Novel ideas and vision become more important than execution.
LLMs and AI agents make it easier than ever to become a creator.
True domain expertise becomes even more valuable.
Full article:
The role of AI in the Creator Economy
Creators will not compete with AI, they will thrive with it.
This piece is a collection of my thoughts on LLMs, AI agents and the role they will play in the creator economy.
Novel ideas and vision become more important than execution
We have all heard the saying “ideas are cheap, execution is everything”. But as larger portions of the execution become automated, the value of the idea, and the vision that drives it might become the true gem of value creation. This stems directly from the known limitations of LLMs, the models backing what most people think of as “AI”. Let’s cover the three limitations which I believe will make novel ideas and vision one of the true drivers of value.
Incapable of producing novel ideas
Individual LLMs cannot produce novel ideas. A paper published by Google’s Research team stated that there was “little evidence that the model [...] is able to generalise beyond its pre-training data”. What does this mean? It means that LLMs cannot derive new connections between concepts contained in its training data or, create new concepts beyond its training data. In other words, the model only knows what is already known. While there is tremendous value in having access to an AI that knows everything that is known, it becomes a lot less useful when it comes to producing novel content or ideas.
In a world of democratised LLMs, noise will be abundant. Most content will be average, and blend in a sea of re-mixed old content. Humans crave novelty, and will crave it even more in a world of abundant repetition.
Tangent: humans crave novelty because we ourselves suffer from repetition and routine. What is novel to the mind, is fertiliser to the soil of consciousness, literal “food for thought”. Repetitive thoughts make us feel stuck, and turns our mind into stale water. We crave novelty for a healthy mind.
In a world dominated by widely distributed old ideas, what will really stand out will be novelty and authenticity. This has been true long before AI was in the conversation, but will be more relevant than ever in the coming years.
An out-of-date view of the world
The other limitation of LLMs is their out-of-date view of the world. LLMs get a snapshot of the world as represented in their training data. They cannot keep up with current events, and as humanity pushes on the edges of its collective knowledge tree, the AI awaits, to be taught again.
Humans ingest the knowledge of the world, as they see it. Our minds, in a mashup of all our individual experiences, produces thoughts, most of which are old, some of which are new to us, and a few of which are new to the world. As long as the world doesn’t stop spinning, and humans spinning with it, LLMs are stuck with an out-of-date view of it and humans will remain the only source of novel ideas.
LLMs hallucinate, but they don’t dream
We established that humans crave novel ideas. We have also established that humans can uniquely produce novel ideas, thus implying novel ideas will remain scarce. For any scarce good that also happens to be in high demand, we can only conclude that novel ideas will remain highly valuable.
But what if we were able to give LLMs an unfathomably large context window? A context window so large that it encompasses all of our present. And what if the LLM was able to keep it all in memory, as it interacts with that present. Would that not make it on par with humans?
Even then, humans would have something LLMs don’t: humans have dreams. We experience the past, that we process in the present to project into the future. In that world, there is only one thing more valuable than a novel idea, and that is a vision for the future. It is what strings together novel ideas about a world that the present has yet to capture.
Hence, the creators that thrive in this new world, are creators that can produce novel ideas, together with a vision of the future they want to shape. Equipped with those two, creators can be empowered by LLMs to see their vision through.
##It has never been easier to be a “Jack of all trades”
Most creators are solo-preneurs, especially as they get started. In addition to practicing their craft, whether that be building a new product, writing a book, or recording a new song, creators also need to do everything else that will bring their creation to the world. Whether that is negotiating a deal with a book publisher, building a website to display their work, or market themselves to the world. Being a solo creator is hard work, and the time left to create becomes more limited. LLMs could change that.
The wealth of knowledge of LLMs, reduces the barrier to entry for any new skill a creator may need to pick up along his journey. Personal branding, building a simple website, improving your language skills, you name it. If it’s been done before, the LLM probably knows how to teach the basics for it. Nothing is out of reach anymore.
In addition to that, the advance of AI agents will automate away anything that can be automated. As AI agents become more integrated into digital products, these autonomous agents will be able to perform any non-creative task which burden creators today. That will free up even more of the creators’ time so they can put their attention on what truly pushes their ideas forward.
In short, LLMs will make it even more accessible for anyone to become a creator, and as the barrier to entry lowers, we will see the creator economy explode.
Jack of all trades, but master of none
Trivial tasks will be automated away, but difficult tasks are likely to remain out of reach for a long time. The role of real expertise in a world dominated by AI will become increasingly important. As trivial and repetitive tasks get automated away, and more people reach difficult problems, two things will happen:
- The number of difficult tasks will grow
- The pace at which we reach the difficult part of a task will grow
In other words, the demand for problem solvers for difficult tasks will increase. Anyone equipped with the skills and expertise to solve these difficults tasks, will become more valuable. Expert knowledge within any domain will be increasingly saught after. Where AI will take you 90% of the way, that last 10% will require the involvement of a truly skilled person.
But who’s going to be talking to the experts?
Reversed Roles
Today, one might turn to an LLM to learn about a topic they know little to nothing about. Tomorrow, an LLM might turn to a person to learn about a topic they don’t know enough about. Imagine being able to ask an AI agent to solve a complex problem for you. The agent would do the best it can to come up with a solution, and then hire a domain expert on your behalf to work on improving it. The agent would posess enough knowledge about the topic to understand and adjust to the expert’s input, while the expert has much less effort to do trying to bridge the knowledge gap with their “client”.
Let’s say a creator requires a marketing campaign. While the AI agent may know every marketing strategy in theory, it may not have enough up to date information to truly tailor the campaign to your niche. However, while communicating with an experienced marketer, the AI agent could be steered in the right direction to accomplish the task even better.
ai-agents-talk-to-domain-experts
Conclusion
Value will accrue where AI can’t compete. On one hand, creators with novel ideas and a vision. On the other, highly technical and specialised people.
Creators will not compete with AI, they will be empowered by it. With the advance of LLMs for education, and AI agents for automation, becoming a creator will be easier than ever before.
The worlds knowledge will democratise. Getting to a basic level understand of any domain will be within everyone’s grasp. But those that become truly experts of their domain will be in great demand.
In short, there's never been a better moment to start dreaming up and building the future you want to see.
https://habla.news/a/naddr1qvzqqqr4gupzp9dxjvnyfxf34hdr9elq7cn4hmqw8patamjtk44na985df4vgqhzqqxnzde3xuengv3cxy6ngdf32dgxjp