🤝 Why your learning habit is actually hurting you
Dear curious mind,
Today's insight challenges everything we've been taught about learning and personal development. In a world where AI can deliver personalized, contextual information instantly, the game has fundamentally changed. The winners aren't those who know the most, but those who act most effectively.
Ready to break free from information addiction and start making real progress? Let's dive in.
In this issue:
💡 Shared Insight
Just-in-Time Knowledge: Why Now is the Perfect Time to Stop Hoarding Information
📰 AI Update
ERNIE 4.5: Baidu's Open Model Family That Runs Everywhere
Qwen VLo: Alibaba's Closed AI Model with Image Editing Capabilities
FLUX.1 Kontext [dev]: Black Forest Labs' Open Image Editing Model
🌟 Media Recommendation
Book: "Lean Learning" by Pat Flynn To Break Your Information Addiction
💡 Shared Insight
Just-in-Time Knowledge: Why Now is the Perfect Time to Stop Hoarding Information
You know that feeling when you bookmark an article "for later" and it joins the digital graveyard of thousands of other saved links you'll never revisit? Or when you buy another course promising to teach you everything about a topic, only to watch it collect digital dust while you search for the "perfect" next resource?
I've been there. We all have. But what if I told you that this entire approach to knowledge, the hoarding, the preparing, the endless researching, is not just outdated, but actively holding you back?
For decades, we've been conditioned to gather and store information "just in case." We'd read dozens of books, bookmark hundreds of articles, and attend countless webinars to prepare for some future moment when we might need that knowledge. This approach made sense when information was scarce and difficult to access.
The internet began changing this dynamic by making vast amounts of information available at our fingertips. But we still fall into the trap of information hoarding, creating digital libraries of PDFs we'd never read again and saving bookmarks we'd never revisit.
Today, AI is fundamentally reshaping how we should think about knowledge and learning. We're entering the era of just-in-time knowledge, and it's the perfect time to embrace an action-driven mindset.
This insight was sparked by reading Pat Flynn's "Lean Learning" (highlighted in this issue's media recommendations), which advocates for focusing on learning only what you need to take the next step forward. It got me thinking about how AI is fundamentally changing not just how we access information, but how we should approach personal knowledge management entirely.
The Shift from "Just in Case" to "Just in Time"
AI assistants like ChatGPT, Claude, and others have become incredibly sophisticated at providing personalized, contextual information exactly when you need it. Need to understand a complex coding concept? Ask AI to explain it in your preferred programming language with examples relevant to your project. Want to learn about marketing strategies? Get customized advice based on your specific industry and target audience.
This isn't just about having a search engine that talks back. AI can:
Adapt explanations to your level of expertise
Provide step-by-step guidance tailored to your specific situation
Answer follow-up questions as you work through problems
Help you apply concepts immediately rather than just understand them theoretically
Action Over Accumulation
The most successful people in this new landscape aren't those who consume the most content, but those who apply information most effectively. Instead of spending weeks researching the "perfect" approach, you can:
Define your immediate goal clearly
Get just enough information to take the first step
Take action and learn through doing
Iterate and improve based on real feedback
This approach isn't about being lazy or cutting corners. It's about recognizing that in a world where personalized information is instantly available, the bottleneck isn't access to knowledge, it's execution.
Your AI-Powered Learning Partner
Think of AI as your personal tutor who's available 24/7, never gets tired of your questions, and can adapt to exactly how you learn best. Whether you're learning to code, starting a business, or mastering a new hobby, you can get targeted guidance precisely when you need it.
This doesn't mean abandoning critical thinking. Instead, it means being strategic about when and how you acquire knowledge. Save your mental energy for the work that matters: applying what you learn, creating something new, and making progress toward your goals.
Rethinking Personal Knowledge Management
This shift has significant implications for how we approach personal knowledge management (PKM). Traditional PKM systems required significant effort at carefully organizing notes, creating elaborate tagging systems, and maintaining our digital gardens. But what if we flipped this approach entirely?
Instead of actively hoarding and organizing information, we can shift to passively collecting our experiences and interactions. Every book highlight, every article we read, every conversation we have, and every project we work on becomes data points that paint a picture of how we think, what we value, and how we learn best.
Your notes and digital traces aren't just information repositories anymore, they're training data for your personal AI assistant. When you ask AI to explain a complex concept, it can draw from your collection of highlights and notes to understand that you prefer concrete examples over abstract theories, or that you learn better through visual metaphors than technical jargon.
This transforms PKM from an effortful practice of information curation into a passive background process of experience collection. Your AI assistant becomes more valuable because it has access to more information to understand you better.
