Getting Super-Powers with AI: Feedback, Big Questions, and a New Take on Notes

Why Asking for Feedback from GPT Has Helped Me Create Better Prompts

I’ve spent quite a bit of time collaborating with GPT on various tasks, from brainstorming ideas to solving technical problems. Over time, I realized something: how I ask questions directly affects the quality of the responses I get. It wasn’t just about getting the right answer, it was about asking the right questions.

That’s when I started using a technique that’s made a real difference for me: asking GPT for feedback on my prompts after each session. This simple habit has helped me improve my ability to craft more precise, effective questions, and it’s one I recommend to anyone who uses AI for complex problem-solving or creative work.

?

Why Prompts Matter So Much

When I first started working with GPT, I assumed that if I got the general idea across, the model would take care of the rest. But I quickly learned that vague or broad prompts often lead to answers that are equally vague or off-base. If I asked for “more detail” without specifying a focus, I’d get all kinds of information that didn’t quite help.

By taking the time to ask GPT, "How could I have phrased that better?" or "What part of my request was unclear?" I began to see patterns in the way I was asking questions. For example, sometimes I was being too broad, and other times I wasn’t providing enough context.

Learning Through Feedback

What I love about this approach is that GPT can provide immediate, actionable feedback. After a session, I’ll review how the conversation went and ask GPT, “Was there a better way to phrase that last prompt?” This has allowed me to see where I was getting things right and where my prompts could be refined.

Over time, this has made a noticeable difference in the quality of our exchanges. I’ve learned to be more specific when necessary, and more open-ended when exploring ideas. As a result, I spend less time clarifying or rephrasing, and more time digging into the solutions and insights that really matter.

Making It a Habit

I’ve made it a habit to ask for feedback after every significant interaction with GPT. It doesn’t take much time, but the long-term benefits are huge. Every session is an opportunity to gain experience about how I communicate, not just with AI, but in general. It’s like having a personal coach for my thinking process, helping me refine how I approach problems and break down questions.

Confidence and Creativity

I’ve noticed that as I’ve gotten better at crafting prompts, my confidence in using GPT has grown. I now approach sessions with a clearer idea of how to structure my questions, which opens more creative possibilities. Rather than feeling like I need to carefully tiptoe around a topic, I can dive in knowing that if the first prompt doesn’t hit the mark, GPT will help me adjust.

I’ve also started using this feedback technique to experiment with more complex queries. The back-and-forth dialogue helps me stretch my thinking and try out novel approaches without worrying about wasting time.

Final Thoughts: Continuous Improvement

Asking GPT for feedback has become an essential part of how I use AI. It’s a simple way to improve how I communicate, leading to more productive conversations and better outcomes. The more I refine my prompts, the more I can tap into the full potential of AI, whether I’m solving a business problem or exploring creative ideas.

If you’re using GPT regularly, I highly recommend giving this technique a try. You’ll be surprised at how quickly it helps you sharpen your questions, and ultimately, get the answers you need more efficiently.

?

Solving Complex Business Problems with ChatGPT and the o1-Preview Model

At the recent European Fabric Conference in Stockholm, Sweden, I had the pleasure of attending a session where a user posed an intriguing question to their Copilot: “How do I increase my profits by 20%?” It’s the type of question that immediately grabs attention, not because of its ambition, but because of its complexity. Increasing profits by a specific percentage requires consideration of numerous variables, from cost management to market opportunities, and it's not something a single, simple answer can address.

This is where the capabilities of the o1-preview model in ChatGPT shine. This advanced version of GPT is designed for deeper reasoning, handling layered queries, and working through complex business challenges step by step. I’ve been running a parallel session with ChatGPT to explore this exact query, and it’s a powerful example of how the o1-preview model can guide decision-makers through nuanced problem-solving.

Why This Matters for Complex Queries

With traditional AI models, you might get a quick suggestion, perhaps to cut costs or boost sales. However, these surface-level solutions often miss the mark when it comes to sustainable growth. The o1-preview model, on the other hand, approaches such a query with a more thoughtful and strategic breakdown. By dissecting different profit levers, such as optimizing pricing strategies, improving operational efficiencies, and exploring new revenue streams, it builds a comprehensive framework that empowers the user to make informed decisions.

For example, in my session, ChatGPT started by asking clarifying questions about the company's industry, market conditions, and current operational metrics. It then proceeded to suggest various strategic avenues, each tailored to these specifics, such as re-evaluating customer acquisition costs, exploring new geographic markets, or introducing subscription-based pricing models. Other considerations I added include lead times, resource availability, and sales cycle lengths, all of which must be fed to the model since it has no knowledge of these things. Overall, the model still a bit too optimistic with how quickly humans can execute but it's great start.

