides you do get to see is this clip. And I'm going to show you guys this clip right here. So basically what this guy is talking about is he's talking about the compute in terms of the future models. And the image shows a graph which looks like it's a giant, giant increase in terms of the capabilities. And from what we've seen previously, this looks like it's going to be pretty insane with regards to the kinds of, you know, improvements that we're going to see. So what I'm going to be showing you guys now is, of course, the clip of the guy basically stating that, you know, we are pretty much nowhere near a plateau, which is definitely, definitely not surprising due to the major things that I've discussed previously. But after that, I'm going to show you guys a clip where Sam Altman actually speaks about what's coming next with GPT5, because this was at the Microsoft conference build software.

But we are riding like a fundamental wave in, in the development of this AI platform where if you just sort of look at compute over time, like how much, how much GPU cycles or accelerator cycles that we're using to train the very biggest models in the world since about 2012, like that Rate of increase in compute, when applied to training has been increasing exponentially. And we are nowhere near the point of diminishing marginal returns on how powerful we can make AI models as we increase the scale of compute.

So, yeah, you can see that this is where, you know, he literally said that there is, you know, nowhere near any kind of, you know, slowdown in terms of what we're going to be able to achieve. And of course, one of the crazy things that we can see is we can also see this image here. And he literally states that the next sample is going to be pretty incredible. So we can see that, presumably, although there aren't any numbers on the chart, we can see that GPT5, according to this chart. In fact, it probably should ask, like an AI, how big, much bigger this looks. The only problem with this image is that we don't know what kind of chart this is. Like, we don't know if this is, like an exponentially increasing chart. We don't know if it's a log chart. But, yeah, so I can't say if it's going to be three times, two times. Like, we really don't know. But what we can see is that the chart does look like it's going up in that, you know, exponential way. So this is definitely pretty, pretty crazy that this was, you know, even something that was talked about. So I think that this is something that's pretty, pretty cool. And I think it's, you know, a key indicator on where we're going to go next. So now I'm going to show you guys the clip where Sam Altman actually talks about how the future models are going to be and what they're going to be like.

And obviously, great to see you, but developers have been such a core part of what's been happening this last year and a half. There's millions of people building on the platform. What people are doing is totally amazing. And the speed of adoption and talent and figuring out what to build. With all of this over, what has really not been very long, like, when we put GPT3 out in the API, some people thought it was cool, but it was narrow, where the utility happened very early. And seeing what people have done with GPT4 and seeing now what's happening with GPT4O, even though it's new and hasn't been out that long, is quite remarkable. I've never seen a technology get adopted so quickly in such a meaningful way. What people are building, how people are finding out how to do things that we never even thought of possible. Which is why it's always great to have an API that's been very cool to see.

Yeah.

And I think, you know, what you just said is, like one of the most important points to me. Like, there's a version of AI that could have existed that is, you know, like a bunch of smart people, like, building, you know, things at extraordinary scale and then just building it into a bunch of products where everybody gets to passively use them. Like, the really brilliant thing that you all have done is, like, taken the exact same set of Things and like decided to make it available to like any developer who's able to sign up for an API key.

Yeah, we try to be really thoughtful about what makes a good API for this. There's going to be all kinds of ways people can use this, but the more this can just be a layer that gets built into every product, every service, the better. And we've tried to make it such that if you want to add intelligence to whatever you're doing, any product, any service, we make that very easy.

Yeah. And again, I think the progress has been stunning. So you know, I think, you know, the setup for like introducing you onto the stage here was I saw that big blue whale. Yeah, like, you know, you're making good use of the whale sized computer right now. And so like without like getting too specific, which we can't be obviously like what are the category of things that people should be expecting over the next, you know, K months?

So the most important thing, and this sounds like the most boring, obvious, trite thing I can say, but I think it's actually much deeper than it sounds. The most important thing is that the models are just going to get smarter generally across the board. There'll be a lot of other things too, which we can talk about. But if you think about what happened from GPT 3 to 3.5 to 4, it just got smarter and you could use it for all these things. It got a little more robust, it got much safer. Both because the model got smarter and we put much more work into building the safety tools around got more useful. But the underlying capability, this amazing emergent property of like, we actually are seeming to increase the general capability of the model across the board, that's going to keep happening. And the jump that we have seen in the utility that a model can deliver with each of those half step jumps in smartness, it's quite significant each time. So as we think about the next model and the next one and the incredible things that developers are going to build with that, I think that's the most important thing to keep in mind. Also speed and cost really matter to us. So with GPT4.0, we were able to bring the price down by half and double the speed. New modalities really matter. Voice mode has been actually a genuine surprise for me and how much I like the new voice mode and when people start integrating that, I think that'll matter. But it's the overall intelligence that'll be coming that I think matters the most. So you, for a while now have.

