We are watching a brand new age emerge.
When it comes to AI technology, I think we are in the ARPANET days. We are still a long way from the emergence of the World Wide Web. This means that it is still very early and the time to get on board is now.
There is little doubt that chatbots are changing life on the Internet. This technology, at least from the general public perspective, is less than 18 months old. Nevertheless, it is safe to say the advancement is mind-blowing. Each update by the major players shows massive advancement compared to what was before.
This truly is an arms race.
We are also seeing some huge volume. Here is one I use that is not one of the major platforms. Yet notice the number of requests.
I am sure this pales in comparison to what Google Gemini, Claude, or ChatGPT get. Whatever the numbers, it is safe to say they are large.
Chatbot Technology On Hive
There are a couple of projects that I am aware that could be forming in this direction. There might be others on Hive that I do not know about so we will confine the discussion to what is known.
Basically, every application is going to require this technology in the future. If one does not have it, this will be akin to your site having to be found by the numerical address (192.168....) as opposed to just using the name.
We have to keep in mind, Chatbots have the potential to solve a major problem on Hive: search.
One doesn't have to spend much time to learn how difficult it is to find information. How often has we looked for something we wrote, and still had a tough time uncovering? If it is not indexed in Google properly, which things often are not, then we have issues.
Speaking of Google search, we know there are "penalties" when looking at Hive content due to multiple site posting. One of the features of Web 3.0 is the data is open. Thus, any UI is free to pull the information. That means everything posted to Hive is basically on at least 6 or 7 websites instantly.
Chatbots solves this depending upon how they are set up.
With the typical chatbots, we are seeing the answer to the request provided along with references to where the data came from. Not all do this but many do. Here is where Hive can have chatbots providing information along with links to the content where it was found.
Specialized Information
Before getting into how this could unfold, we can discuss the two initiatives that are working in this direction.
The first is LeoAI, which is being developed by the Leo team. This was started months ago and began the training by feeding in the 1 million or so Threads. From what I understand, the articles are the next set of data to be fed into the machine learning engine.
Then we have this which is a portion of a conversation I had.
As we can see, the CTT/SpkNetwork team is also in the process of putting something together. This makes sense since, as stated above, most applications are going to require this in the future.
What is important is to focus upon the specialized information.
Basically, the set up looks something like this:
The teams start with some LLM that is available. For the sake of this discussion, I will use Llama since it is open source. There are a variety of options so there is no guarantee this is the exact path undertaken.
Llama is a Meta development. Again, we will make a presumption and state it was trained on data from Facebook and the assorted applications. Obviously, this is a ton of data providing a massive cross-section of information. Perhaps there was other data fed in such as Wikipedia.
Once the basic model is available, the Hive-based teams will feed its data. This would naturally be what is posted on-chain.
Of course, they are free to pick and choose what they want to use. It would make sense to feed whatever is available but there are always other factors such as cost of training.
Nevertheless, this is where some of the specialized information comes from. It is also what will separate these chatbots from the major ones.
Hive Answers
We all know that it can often be difficult to find answers to basic questions about Hive.
The first problem is the name. In most instances, if we type "Hive" or "Hive blockchain" into search, we get information that is not related to this ecosystem. Under this scenario we either get information about bees, Apache's software, or the Bitcoin mining company.
Here is where these specialized chatbots can solve a problem. As they are trained on the different Hive based data, these answers will be available. Imagine being able to simply use the chatbot to ask "what are resource credits" and getting the answer, as it relates to Hive.
For example, we might find an answer to the resource credit question but what about MVests?
That is some of what could be possible.
While some of the general information might be available, pulling the data from Hive means we can get a great deal more specific. This is the natural outcome when the learning engines are fed a plethora of Hive based information.
There is another layer to this that opens up: individual accounts.
Since we are feeding information relating to accounts, we can see how that data could be structured. We see this in some of the front ends pulling the number of posts or comments made. The data is recorded on-chain so it is a matter of pulling it. Obviously, we are at the mercy of the application developers and what they program in.
What if, however, one could use the chatbot to request the number of posts one made within a certain time period? How about if one could ask it to provide a list of the Top 10 posts based upon number of comments? Perhaps we ask who has commented on my posts the most since I started writing.
The point B=being there is blockchain data that is not picked up by the likes of Google. It is all there but it has little use outside of Hive. Here is where we start to see this being a resource center for those who are utilizing the ecosystem.
Providing A Service
The final aspect pertaining to this that I want to cover is the idea of offering a service.
We can see where the future lies for some of these LLM developers. OpenAi, for example, is going to take a lot of their most advanced technology and target businesses. This is where the money is. At the same time, many can use it to leverage other aspects of their digital platforms.
X is incorporating Grok and allowing anyone with the premium membership to utilize it. Does this enhance the value of that subscription service? Most likely. The same concept could be applied by Amazon with its Prime membership.
On Hive, applications have a service to offer the users. How this is factored into the economic structure is up to the individual teams. That said, there are many ways to use this to develop a revenue stream, something that is important if we are looking at some type of "DAO" concept.
This is where innovation and creativity enter. Perhaps there could be other aspects built in, that excite the user base. Naturally, with Web 3.0, we also have to highlight the fact that many users have stake in the economics (tokenomics) of the platform.
The Future
We have not seen the end of the capabilities of chatbots. This is something to keep in mind.
Whatever Meta adds to Llama is going to eventually enhance what is available to everyone. The different initiatives that are using that software as the basis are going to be able to add the newer features to what they develop.
The best way to look at these LLMs is as operating systems. One chooses a system and then any updates that are added can be incorporated. For mobile, we basically have IOS and Android. This the foundation and the developers build applications on top of that.
We are dealing with something similar. Something like Llama is the basic foundation and the developers can take that and add to it from there. This means creating different features that are important for their platforms.
This is what we can do on Hive.
Fortunately, there are a couple teams already looking into this process. We will have to see if there are others.