Tim Morris’ Post

Transcript

Hi everyone, welcome to another video where I run through some of the key things in the AI space that I'm paying attention to. I'm also very amused. I'm wearing a postdoc T-shirt here that says everything I know I learned from the Internet, because everything we're covering today I did indeed learn from the Internet. Anyway, getting into it, let's have a look at the first thing I'm paying attention to, and that is pretty much as always, at the evolution of thinking around the application layer of AI. I find this particularly interesting for a couple of reasons. First of all, we are building an application here at agency, so super relevant to us. But also I really find it interesting how the thinking around where the value in the AI stack will come from has evolved. So there's an old saying in a gold rush, you don't want to be digging for gold but selling picks and shovels. And people kind of took that thinking to the AI space pretty hard early on. And so it was all about for quite a while, the semiconductor chip space, you know, the things that are powering all of this AI. Then it moved to the infrastructure layer with the foundation models. And for quite a while, the application layer was kind of derided as just a ChatGPT wrapper. I think people thought that AGI was going to get here super, super quickly and that the foundation models would be able to handle absolutely everything. But what we're seeing is that that isn't the case. It needs the application layer to make these foundation models useful in the real world. And people are really, really start to dial in on this. And so here's a great. Cycle that I'll share along with the video, but the one particular thing in here that I like is this image here right where it makes a comparison between the evolution of the cloud over 10 years and where we think the evolution of Genai will go over the next 10 years. And so on the left here we've got where the value was derived in the cloud 10 years ago and so we can see 10 years ago, you know a lot of the values at that semiconductor space, the chips that run the whole thing and the data. Warehouses and then there was a little bit of value at the infrastructure layer and then only it's high and a little bit of value at the application layer. But then today that triangle has inverted, right? So there's a lot more value being created in the cloud at the application layer today than there was 10 years ago. And they're kind of making the argument which resonates with me that we're going to see the same evolution in Gennai. And so over the last couple of years, a lot of discussion around, you know, the, the chip side of things. NVIDIA, the data centers, then a lot of discussion around the infrastructure side, which is the foundation models, not much attention being paid to the applications, but that is going to invert again over the next 10 years and we'll see that the biggest opportunity is in the application layer. So super interesting for me and an area I'm obviously always paying attention to. Link to this article along with the video. Now the second area that I'm paying a ton of attention to is podcasts. Absolutely love podcasts. You can see my Spotify account here and you can have a bit of a. Insight and other kinds of things are listening to and as you would expect, it's pretty much all AI and a bit of business stuff, which is cool. But the one podcast I'm really, really loving at the moment is 20 VC by Harry Stebbings. I'm sure a ton of you would have listened to this podcast. I think this is just such a phenomenal interviewer, interviews business leaders and a whole bunch of topics, but is really, really dialing in on the AI space as well. And so just one particular episode. Listen to just recently. Yeah, here we go. This was a great episode with a huge AI bent to it, but it was talking about perplexities, growth machine, what worked, what didn't work, and how they've been able to achieve such amazing growth in the application layer of AI as well. Maybe that's why I gravitated towards it. But yeah, such a great episode, but there's tons of really good ones in here. I felt this Bolt episode was super, super interesting as well. You know, if you're in double sided markets, then that's an amazing listen. In fact, I would recommend pretty much all of the episodes in 20 VC, super interesting from a general business sense and particularly relevant to AI in some of the recent episodes. And then the third thing I'm paying close attention to is the evolution of agents and the leaked release of an open AI agent code named Operator that can take control of your computer. So we've seen a couple of other foundation models releasing or at least previews. Of AI agents that can take over your computer. We saw the Anthropic 1 not long ago. Google's also reportedly working on one. But now Open AI has unofficially entered the fray. So this was a leak about a week ago that they are launching an AI agent called Operator that can take control of your computer in January. Very interestingly, supposedly it's only going to be available via the API to begin with. And so that is put it in the hand of us developers that can go and take that and actually plug it into the applications we're developing and figure out how to make this really, really useful for users, in our case, business users. And I think that's the way this is going to go, right? There is no way a business particularly is going to just let their team adopt an AI agent that can take control of everyone's computer, do stuff on it without some very specific guard rails in place and also ensuring it is working in the best interest of the business. And so this is something I'm paying a lot of attention to because I actually see it as a massive advancement in capability, not just in agent performance, but in what we'll be able to do. With our agents. So yeah, watch this space. I don't think any of this is official, but you know, where there's smoke, there's fire. And so looking forward to January and seeing some cool new stuff coming out from open AI. So there you go. Those are a couple of things that I am paying attention to in the AI space. As always, the evolution of the application layer. Get a lot of inspiration and great lessons from podcasts, particularly that 20 VC podcast. Absolutely wonderful. And yeah, really paying close attention to the evolution. Of agents that can operate in the real world. Thanks a lot.