Dipping My Toes into the Climate Change Ocean
Hey everyone! So, I've decided to embark on a bit of a learning adventure. The topic? Climate change. Yeah, I know, it's a big one, and it seems like everyone's got an opinion, often a very loud one. My goal here isn't to solve it all (if only!), but to get a much better grip on the science behind it all, and also to understand where and why the controversies pop up.
This isn't just a casual browse through Wikipedia, though. I'm trying to be a bit more methodical. Think of it as building a personal knowledge base, and for that, I'm enlisting some digital helpers.
My AI Research Squad
I'm planning to throw a bunch of AI tools at this to see how they can help me gather and process information. My current lineup includes:
- Claude
- Perplexity
- Our very own Gemini
- And, of course, ChatGPT
To keep all the eventual notes, snippets, and insights organized (or at least try to!), I'll be using NotebookLM. The idea is to create a structured repository of information that I can easily search and refer back to.
The First Baby Steps: Talking to ChatGPT
So, where does one even begin with such a massive topic? My first port of call was ChatGPT. I had a couple of initial conversations to try and lay some groundwork. You can think of this as the "scoping" phase.
One of the first things we tackled was trying to define the project itself. I was interested in exploring not just the standard climate change narratives, but also the role of things like heliophysics (the physics of the Sun) and space weather, and how they might connect to what's happening with Earth's climate. This led to brainstorming a project name. ChatGPT came up with "HelioScope," which I kind of like!
Here's a snippet of the kind of "project blueprint" idea it threw out:
Ambitious? Maybe a little for a one-person show, but it sets a cool direction!
Beyond just a name, we also worked on crafting some initial "instructions" for how an AI should approach this research. The goal was to get detailed, focused, and efficient prompts. For example, for a potential custom AI setup, ChatGPT suggested instructions like:
It even helped me condense these instructions for Perplexity, which has a character limit for its AI prompt setup. It's pretty neat to see how you can guide the AI's focus.
What's Next?
The document I started with (which was just these initial ChatGPT chats) is now uploaded to NotebookLM. The next step is to start feeding some of the more detailed prompts we developed to the other AIs, particularly Claude, and see what kind of information they dig up. I'm expecting this to be an iterative process – the findings from one prompt will probably lead to new questions and new prompts.
It feels like I'm just scratching the surface, but it's an exciting start. The aim is to build a wider and deeper understanding, piece by piece. I'll try to post updates as I go, sharing what I learn and any interesting (or frustrating!) bits along the way.
Wish me luck on this data safari!
No comments:
Post a Comment