In today’s AI Adventures, I fed ChatGPT4 an article summary created by another AI tool, Perplexity. Perplexity does a great job with summarizing, is available right from any page I’m reading, and is a great tool for research and understanding. You can use it to ask questions about the source page you are on as well. Like all AI tools, the better the prompt (“summarize and include bullet points for the main talking points”), the better the output.
Once I fed the summary to ChatGPT4, I could ask it to summarize what it just read in order to confirm that the context was set.
Once I knew we were on the same page, I asked ChatGPT4 for a list of possible ways to implement the theory proposed by the article.
It wasn’t a bad list; it did need a bit more context, however.
Okay, better. But I wanted it to be even more relevant to my life specifically. So I fine-tuned the recommendations by feeding ChatGPT4 my list of core values with a short description of each value. Then I could ask for recommendations based on that knowledge.
Jackpot! Much better response. Nothing new given a lifetime of fascination with self-improvement but still helpful to see if presented in this manner.
That’s still a lot to tackle at once. Can we refine this further?
That list is the work of a lifetime, at least the type of lifetime I’d like to experience. Of my experiments, this one was the most engaging. I look forward to thinking with ChatGPT more.
Some takeaways from today’s session
1. Combining AI tools can be incredibly useful. Each one has strengths and weaknesses and experimenting to see which one handles a specific scenario best is worth the time.
2. Don’t trust that what any of these tools tell you is true. Confirm everything. I read the article first before using AI to help me work with its content so I would know if ChatGPT4 was tripping. I did the same exercise with ChatGPT3.5 and it took far more far-tuning. It went into complete hallucination mode and needed several course corrections.
3. Establish context and confirm that the context is accurate before going deeper into the material.
4. Continue to provide context with prompts as helpful, such as adding in another reference to deepen the matrix.
5. Go past the novelty and get to a place of value. It’s easy to get caught up in a bit of awe at these tools. But the real paradigm shift is in using them to gain clarity, perspective, and a deeper level of understanding. You know, the human stuff!
- Background Ops #1: Strict Limit by Sebastian Marshall
- Featured image from Dall-E based on this prompt: Create a cartoon image of a person lying on a couch in an office. Next to them is a chair. In the chair is a computer. The computer has a face and is looking at the person on the couch.