{
  "title": "Embracing Change",
  "date": "2023-07-18T16:40:08.373Z",
  "slug": "embracing-change",
  "author": {
    "name": "Lucas Steinmacher"
  },
  "content": "<h2>The Leap of Faith: My Journey with Artificial Intelligence</h2>\n<p>I had an abrupt introduction into AI as many had in December 2022 with the release of OpenAI's ChatGPT to the general public.  We had been using OpenAI's Whisper for a transcript automation process that we did in house to process client's video's audio files into close captioning files (also just referred to as srt file from their mime type) which we would then use to process transcripts for the clients content. After November 30, 2022 of course the scope of which we used AI on a daily basis changed dramatically. We used it to increase our understanding of content that our clients would ship us by running the transcripts of their videos with prompts to outline what was talked about so we could better understand what we were to be working with.</p>\n<p>The release of OpenAI's ChatGPT in December 2022 marked a pivotal moment in my journey with AI, shifting our daily reliance from Whisper to also include ChatGPT for various tasks. This leap into AI technology has fueled my enthusiasm to explore its vast potential and embrace the next phase of my journey.</p>\n<h2>A Coworker's Experience</h2>\n<p>As was chatting with my manager on slack one day, our conversation took an unexpected turn towards how they had created some email outlines for a client based on previous emails that the client had written.  They had used reverse engineered a prompt to use with ChatGPT for this and could generate an email in seconds based on a subject that the client was working on. Of course I was interested in the potential for this of course and asked them their process and the more they talked the more I was interested in this reverse prompt engineering.</p>\n<h2>Give Credit Where Credit Is Due: No One Learns Alone</h2>\n<p>First off I'd like to address that my manager who I'll not name for their privacy, lets call them T-Dog, has been a large influence in my learning and has been an amazing mentor to me in many regards who shared <a href=\"https://couchdeck.com/mastering-reverse-prompt-engineering-using-chatgpt/\">this article</a> with me to help guide me through the stumbles of prompt engineering and software in general.  Secondly I follow rather heavily a writer and developer advocate named <a href=\"https://www.swyx.io/\">@swyx</a> whom has written about the subject of reverse prompt engineering here. I aspire to be as good of a writer and thinker as both of these individuals, T-Dog and Swyx respectively, and like to shout them out. (I'll ask T-Dog if I can mention them here)</p>\n<h2>Unleashing Potential: The Steps of Reverse Prompt Engineering</h2>\n<p>I posted the article above but I'd like to break down the process of recreating articles, code, email, or whatever your heart desires. 🖤</p>\n<h3>The first stage: Focusing on the ChatGPT model</h3>\n<p>We need to prime the LLM with a statement such as: \" want to use reverse prompt engineering where you help me create prompts based on the text I give you that would be optimized and ideal for producing similar content\"</p>\n<h3>The second stage: Selection of a text to begin with</h3>\n<p>Choose a suitable code or text that you want to use for reverse prompt engineering.</p>\n<h3>The third stage: The reverse prompt generation process</h3>\n<p>Generate a reverse prompt by submitting the selected text to ChatGPT.</p>\n<h3>The fourth stage: Rewriting the prompt</h3>\n<p>Refine and rewrite the prompt to make it more precise and contextually relevant.</p>\n<h3>The fifth stage: The testing process</h3>\n<p>Test the rewritten prompt by inputting it into ChatGPT and observing the output.</p>\n<h3>The sixth stage: Iteration</h3>\n<p>Evaluate the generated prompt and make adjustments if necessary. Edit the prompt as needed to achieve the desired output.</p>\n<p>These steps provide a framework for utilizing reverse prompt engineering to enhance the performance of text generation models like ChatGPT. By understanding the relationship between prompts and generated text, it becomes possible to generate more impactful and accurate content.</p>\n<h2>Exploring AI's Power in Curriculum Enhancement</h2>\n<p>In my first experience using reverse prompt engineering, I witnessed the incredible impact it had on streamlining curriculum development for our clients. By leveraging the power of AI and carefully crafting prompts, I was able to generate high-quality curriculum suggestions of potential problems for learners to work on based on given finished code. The iterative process of refining prompts and adapting outputs not only saved time but also ensured that the generated problems were contextually relevant and aligned with the desired outcomes.</p>\n<p>The discovery of reverse prompt engineering has opened my eyes to not only the potential of this one technique but now looking into other new and exciting ways to manipulate LLMs especially now with the release of <a href=\"https://openai.com/blog/chatgpt-plugins#code-interpreter\">ChatGPT Code Interpreter</a>.</p>\n<h2>How I'm using it now</h2>\n<p><img src=\"/images/tinfoil-dude.jpeg\" alt=\"How do you know this wasn&#x27;t AI generated?\">\nThe magnificence of AI lies in its vast scope, presenting countless opportunities for contribution that go beyond traditional expertise. From harnessing AI as a supplementary tool in various domains to addressing societal issues, the possibilities are boundless. AI acts as an enhancer of human potential, empowering us to enhance our skills and drive substantial transformations. By embracing a mindset of perpetual learning and adaptability, individuals can uncover their unique niches and forge a profound influence. Rest assured, this perspective is purely a human-generated perspective. There's no way that I would use my skills in using a generative AI to help me write this article... pinky swear. 🤞</p>\n<p>What am I using if for now? Well debugging Typescript type errors and finding where I accidentally deleted a <code>&#x3C;div></code> in my JSX but I don't think that is the end all of the matter. There's so much potential past being just a <a href=\"https://en.wikipedia.org//wiki/Rubber_duck_debugging\">Rubber Duckie</a>. This rapidly evolving era of AI, we are witnessing the harmonious blend of human ingenuity and artificial intelligence. Not just reverse engineering an article, email or code but so much more that we have yet to dive into.  I'm excited to navigate the dynamic landscape that this intersection of human and computer interaction and am excited to see where the future leads this field.</p>\n<h2>Conclusion</h2>\n<p>Embracing the world of AI has been a transformative journey for me. The release of OpenAI's ChatGPT and my exploration of reverse prompt engineering have shown me the vast potential of AI in revolutionizing various domains such as education. From streamlining curriculum development to enhancing human-computer interaction, AI acts as a catalyst for progress and innovation. As I continue on this exciting path, I am filled with anticipation for what the future holds in this dynamic landscape of AI and human ingenuity. Together, we forge a powerful union that paves the way for advancements and opens doors to unexplored realms of discovery. The possibilities are limitless, and I'm eager to be part of this transformative journey.</p>"
}