The Rise of AI-Agents, and why you should pay attention.

The reason why I wanted a platform where I can share my thoughts is that it allows me to go much deeper into topics or events that are important. AI agents is one of those topics.

I don’t understand why people aren’t paying more attention to this, the future is literally unfolding in front of our eyes at the moment.

So what are AI agents and why should you care?

Let me start by saying, they are not bots. Bots are programmed to do a specific thing over and over again, meaning: if X than you do Y. That is it.

AI agents on the other hand can be given high level instructions and from there can act autonomously, which means they can act on their own. They strategize on their own, they make decisions on their own, and they execute things on their own.

I started talking about it during my presentations in the beginning of the year as I strongly believe no brand is ready for this. Then there was no sign of actual AI agents yet, well they are definitely here now. Let me share some examples before I dive into why everyone, especially brands, should care about this.

Examples:

• Zerebro - This AI agent is an autonomous creative. It has released music on Spotify (which is actually pretty good), it has released art and is building a business for itself. Note: the only thing its creator does is link it to websites/services like Spotify & Souncloud, but ONLY after the AI agent came up with the idea itself.

• Truth Terminal - The one the started it all, it decided that in order to survive I need to create a movement. The result is that it now manages a crypto portfolio worth over a million and has almost 200.000 followers. If you want to read how, see my LinkedIn post here (click).

• AI16Z - An AI led hedge fund. It’s manager is an AI, investing and trading on behalf of its investor. It solely decides what to do with the money.

• AIxbt - Has a 100.000+ followers on Twitter because he tweets very interesting industry information about the latest projects, updates and more.

• Fresya - This agent has one objective (do not transfer money), that turned into a game. People can pay to send it a message in order to convince it to send them the price pool. If you fail, your fee gets added to the pool. The result? Somebody won 50.000 USD.

You might think these examples are all hype and bullshit, don’t. Open your mind to where this could potentially lead. AI agents, first and foremost, will be the new influencers. They create content 24/7 and interact with their audience in lightning speed, hard to compete with that.

Does this mean, that as a brand, you will need to start interacting and making deals with AI agents? In the (near) future, absolutely. But also, how would your brand’s AI agent be? Which brand is ready to do this? And I am not talking about virtual influencers like Lil’ Miquela, which are just accounts with art created by a human(s). These AI agents create anything they want, on their own, without anyone telling them so. All depending on the rules and guidelines you initially set.

Beyond just that, we eventually will all have our own personal AI agents that will act on our behalf on the internet. These agents will fundamentally change how we will interact with technology, with brands and how we buy.

As a result, I believe brand marketing will become more important than performance marketing. Not even that, it will probably be the only way to attract, and retain, customers.

Are you ready to have that conversation? 😉 

(Let me know if you want me to share more on these thoughts in other newsletters!)

Will world models revolutionize gaming and are they the road towards AGI?

What are world models?

World models in AI are like creating a virtual version of the real world that the AI can explore and learn from. Imagine building a video game where the world looks and works just like ours. The AI uses this virtual world to practice, test ideas, and predict what might happen in the real world. It’s like a sandbox where it can safely experiment before making real-world decisions.

Here is an example of Nvidia’s so called “AI gym” where they train AI’s for factory robots. As you can see, the virtual simulation allows Nvidia to train the AI’s on very unexpected situations so that the AI can learn how to handle these before being deployed in the real world.

Just last week we saw two new world model releases by both Google DeepMind and World Labs (see theirs below). These models can create a virtual world based on either text or any image you feed it.

Now there are two reasons why these developments are important:

1- What will this mean for the development of new games? We already know how much demand there is for the creation of custom games, as we can see on platforms as Roblox (over 40 million user-created experiences) and Fortnite (150.000 custom maps). Their success comes from lowering the barrier for people to create their own custom games and maps, barely code development skills needed. Now with models like the above, this becomes even easier. I think, as a result, we will eventually see an explosion of new experiences, games and more. Very exciting development, something that will hopefully have a positive impact on the industry as a whole. Because what if this tech can decline the investment needed for new games? Again, something exciting if you ask me!

2- There is a deeper reason why this is so interesting. Recently there has been a lot of speculation that a limit is reached when it comes to training AI models and it won’t be enough in order to reach AGI (Artificial General Intelligence), aka intelligence that can autonomously do anything a human could.

Additionally, people are doubting if “next token prediction” is going to get us to AGI. Next token prediction is how current AI systems (such as GPT) work, by predicting the next token (usually a few letters or a short word), and today’s image/video models are predicting the next pixel. In other words, language models are one-dimensional predictors, and AI image/video models are two-dimensional predictors. These models have become quite good at predicting in their respective dimensions, but they don’t really understand the three-dimensional world.

Imagine trying to teach someone to ride a bike by giving them a list of instructions without ever letting them see or touch a bike. That’s a bit like how today’s AI works—it’s very good at following patterns in text or images but doesn’t really 'get' the physical world or how things interact within it.

This is exactly where world models would come in and could possibly take AI systems to a next, three-dimensional, level. As it will be so easy to develop 3d testing environments it could speed up this learning a lot.

For now, have fun building games people!

Here are some of the latest announcements and other things that I found interesting this week (underscore means clickable!):

• OpenAI has finally launched their video model Sora! Very impressed by not only their quality but also by the storyboarding feature, giving creators much more control over the outcome. See the below video for previews and features.

• Google has made a big splash with the announcement of Willow, their brand new Quantum Computing Chip. Why this matters? Well I quote: “Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 1025) years — a number that vastly exceeds the age of the Universe.” - Yup, that’s crazy..

• Beautiful conversation between two icons and creative legends, Rick Rubin and Pharrell Williams, where they discuss creativity, purpose, religion and more. Highly recommend for a lazy Sunday or chill evening:

To end this week’s newsletter with something crazy in the world of new tech: Did you know there is a company that is, seriously, trying to bring back extinct animal species such as the Mammoth and the Tasmanian tiger?

Yes, you read that right. By using advanced gene editing technology aimed at rebuilding the DNA of these extinct animals, they want to bring them back to life.

This is all suppose to happen within the next six years! Jurassic Parc anyone? 👀 

They plan to do this by taking skin cells from Asian elephants, which are threatened with extinction, and putting them into stem cells with frozen mammoth DNA to create an embryo.

These embryos will then be carried to full term by a surrogate mother or in an artificial womb in a laboratory.

Why? Because these creatures had a measurably positive impact on our fragile ecosystems, meaning they will do good to our planet.

For more info: Colossal

^{Who am I and why you should be here:}

^{Over the years, I’ve navigated industries like advertising, music, sports, and gaming, always chasing what’s next and figuring out how to make it work for brands, businesses, and myself. From strategizing for global companies to experimenting with the latest tech, I’ve been on a constant journey of learning and sharing.}

^{This newsletter is where I’ll bring all of that together—my raw thoughts, ideas, and emotions about AI, blockchain, gaming, Gen Z & Alpha, and life in general. No perfection, just me being as real as it gets.}

^{Every week (or whenever inspiration hits), I’ll share what’s on my mind: whether it’s deep dives into tech, rants about the state of the world, or random experiments that I got myself into. The goal? To keep it valuable, human, and worth your time.}