from discord to deepthought

**File 2**

Okay, so everyone's talking about AI, right? But we're not going to go down that same robot uprising road today. Right. We're doing something way more interesting. Yeah. We are going deep on how people like you listening right now are actually shaping the future of AI like right now. Yeah. Specifically, large language models like Perplexity.

And we're looking at a really fascinating example of user-driven innovation in this deep dive. We're talking about a community of Perplexity users that basically came together and said, "Hey, what if we could make AI even smarter, more organized, more, uh..." Hold on, thoughtful AI. That's where I need you. Yeah. To break it down for me. Right, because when we think of AI, we think of algorithms, we think of data sets, not necessarily thinking. But this community actually developed a framework called Scratchpad think.

And it's basically like giving the AI a notepad to work through complex ideas before it gives you an answer.

Okay, so it's not just spitting out information, it's actually like processing it in a more structured way.

Yeah. Tell me more.

Exactly. Imagine the AI like jotting down notes, making connections, even questioning its own assumptions before it presents you with this like really well-reasoned response. It's like going from a quick Google search to like having a conversation with a really, really smart research assistant.

And this Scratchpad think thing, this isn't something Perplexity built in, right? This is coming from the users themselves.

That is what is so remarkable about this. This isn't some top-down directive from some big tech giant. This emerged organically from users that just wanted more from their AI interactions. They experimented, they collaborated, they essentially crowdsourced a better way to think with AI.

That's incredible. So we're not just talking about getting better answers here, we're talking about actually like influencing how AI evolves.

Precisely. And this is where it gets really, really interesting. This isn't just some abstract concept or anything. People are using Scratchpad think to like dive deep into complex topics and getting amazing results.

Okay, you've got to give me an example. Show me how this actually works.

Okay, so imagine you're fascinated by quantum physics, right? But you're no scientist. Yeah.

Obviously.

So you go to your AI and you're like, "Hey, explain quantum entanglement, but like make make sense."

Yeah, that's my usual request. Explain it to me like I'm five, but also like I have a PhD.

Exactly. So now with Scratchpad think, instead of just like bombarding you with a bunch of jargon and equations, the AI starts breaking it down.

Okay.

It might say, "Okay, to understand entanglement, let's first think about particles," and lays out some basic principles.

So it's like it's building a foundation instead of just throwing the textbook at me.

Exactly. But here's where it gets really cool. The AI using the Scratchpad think framework might then say, "Now, imagine those particles are connected in a way that," and then introduces this concept of entanglement in a more relatable way.

So it's not just explaining it, it's like guiding me through the thought process like a teacher almost.

Precisely. And as it's going, it might even say, "You know, this is where it gets really mind-bending," or "Let's consider a different analogy to really grasp this."

Wait, the AI is acknowledging the mind-bending nature of quantum physics? That's wild.

It is. And that's the power of this framework. It encourages the AI to be more transparent, more engaging, and ultimately, more helpful in guiding your understanding.

Okay, I'm starting to see why this is such a big deal. It's like it's changing the entire dynamic of how we interact with AI. But how are people actually using Scratchpad think? Are they literally typing this whole thing out every time they...

That's a great question. Some people are embedding it directly into their AI profile, so it's always running in the background.

Oh, interesting.

Others are creating these dedicated collections. Think of them like special workspaces where the Scratchpad think framework governs every interaction.

So it's like stepping into a room where AI is hardwired for deep thinking.

You got it. And the cool thing is they're constantly refining how they implement it, experimenting with different approaches to see what works best.

This is incredible, but I have to ask, what's the catch? This sounds almost too good to be true.

Well, there are always challenges with any new technology. So there are challenges. What are we talking about here? Like glitches, existential crises, what...

Well, nothing quite that dramatic. But this framework, remember, it's all about pushing the AI to think more deeply, more independently. So naturally, there are going to be some questions about, you know, how to guide that process responsibly.

So, how do we make sure the AI doesn't go off the rails, so to speak?

That's what's so interesting about this community. They're not shying away from these questions at all. They're like, "No, let's confront this head-on."

I like that.

Yeah, they're constantly refining that Scratchpad think framework. Finding ways to balance the AI's ability to think independently with the need for human guidance.

So it's like a constant dance between pushing the boundaries, but making sure things don't get out of hand.

Exactly. They're experimenting with different safeguards, different ways to ensure that the AI's reasoning stays aligned with human values and objectives.

Oh.

So work in progress, but the collaborative spirit of this community is really inspiring.

And to me, that's what's so fascinating. It's not about making AI smarter just for the sake of it, it's about doing it in a way that benefits everyone.

Precisely. And that's what this Scratchpad think framework really shows: the power of collective intelligence. When people come together with a shared goal, amazing things can happen.

It's like they've unlocked this whole new level of collaboration, not just between people, but between people and AI.

Exactly. And who knows what incredible discoveries they'll make next. This is just the beginning.

This deep dive has really opened my eyes to the possibilities here. So to sum it up, for everyone listening out there, we've gone from "AI is going to take over the world" to "AI is taking notes and asking insightful questions."

It's a pretty good direction, right? I like it a lot better.

It's a powerful reminder that we all have a role to play in shaping the future of AI. So next time you're interacting with an AI, don't be afraid to ask it to think out loud, to show its work. You might be surprised by what you discover. Until next time, keep exploring, keep asking those big questions, and keep diving deep.