ChatGPT is making headlines once again.

The company announced that it released its AI search, which allows it to pull data from the Internet, something that it was previously blocked from doing.

This is an outright threat to traditional search engines such as Google. Of course, it should be noted that it too is diving into AI search, as is Meta.

Ultimately, we are going to see all kinds of AI search. There is no way the traditional search engine survives.

OpenAi spend a great deal of time (and money) getting partnerships with different publishers. This is the feeder system into the real time search capabilities.

Image generated by Ideogram

AI Search Is Spreading

The key to all this is updated information.

Most of us have done a prompt on a search engine about something rather recent and gotten a reply stating that "I am only trained up to such and such date". ChatGPT was certainly in this camp.

The reason for this is the lack of real time data. ChatGPT would update each time a new version was rolled out. However, it lacked the ability to expand on the fly.

AI search is changing this.

Ironically, this is an advantage that Google, Meta, and XAI have. Since they have very popular social media platforms, information is being fed in their daily. Is there anything that takes place in the world that isn't on one of those platforms? X states that it is the number 1 source for news. Whether this is correct or not is immaterial. The fact that it potentially could be is the telling part.

What is important to remember is we are dealing with first iterations of this technology. AI search is embryotic. It is likely that a couple more generations are required to see massive impact.

Where this gets really interesting is with smaller language models. When entities use a Llama3 as the main model, it gets the most up to date training available. It also provides whatever else was built by Meta (or the community).

That said, there is also the data the SLM is trained on. Depending upon the search requirements and data input mechanisms, this also can pull a great deal away from traditional search.

For example, let us say there is an AI model that deals with the EPL. People are feeding data in daily surrounding that. Hence, we have a database of historic information about clubs, players, and managers. articles are fed in along with statistics. Anything remotely related to EPL is in there.

What do you think the search will look like there. Basically, this would be the most knowledgeable EPL search there. It will blow away Google while also usurping the LLMs AI search.

Public Versus Private Search

Ultimately, all of this boils down to action. Dealing with information is one piece of the puzzle. However, what about other information that takes place online.

We know recommend engines are the norm today. AI models actually enhance these. Over time, as more data is fed in, and the weighting is set by the SLM to focus on that, we see how users information becomes available.

On social media, do you know who you voted for the most. How many likes do you pass out daily? What Tweet got the most comments? What YouTube video got the most likes. Who liked your stuff the most?

All of this becomes available with AI search. Data of every nature is captured and fed in.

OpenAI is looking towards the future. When asked about autonomous features, this is what the reply was:

Even more eyebrow-raising was the hint of autonomous features for ChatGPT. One Reddit user asked if the chatbot could do tasks independently, to which Weil responded by saying that would be a “big theme” in 2025.

Source

As organization move in this direction, we are going to see both public and private agents developed. The idea of autonomy, in this era, most likely means AI agents. It will not be surprising to see OpenAI build these into the search functions.

This is going to spread like wildfire. Do not be surprised if traditional search is dead by end the end of 2025.

---

What Is Hive

Posted Using InLeo Alpha