ChatGPT craziness what it can do and what it cannot

ChatGPT craziness what it can do and what it cannot

Many discussions have been about how far ChatGPT can go to create entire applications. Let’s be clear. If you do not know to program or have some basic programming skills, you will not build a single application, or if you do, it will be a very simple app without lots of logic.  ZDNet explains in a blog that “ChatGPT can write useful code, but it can also get completely lost, fall into a rabbit hole, chase its own tail, and produce absolutely unusable garbage.” If you ask ChatGPT to write an entire application, it will fail. It will fail if you don’t know anything about coding and want ChatGPT to build something. 

ChatGPT requires you to know to program if you want it to help to assemble applications

ChatGPT can help and succeed when you want it to help somebody who already knows how to code to build specific routines and get tasks done. ChatGPT does not help create wireframes (yet), but I am sure that will also be there soon. Another blog entry from Zdnet explains that in terms of the scope of an app, it should be seen as a “book or set of books.” ChatGPT can help you with functions, methods, and subroutines. Within a book, these would correspond to a paragraph or a small article, or even a chapter in a book. ChatGPT can help you at the paragraph level and help you to structure your chapter’. ChatGPT can also help you to debug your code, as is discussed in the blog article by ZDnet.

Creating plans with actions has already helped me tremendously in my work 

Comparing the experiences with my ChatGPT use for my work resembles quite a bit of what is discussed in the ZDNet articles on how ChatGPT can be used in building applications. I have found ChatGPT extremely useful in crafting plans with action items. I usually ask ChatGPT to create “paragraphs,” but I am still the person and brain to put together the entire document and add my own experience and value add. ChatGPT does NOT think. It assembles together information based on the large language models (LLMs), but I still have to understand the “wholeness” of what I am trying to accomplish. ChatGPT takes care of the details that I either don’t remember or, in many cases, should NOT need to remember.

New ways of working both in professional and academic life

If I compare ChatGPT to when I got my academic degrees, ChatGPT has taken over some of the things that I had to do in a very labor-intensive way. To get references to content, I had to walk to the library, buy a book or subscribe to academic journals. Every source of information did what I had to do the hard way. In today’s world, Google or Bing search helps students to find information, and now with ChatGPT, the student can craft initial plans for articles or a dissertation. ChatGPT is almost like an automation engine collecting information from multiple books, the Internet, and articles, assembling it, and suggesting useful approaches. I will then judge whether the “assembly” is useful or whether I need to give it more specifications. The question remains whether a student working on papers or dissertations will put enough time in his/her own thinking and not rely too much on generated content. That is another topic for another blog article.


Leave a Comment