The Bottom Line
We're living in a unique moment in history where personalized, contextual knowledge is more accessible than ever before. The competitive advantage no longer goes to those who know the most, it goes to those who act most effectively.
Stop trying to prepare for every possible scenario. Start defining what you want to achieve, get the information you need to take the next step, and then actually take it. Let your digital experiences accumulate naturally in the background, creating a rich context for your AI companion to understand how you think and learn.
The future belongs to the action-takers who leverage AI that truly understands them, not the information accumulators with perfectly organized but unused knowledge bases.
📰 AI Update
ERNIE 4.5: Baidu's Open Model Family That Runs Everywhere [Ernie blog]
The open-weight model landscape just got a major boost with Baidu's open release of ERNIE 4.5, a comprehensive family of large language models which are competing with the current state-of-the-art models. ERNIE 4.5 covers three different usage scenarios:
On-device models with just 300 million parameters that can run directly on your phone, bringing AI capabilities to mobile devices without requiring internet connectivity
Local GPU-friendly models under 30 billion parameters that enthusiasts and developers can run on consumer hardware, making advanced AI accessible to individual researchers and smaller companies
Enterprise models with up to 424 billion total parameters that run on powerful GPU servers and compete with the best open-weight models like DeepSeek-V3, and in some scenarios even challenge proprietary giants like GPT-4
The models are created with a Mixture-of-Experts (MoE) architecture and support either text only or text and visual understanding. The latter also support thinking and non-thinking mode.
Perhaps most importantly, in contrast to previous ERNIE models, all ERNIE 4.5 models are released under the Apache 2.0 license, meaning they're completely free for commercial use. This represents a significant contribution to the open AI ecosystem, giving developers and researchers powerful alternatives to proprietary models. To get a first glimpse of the performance, take a look at the benchmark results shared in the release blog article. So far, you need the deployment toolkit FastDeploy to run the models.
Qwen VLo: Alibaba's Closed AI Model with Image Editing Capabilities [Qwen blog]
Qwen VLo represents a significant advancement in multimodal AI, offering both understanding and generation capabilities.
This model rivals the image editing power from gpt-image-1 in ChatGPT as it
enables users to generate, edit, and refine high-quality visual content from image and text inputs. Furthermore, the model provides an enhanced precision in understanding image inputs.
Unfortunately, in contrast to previous Qwen models, Qwen VLo has not (yet) been released with open weights, which may limit its accessibility and adoption within the developer community.
FLUX.1 Kontext [dev]: Black Forest Labs' Open Image Editing Model [Flux Blog]
Initially, the FLUX.1 Kontext model was available solely through Black Forest Labs' official API, as covered in an earlier issue.
Now, Black Forest Labs has also released a smaller, 12B parameter open-weight version for consumer hardware. Despite its reduced size, the model retains impressive inference results, making it the first capable generative image editing model available as open-weights.
🌟 Media Recommendation
Book: "Lean Learning" by Pat Flynn To Break Your Information Addiction
In our current world, we're drowning in information. Never did a book make me consider stopping reading non-fiction books. Pat Flynn's "Lean Learning" challenges my approach to media intake in general, not only books, but also reading articles, watching YouTube videos and listening to podcasts. But before I share my key takeaways from the book, I want to let you know that I liked the book so much that I finished it despite its core message made me question if this was the right thing to do.
Flynn argues that the internet transformed information from a scarce commodity into an abundant resource, and AI has made it even more accessible. Yet, many of us still operate under the old paradigm and hoard knowledge "just in case" instead of learning "just in time" to solve specific problems.
Key Takeaways
The ITWEWWILL Question: Flynn applies Tim Ferriss's powerful question: "If this were easy, what would it look like?" This becomes your go-to tool for breaking down complex tasks into manageable first steps.
Action Over Information: The book's central message "Information is silver, action is gold" directly challenges our tendency to endlessly research capabilities without actually using them to solve real problems.
Just-in-Time Learning: Instead of comprehensive courses, learn specific techniques when you need them. Today's AI assistants like ChatGPT make this approach more practical than ever.
My take: This felt like the entrepreneurial guide I didn't know I needed. Flynn's approach aligns perfectly with how AI can supercharge just-in-time learning, making his framework even more powerful today, with AI being able to provide you tailored information based on your specific requests. If you've ever felt paralyzed by the endless stream of information and possibilities, this book might help you to start taking meaningful action. It really did make me question my media consumption and with that inspired the shared insight of this issue.
Disclaimer: This newsletter is written with the aid of AI. I use AI as an assistant to generate and optimize the text. However, the amount of AI used varies depending on the topic and the content. I always curate and edit the text myself to ensure quality and accuracy. The opinions and views expressed in this newsletter are my own and do not necessarily reflect those of the sources or the AI models.