Web 3.0: The Agentic Internet

in #hive-16792226 days ago

We are heading towards the Agentic Internet. This is something that is going to happen regardless of the ownership structure.

Naturally, anyone who regularly reads my articles knows I am a fan of Web 3.0 because of the ownership structure it provides. Under this design, the economic means of production are no longer relegated to either the private (corporations) or public (government) sectors. We have a third option: communal ownership.

Of course, for this to come into being, we have to develop Web 3.0. We will not go into all the levels required but, needless to say, it all starts with data.

This is what the digital world is built upon.

It is also something that is holding true with the Agentic Internet.


Image generated by Ideogram

Web 3.0: The Agentic Internet

What is the Agentic Internet?

Basically, we are looking at an internet that is run by AI agents. We are starting to see the early stages of this being developed. Over the next few years, we will likely see billions of agents created. This is going to spread like wildfire.

Some platforms are already traveling down this part. They are providing the ability for people to develop agents.

Coinbase is one such company:

Coinbase unveiled its latest tool, promising to bridge artificial intelligence with cryptocurrency through new developer capabilities. Lincoln Murr, the Coinbase developer behind this project, announced that the new platform, "Based Agent," is a tool that allows developers to create AI-powered crypto agents in less than three minutes.

Source

This is important because I do not think we can separate AI and cryptocurrency. They go together. Crypto is going to be the medium of exchange for AI agents.

Then we have Google. It released Project Astra. This is what that is focused upon:

Project Astra, originally unveiled at Google I/O 2024 in May, aspires to create AI agents capable of multimodal understanding and interaction with the physical world in real-time.

It seeks to give AI agents multimodal capabilities. What this means is learning from more than just text. Instead, we are looking at something that will be able to interact with audio, video, and images.

Any further updates are now delayed until early 2025.

It was recently announced that LinkedIn added an AI hiring assistant:

The platform's new AI-powered Hiring Assistant is designed to help companies find the right talent by handling time-consuming administrative tasks like candidate sourcing and application reviewing.

Source

At the moment, this is an aid in handling administrative tasks. However, do not think for a second that we are not going to see this evolve into a full blown hiring agent which makes the decisions on behalf of companies.

In other words, the world is going to AI agents.

What The Agentic Internet Looks Like

Welcome to the death of the website.

When something is with us so long, it is hard to let it go. It is like a child going away to college. The years of the youth are gone forever.

The Internet is about to move to another stage of evolution. websites were the adolescent years. We are embarking upon adulthood.

Basically, we are looking at the entire framework of the Internet changing. Hyperlinks are the mechanism for travel. The webpage is the basic element, with a series making up a site. We travel from site to site (actually pages) to get what we require.

That will no longer be the case.

The Internet is going to be mostly automated. Agent-to-agent interaction will occur. Webpages will be akin to 8-track tape. While people listen to music, almost nobody uses an 8-track tape player.

We are still going to garner information. The difference is we will not be basing things on webpages.

Instead, we are looking at the digital platform. It is here where the agents will do all their work. It is also why Web 3.0 is so crucial. Do we want this run by the likes of Google?

We gain a bit of insight look at AI Overviews by Google.

AI Overviews, which employ Google's Gemini AI models to summarize answers to search queries, will now be available to more than a billion monthly users.

The paragraphs written by the AI Overviews feature are concise and include links to the sources used to compose them, usually articles from publications, business websites, or related sources. Google has claimed since they debuted that AI Overviews have led to increased engagement with diverse sources on the web.

Source

We can see how this is going. Ultimately, Google is going to use AI to summarize all the information is can put its crawler on. Links might still be provided but the click through rates will decline.

This is especially true when AI agents are the norm.

An AI agent is not going to gather its information by engaging with a website. Instead, it will simply interact with another AI agent. In this instance, Google's will provide what is needed.

Do you see the power that Big Tech can amass? Without decentralization, we can see the major platforms taking on even more power than today.

Personal AI Agents

Everyone is going to have an AI agent. The same will be true for every business.

An AI agent can be thought of as an entity with skillsets. Each will be trained on certain task, capable of performing what is required. The skills expands as more data is processed through, enhancing the abilities.

When it comes to a personal AI agent, it will know the individual, being fed everything that is done. Since all data will be tied together, the agent will keep updating the profile.

From this point, the agent will deal with the outside world. Of course, that will be made up of agents, representing everything under the sun. If you can think of a skill that is performed online, there will be an agent for it (likely many).

The Agentic Internet is one of automation.

Up to this point, we are focusing on the digital world. Eventually, we are going to see this tied into the real world. That is, however, a matter for another article.

By 2030

By 2030, AI agents will be the norm. There will be billions of agents operating, meaning the overwhelming majority of Internet traffic will be non-human.

This might seem surprising to people yet there is history there.

To start, we know that more than half the Internet traffic is initiated by computers. Humans are responsible for less than half the activity online.

When we look at the stock market, we see how this unfolds. Due to the advancement in technology, roughly 85% of all traded are "decided" by a computer. On top of this, all trades are processed by computers.

In spite of this, most people have not shied away from the stock market or see it as less valuable since humans aren't running it.

Another area we can focus upon is the paying of bills. How many of us have some type of auto bill pay we use?

That is what AI agents will do. Stretching beyond bill paying, it will get to the point where it will automatically order our groceries. These will likely be delivered by some type of automated (or robotic) system. Perhaps, at some point in time, our personal robot will receive the order and put the groceries away.

While that might be a bit further down the road, we can see how agents will take control. Therefore, Web 3.0, in my view, is crucial. The data race is on. We know the ability of the agents, in part, is due to the data available to it.

Once again, we keep coming back to this point. Who is acquiring the data and has access to it?

It is crucial that AI agents are on blockchains. We need the activities of these agents generating data that can be placed where others can use it. This is what permissionless networks do.

The tens of billions of agents will exist by 2030. My only question is where do they reside? Are we looking at people using agents tied to Google, Meta, X, or Amazon? Or do we have an accompanying set that are available on blockchain?

Our answer will come from the development we undertake today. If people start thinking in this direction and, more importantly, built it out, then we are looking at the opportunity for an alternate power structure to form.

The Agentic Internet is on its way.


What Is Hive

Posted Using InLeo Alpha

Sort:  

I am not sure how quick it accelerates but it is something to consider.

The laborforce participation rates peaked in the US in 1998 and globally in 1990. So as a percentage of the working age population, it is already down.