Wednesday, September 6, 2023

To Bard or not

TL;DR -- We're approaching a year since the release of OpenAI's system. There are many others in the mix. After a little review of the situation, we [will] look at an interchange with Bard (of Google) with respect to core items that are usually ignored. [As such, we] We will demonstrate that such oversights are not tenable now (the piper always arrives; karma unfolds its retribution or reward). 


Technology is a focus that we have adopted. We see it and its influences all around. Of late, machine learning (ML) has risen to attention after decades of preparatory  (saw this; forgot to fix; Boston influences caused the "r" drop ;>) work. Well, a side note would emphasize the improvement of hardware over that time period as being of primary importance. 

Actually, we can thank gaming and its emphasis on graphics that is fast and representative beyond merely providing stick figures [plumbed out and shaded]. Recently attention has grown even more in this area with the stock of Nvidia bursting out as a consequence of its last quarter's boom in revenue. Nvidia made chips for the graphics processing used in gaming (GPU) which are now [which now] have another alternative that has potential. That relates to a better way to handle the processing of numbers. 

ML requires huge expenditures of several types, including computer time. One expectation is that the new Nvidia chip will get things done faster or allow more work within the same timeframe. But, what is ML and why is it getting attention? That is one topic that we will cover. Is this topic related to artificial intelligence (AI) which has seen so much coverage in the press? Yes. 

Right now, there are two items to present. Over the past few years, we have had several posts which covered various aspects of this new way of doing business. We have plans to continue that emphasis but more thoroughly and with regularity. 

  • Technical - The history goes back to the 1950s in the early days of computing. A major division of approaches then related to our understanding of intelligence. The next bullet looks briefly at the overview issues. For now, here are two recent posts: Getting technical; Year to date, 2023. ChatGPT of OpenAI was released in November of 2022. It attracted the attention of millions of people who started to experiment with the system. Now, several months later and after an intitial negative reaction by the academic world, we now see this system being used as a tool by students under the auspices, and guidance, of professors. One important topic will be how to get academic knowledge covered by these systems as the first pass ignored such a requirement. 
  • Cultural/Philosophical - Given the intelligence focus and that this deals with the core of human abilities, we have to relook at how philosophy is important to computing. Such a recognition was lost which we will discuss. Too, this does apply to the U.S. and its history and people. At the same time, we are talking worldwide influence due to the role that the "cloud" architecture plays in this area. All of this is both technical and more general. ChatGPT has a peer from Google named Bard. We have mentioned work with both of these. There are others. But, consider Bard looking at Kant who was a pivotal entity (in his time and now). The pre-Kant look considered ethics and its importance. The post-Kant look bring to bear some serious issues, some of which seem to induce pessimism. An example of that is the dystopian view that seems to have erupted with strong emphasis of late, as seen in certain quorums found on social media. 
As well, Kant brought "critique" to bear in his look at metaphysics. In regard to ML (or AI, for which we have used AIn't) there have been people on both sides of the arguments. We particularly have like Dreyfus' position (see this post: Alchemy and AI). 

This post is brief. Going forward, the breadth of the topic will require several threads to be followed, simultaneously. We will try to keep that coherent. On the other hand, the fact that the computer and its AI is ubiquitious demands that we pay attention to details as well as motivations. 

Sounds difficult? Not really. Think of the 400 years that have been documented here of which only a minor part has been studied. That will be changing. Again, new ways will be central to the new work. 

So, on to the future. 


Note: This was typed up in about 15 minutes (late in the evening of 09/06/2023 with a limited amount of time available) and pushed to Quora and Facebook. The typing, per usual, is by touch while composing in the mind using the editor provided by the blog. Use Word (or other) and cut and paste? There used to be issues with text markings and such. But, the style is so not what is being done by xNN/LLM that I'll continue (a human is doing this; cannot be replicated; we'll test that someday). This am, I looked at the post via smartphone. Saw a few errors (nothing major). Per usual, I make corrections later and then change the "Modified" date. Why this topic and choice? Technology's lure and promises versus its downsides which I have observed up close for decades. No magic there, folks.  

Remarks: Modified: 09/07/2023

09/07/2023 -- A few changes, wording, grammar. 

No comments:

Post a Comment