DailyPost 3006
CAN WE CREATE PATHBREAKING LLM USE CASES?
ChatGPT and artificial intelligence have become one thing for the general user. What it finally connotes is no one’s baby, the fact of the matter is that it is able to cut his drudgery to a great extent. Life becomes effortless with the near accepted understanding that the output out of ChatGPT would both be correct and serves the purpose as well. It has been proving it so, and there lies the exponential acceptability. ChatGPT is a large language model coming out of the GPT fold. This is not the only one, other companies have followed suit and are giving it a tough competition.
Generic usage has now become a given. What about use in hundreds and thousands of use cases in any number of areas of operations, in specialised areas and in many areas where one had started believing itself that technology would not be of any great assistance. Developing technology is just one part of the story, the bigger and the more critical part of value for both the user and the enterprise are the use cases. It would prove the specific utility and take the product / use case to a different level.
As we all know the IT industry runs on use cases, and that is the first question that is asked of any technology to prove its robustness and prove its specialised utility while it being a generic platform. This is what general purpose technology is all about and has been true of each and every general purpose technology in human history. Such use cases are falling by the dozens in the western world and we seem to be fully devoid of that. In nearly exactly the same manner we have been missing out of one technology after the other.
To get into developing a use case is not as easy as it is made out to be and it is precisely for this reason we are literally struggling in our cherished dream of making use cases and then taking it to the wide wide world. The training on the data is the most critical and so is the problem. The two are intertwined and it is not like any other problem statement. Data in the generative pre-trained model is critical, so as to have a clear chance of proving or disproving the model. The lack of LLM model use cases is coming in the way of its professional usage. We can’t keep expecting a readymade final product till the last mile.
WE ARE STILL A LONG WAY OFF FROM EVEN THINKING OF LLM USE CASES.
Sanjay Sahay