We still do not have a great deal of information regarding LeoAI. This is something that could hold tremendous potential.
Grok, the LLM that was released by X.ai in November was moved to open source. Over the last 24 hours, there are some people playing with it and recording videos. So far, the results are limited since, it appears, the only way to truly access it is through X. It is too big to run without a massive amount of compute.
That said, I did gleam onto something that I found very valuable regarding a model for LeoAI. I do not know what @khaleelkazi and the team have planned but this could serve a basis.
Grok Linking To Tweets
In one video that I watched, the individual was putting Grok through its motions. The usual tests were conducted to see how it performed.
What really caught my attention was the fact that some of the responses were provided and linked to the Tweet that was used. This is something that makes a lot of sense.
We were told that Grok was trained using Twitter data. For this reason, it would be sensible that many of the results would have Tweets that could be linked.
Here we see the individual Tweets suddenly having more value.
Could LeoAI follow something similar?
Threads Linked As The Reference
While I am going to refer to threads, this could equally apply to posts.
What is the potential of having LeoAI provide answers and then show the associated thread or post? This would mean that we have incentive to generate content that has information that an feed the ML engine and help train it.
It could radically alter the mindset that many on here have regarding their threads.
Also, this is another way to generate page views. If the person want to follow the information provided in the response, there is the ability to click through.
Fragmentation To Concentration
The Internet is the epitome of fragmentation.
Information is posted in different areas and linked in many different ways. This is true even on an individual platform. If that is the case, how does anyone find what they are looking for?
Here is where algorithms and recommend engines enter.
While some people might complain about them, it is a shortcut that is highly effective and concentrates data. This is why, when on something like YouTube, if we are searching for a topic and pull up a video, we get other videos recommended that are associated with that topic.
The same concept applies to Wikipedia and links. This platform using direct links to allow people to immediately click through to other pages that are of interest. That is why each Wikipedia page has over 1,000 internal links.
With LeoAI, we can build this concept in. With LEO threads and articles, the AI could provide a direct link to the page. While the click through percentage might be very low, if every response does this, we can see how it could result in some growing numbers.
Then we have LeoAds.
Who puts these pages together? It is done by individual users who could benefit from having more traffic showing up on their pages.
It seems like a win-win-win scenario.
Is this possible? We will have to see what LeoAI is like when it rolls out. So far, details are sketchy. However, this is what Grok is already doing.
It is a concept that Leo should adopt. Why not us our own data for gain of the entire platform?
Posted Using InLeo Alpha
