TL;DR -- Paul Allen and his work has led to a phenomenal advance in query support. We will go on about and further look at this application of machine learning.
--
Our focus, outside of those related to history and heritage, deals with technology. Of late, we have been looking at GenAI and the changes that we can see in the world since OpenAI let loose their premature "omniscient, omnipresent" whatever. ChatGPT was the name; it had 100s of millions users within a short period of time. That set the stage for several developments are were mostly reactionary.
We, basically, analyzed the situation and the various cohorts of the new type of thing. And, found where things were wanting. On the other hand, GenAI made a hit in fiction and chicanery. Of the later, "fake" this and that. In our interchange with the things (several), we noted its superior attitude. Now it all, essentially. Telling humans what they ought to know.
When, mind you, it could not do simple algebra or even arithmetic. There were other obvious failures. The industry (OpenAI and its cohorts) bragged of 80% success. Oh yes, of all of the flights in a day, how would you like to have 20% of those fail. Well, there are types of failures. So, the matter could be discussed usefully, in this case. But, for one maker of planes, the evaluation criterion was 99.999% success which was applied throughout all of the process. The effect was something that could regularly fly millions of miles with no incident.
Then, things happened. Well, we'll see more of that at GenAI (and other types) get put into the world due to the lack of ways and means to contain the hype or to handle the proper analysis. But, we will get there. Have seen progress. Yesterday, in one quorum, ChatGPT said that it was a tool and that it wanted to help me, a human, with my research. Okay. More on that later.
Now, let's look at one of the many ways that mature minds have applied this stuff. I spent the afternoon using Paul Allen's Semantic Scholar. Well, not his; rather, it is the work of the Allen Institute and the University of Washington and other contributors. The experience was encouraging with respect to getting a proper framework going for this type of technology.
Here is their website: https://www.semanticscholar.org/ (we will look at this further). Based upon their query facility, it is a huge step forward.
As mentioned, there are other examples. One of these is Wolfram's use of GenAI with his computational mathematics system.
10/26/2024 --
No comments:
Post a Comment