Been one of the most successful Startup investors in the world. World. And like now you are one of the most successful CEOs of one of the most important companies in the world. So you've got a room full of developers here. You know, like, I think there are 5,000 people in the room and there are about 200,000 people online right now. What's your advice to them? Is like, they think about how to spend their precious time given what's happening in the world. Like, what? What's your advice?

Two things. Number one, this is probably the most exciting time to be doing, building a product, doing a startup, whatever it is that we have seen, at least since the mobile boom and probably, I would say since the Internet and maybe even bigger than that, we don't know yet. But the big opportunities, the big, you know, the ability to sort of build something new and really kind of like change the landscape that comes at the platform shift times. And we haven't had a platform shift in a while. And this looks like it's really, truly a platform shift. And so my biggest piece of advice is like, this is a special time and take advantage of it. This is not the time to delay what you were planning to do or wait for the next thing. This is a special moment and a few years where a lot of stuff is going to happen and a lot of great new things are going to get going. The second thing also about platform shifts is when the mobile phone revolution started or really got going, like 2008, 2009, you would see people say, we're a mobile company. You know, we have a mobile app. And then only a few years later, no one said they were a mobile company because it was like table stakes, an amazing new technology, which I would bias, but we'll put AI in that category. It doesn't get you out of the hard work of building a great product or a great company or a great service. You still have to do it. AI alone is a new enabler, but it does not automatically break the rules of business. And so you can use this as like a new thing to do, but you still have to figure out how you're going to build enduring value in whatever you're doing. And it's easy to lose sight of that in the excitement of the gold rush.

Yeah. So one last thing before we let you go. So you and I and like members of your team and members of the Microsoft Team, have been doing really an extraordinary volume of work over the past year and a half, two years thinking about safe deployment of an awful lot of AI capability, like everything from, you know, APIs and developer tools to end products. And, you know, I think we, you know, have accumulated a really interesting volume of experience, like, experience that's sort of hard to get if you're not doing deployments at this scale. So I, you know, and I think you just mentioned something that's like, really, really interesting. Like, part of, you know, part of the interesting and surprising progression of capabilities of these models means that they're more useful in, like, helping to, like, make AI systems safer. So I don't know whether you had some thoughts you wanted to share there as well.

You know, when we first developed this technology, we spent a lot of time talking about, all right, we've made this thing, it's cool. Are we ever going to be able to get it to an acceptable level of robustness and safety? And now we kind of take that for granted with GPT4. You know, if you use it, it's far from perfect. We have more work to do. But it is generally considered robust enough and safe enough for a wide variety of uses. And that took an enormous amount of work across both teams and fundamental research, like, when we started this, we're like, we've got this thing, we've got this language model. It looks like kind of impressive and kind of not. And even then, how are we going to, like, get it aligned and what it, what does it mean? You know, what is it going to take to be able to deploy it? The number of different teams we've had to build up to go from research and creation of the model to safety systems to figuring out policy to how we do the monitoring, that's a huge amount of work. But it's necessary to be able to deploy these and use them like, you know, when you take a medicine, you want to know it's going to be safe. When you use an AI model, you want to know it's going to be robust and behave the way you want. And have been super proud of the work the teams have done, done together. And I think it's amazing how fast this much work has happened and that we can all now use this and say, oh, yeah, it basically, it basically works. As the models get more powerful, there will be many new things we have to figure out as we move towards AGI, the level of complexity. And I think the new research that it'll take will increase. I'm sure we'll do that together. But we view this as a gate on being able to put these things out in the world, which we really want to do.

Yeah, it's definitely table six. So thank you so much. Thank you for being with us here today. Like, I really appreciate your time. It's awesome to hear from you.

So with that being said, what do you guys think about the future of these models? One of the clear thing that Sam Altman has constantly reiterated is that these models are going to consistently get smarter. And, you know, with the recent statement that, you know, we're nowhere near the kind of, you know, cutoff point to where these models are going to be getting their improvements, I mean, right now I think the hype is definitely real and, and I am really, really excited to see what some of the most notable AI critics do state if these models do have a massive jumping capabilities. Because some people are stating that this isn't going to happen at all. For example, one notable AI critic, Gary Marcus, has said, dear Microsoft CTO Kevin Scott want to put money on what you just said, $100,000. And he's clearly betting that these models just won't get any better at all. So I think this is going to be super interesting to see in the future. If he was someone that was right or was someone that was just literally, you know, I guess you could say being a AI doomer in terms of, you know, trying to state that these AI abilities just won't get any better. I'm someone that believes that they will get better just based on the, you know, the things that we've fundamentally seen. And just because OpenAI may have not published any research. Okay. About what they're getting up to, I think it's clear that, you know, the fact that these other companies that have had such a long time to catch up, in fact not even that long, and they've managed to catch up to, just shows us that in the next cycles, OpenAI are clearly, clearly, clearly going to come up with something incredible. But I do think that this Microsoft build has released some very, very interesting products and some very, very new, interesting things that we're all going to be using in the next future, especially with the new computers. If you didn't see that, don't forget to click the link down in the description.