Handreson
commented
1 year ago As a scientist, I have always approached thesis writing with a data-focused mindset, starting from my bachelor's studies and continuing through to my doctoral thesis. During my studies between 1996-2006, when I completed my BSc, MSc, and PhD, Google did not exist, so most of the information for scientific research came from scientific journals that we accessed through the library. Back then, most of the effort was focused on developing academic writing skills, which often involved making repeated mistakes. However, the challenge at the time was the lack of information. Nowadays, we face the opposite problem of information overload, which makes it difficult to distinguish relevant from irrelevant sources. Additionally, issues such as plagiarism and ethical considerations add to the complexity of modern scientific writing. The beginning era of AI opens doors that we don't yet know exist. This homework has opened doors, introducing new possibilities that accelerate and enhance our work in science as well. Initially, I challenged myself not to rely on Google as my primary source, but instead asked for help from Chat GPT. With each task, I learned to phrase my questions more effectively. This process helped me install Chocolatey and CopyQ and refreshed my memory on how to use terminals. The hardest task was making a plot in GitHub. I wanted to create a Venn diagram, but the plot that was produced did not show the correct numbers. I spent many hours working on it, using both ChatGPT and Google for help, but unfortunately, I was unable to make it work. However, this experience made me think that if it wasn't my own data, I probably wouldn't have recognized the problem with the plot. Eventually, with the help of ChatGPT, I was able to create another plot, as well as a few others, that worked well. However, these plots did not hold as much scientific value as a Venn diagram would have provided. Based on my data ChatGPT proposed a heatmap, scatter plot, box-plot, and other plots but not a Venn diagram. This experience made me cautious with scientific data, as we say in Estonian, "Usalda, aga kontrolli" ("trust but verify"). / I authored this text, but it was also edited by ChatGPT. Throughout the editing process, the text had to be reformulated several times, either by myself or with the assistance of ChatGPT, to ensure that the idea was conveyed accurately./
Asking
Feeling, Writing, Thinking
Doing
Programming & Infrastructure
copyqon Windows Powershell or Linux using Chocolatey, or install Maccy on Mac using Homebrew. This will save you a lot of time throughout the class as you copy and paste things from ChatGPT!