Raw LLM Responses

The best way I can explain this is with computer programming. In the beginning, we had to use just 0s and 1s and code in 8 bits at a time to represent a value. Then came assembly language, where we had very simple short codes to represent actions like add, subtract, and, or, jump, and compare. Assembly language came about because someone wanted to make it easier to code. Then, someone took assembly language and created early language-based code compilers, such as those for COBOL, which allowed you to write in a language that was very English-like and wordy, or languages that let you code in repeatble procuedures, and we got more math abilitiy. We built up libraries to make our lives easier, and as computers got more powerful, more of those libraries started being included in the languages. I remember a time when we had to put pixels on the screen independently. Over time, in particular with Windows and macOS coming out, the whole screen was pixels, and we no longer had to do individual pixels, but we could call code built into the language that would put text or images on the screen in a single line of code. Today, the SDKs we have available to us let us animate a graphic and move it across the screen in a single line of code. Each of these steps made life easier for programmers. Each phase required new skills and a new way of thinking about at the same time, aniquating people who had done things in the previous generations. AI is no different. In my last role as a tech writer, I had to write all of the words that went on the page. I had to do the interviews with subject matter experts and produce the documentation. Today, I'm encouraged to just write a good prompt and get my base article written, then I just have to correct any mistakes. Instead of taking several hours to do all of this, I should be able to do the work in a much shorter time frame. I'm not fully ready to trust AI to do this, in particular when my work is propritary and I know our LLM doesn't