The Power of Iterative Collaboration

What makes the o1-preview model especially valuable is its ability to engage in iterative collaboration. Rather than providing one definitive solution, it encourages dialogue, suggesting ideas, asking for further details, and adjusting its recommendations based on the user's responses. This is precisely the kind of interaction needed when dealing with a multifaceted problem like increasing profits by 20%.

Ultimately, the beauty of the o1-preview model lies in its ability to facilitate deep, meaningful conversations around complex business objectives. It’s not about giving a quick fix, it’s about empowering users to think critically and strategically, enabling them to make decisions that align with long-term goals. If you're tackling sophisticated questions like the one posed at the European Fabric Conference, the o1-preview model is an ideal partner for working through those challenges in a structured, insightful way.

Discovering Google’s NotebookLM: How AI is Transforming My Approach to Notetaking

As someone who's always juggling information, I recently stumbled upon NotebookLM, Google’s experimental AI-powered notebook, and I have to say, it’s impressive. We all know how traditional note-taking apps work: they’re great for storing information, but they can sometimes feel like static repositories. NotebookLM takes that concept and completely flips it. It turns your notes into something interactive, almost like having a personal assistant built right into your documents.

What Exactly is NotebookLM?

Think of it as an AI-powered companion for your notes. Unlike typical apps where you must dig through pages and pages of content, NotebookLM helps you work through your documents by answering questions based on what you’ve uploaded. This makes it far more useful for getting real insights from your notes, rather than just filing things away.

Features I’ve Found Useful

1. Interactive Q&A: You can literally ask it questions about your notes and get back relevant, contextual answers. This is perfect when you’re trying to recall a specific detail or concept without having to read through everything.
2. Summarization: One of my favorite features is how it condenses long documents into summaries. This feature alone has saved me tons of time, especially when I'm skimming for key points.
3. It Actually Understands Context: I’ve noticed that it doesn’t just pull keywords, it gets the underlying ideas. That means I can ask broader questions and still get meaningful responses. This makes a dramatic difference when dealing with complex topics.
4. Audio summaries: One feature of the notebook is the ability to turn your notes into a podcast. I’ve done that here with this newsletter. Check out the summary of this newsletter here: https://notebooklm.google.com/notebook/aeb348d7-41cd-4c44-8cdc-3762a4db6af6/audio

How It Works

The process is straightforward. You upload your notes, and the AI processes them, then you can start asking it questions or request summaries. What’s great is that it’s not just spitting out generic responses, it’s pulling directly from the material you’ve given it. So, whether you’re working on research, business notes, or even brainstorming for a project, the interaction feels tailored to what you need.

Why I Think Google’s NotebookLM is Impressive

• Learning Becomes Easier: It’s like having a study partner that can instantly answer your questions or clarify things you might have missed. For students or anyone working with a lot of information, this tool really ups the game.
• Efficiency Boost: Summarizing large documents or finding specific pieces of info no longer takes forever. NotebookLM automates that part, so you can spend more time on the actual decision-making or analysis.
• Context-Specific Insights: I’ve also noticed that the more content I feed into it, the more useful it becomes. It's almost like it’s adapting to my workflow, which is a refreshing change from traditional note apps.

Use Cases I’ve Tried

• Research Projects: It’s been a lifesaver when I’m working through lengthy articles or papers. Instead of reading line-by-line, I upload everything into NotebookLM and just ask for a summary or specific insights. The time saved is incredible.
• Meeting Notes: I’ve started uploading notes from meetings and projects, and the ability to quickly reference them through simple questions has made follow-ups so much easier.
• Creative Brainstorming: Even for creative tasks, the AI-generated summaries and context make it easier to organize and refine ideas.

Privacy Matters

I know privacy can be a concern with AI, but Google has made it clear that the documents you upload are used strictly to provide personalized help. They’re not used to train the model, and all your data stays private, which gives me peace of mind.

Getting Started is Simple

It’s still experimental, but if you’re curious, you can sign up to try NotebookLM. It’s easy to get started, just upload a few documents and dive in. I found the learning curve practically nonexistent.

Final Thoughts

NotebookLM has quickly become one of the most powerful tools in my workflow. It’s not just another place to store notes, it actively helps you engage with your content. Whether you’re a student, researcher, or just someone trying to get more out of your notes, it’s worth exploring. I’m excited to see how this tool evolves as more people start using it. If you follow Jason Spielman , you can see from his recent post that there's a lot more to come.

I hope you’ve enjoyed this edition of Grounding AI. Please share this newsletter with others who may be interested in this type of content, via this link https://www.dhirubhai.net/build-relation/newsletter-follow?entityUrn=7188273253503664128 .

?