How do you see AI transforming the way we explore new knowledge?

@wiktor_wysocki This reminds me of the AI in the movie Her. The idea of treating AI as conversational partners opens up fascinating applications. ChatGPT, for example, has made great strides in language and voice modeling, creating natural and engaging interactions. I’ve already started using it this way—for instance, while walking to the subway, I might brainstorm ideas or draft an email through a quick conversation. That said, I feel the need for pure AI conversation might not be as strong as it seems. After all, we can talk to friends or colleagues, and only a small number of people might prefer chatting with an AI over humans. However, apps like Speak, which use conversational AI to help people practice English, show that when conversation is tied to a clear purpose—like learning a new language—it becomes much more valuable. By creating life-like interactions, Speak provides a space for people to build confidence and skills in a natural, low-pressure way. This makes me wonder if the real potential of AI lies beyond casual chats. Perhaps it’s about helping us transform scattered information into structured, actionable knowledge in our brain. Imagine a process where AI guides us to internalize information effortlessly, shaping new perspectives that we can apply in daily life. This shift—from simply conversing to actively building understanding—feels like the next meaningful step for AI.

@timliao I agree. AI as a tool has to mature so people stop having to think about using it. This is the case with social media right now; you do not know when you open Instagram (unfortunately). But yes, even then, there could just not be a need to have another colleague to talk to, so maybe your more knowledge structuring and purposeful usage idea is more real.

@wiktor_wysocki Thanks for your reply, Wiktor! The example of Instagram is spot on. I’ve always felt that AI applications are still in their early days. At their core, most AI tools right now are chat-based because large language models themselves are fundamentally designed as input-output language systems. But I fully agree with your point—just like how you don’t consciously think about opening Instagram, AI needs to evolve to that level of seamless and purposeful integration in our lives. It’s exciting to think about what’s next! I’m also really glad to have this opportunity for such a high-quality discussion. I’d love to sincerely invite you to try our latest project, Minduck Discovery. It’s a tool that combines visual thinking， search engine and AI to help people structure information and internalize it into their own knowledge. I think you might find it interesting!

@felix_sattler Hi Felix, thanks for sharing! I found your take on AI improving efficiency through quick answers very interesting. However, my question was more about whether this speed might reduce opportunities for deeper understanding and how AI could inspire curiosity and exploration instead of just delivering answers. Do you think tools like this risk oversimplifying knowledge and discouraging exploration? Or could they be designed to balance efficiency with encouraging users to think and connect ideas? By the way, I noticed the link you shared doesn’t give much of an introduction to the product itself. A landing page with an overview might help more people understand your product at a glance. Would love to hear your thoughts!

@timliao I don’t think breaking knowledge into smaller, more digestible pieces is oversimplifying – it’s just a more focused way of learning. Instead of wading through an entire PDF where only one sentence actually answers your question, you get straight to what you need to know. Plus, the chat-based approach, which our tool and others use, makes exploring content feel more like having a good conversation with an expert. Overall, I’d say these solutions are less about discouraging exploration and more about focusing on exploring what users actually want to know.

@felix_sattler Got it. So, if I understand correctly, your AI solution helps users quickly find the exact information they need from a sea of data, while keeping them focused and avoiding unnecessary distractions. That’s really useful. For example, beyond your PDF case, I often open tons of browser tabs during research, each with important content. It can get frustrating switching back and forth before finally landing on the right one. In those moments, focus is everything! I’m really curious though—after users find the key info they need, does your solution offer any features to help them build their own knowledge system? I think that idea is fascinating because turning information into actual knowledge feels like the next big step for long-term value. Would love to hear more about how you see this!

@meduard You’re absolutely right. In the past, we had to find clues and piece them together before reaching a conclusion, which created a process of personal discovery. The knowledge we gained came with our own experiential context, and when we recall it, there’s always a clear mental pathway. Nowadays, many AI-driven products provide "one-click" solutions to get usable answers. While this is undoubtedly progress, if we rely solely on such tools for the long term, our ability to think independently might diminish. It could lead to a growing dependency on external systems for answers, limiting our cognitive development. I believe the key lies in ensuring that learning involves a process of connecting ideas and understanding relationships. This allows our brains to internalize external information and ultimately enables deep learning.