The Strengths and Weaknesses of Notion
The challenge with Notion is that once a page is created, the product almost immediately asks you to decide where it belongs in the larger system. What is it? Where should it live? Is it tied to a specific task or Project?
You can always start from a blank page and write freely, but Notion's power comes with a kind of constraint. It keeps pulling you toward system design, which is why you often feel the pressure of having to place information properly.
That pressure is also where Notion's value comes from. Its notes have clearer states, relations, permissions, and operational boundaries. This structured way of organizing information makes it easier to turn goals into action and knowledge into reusable objects, and it can also become a shared standard in collaborative work.
That is why you can build a CRM, a Project board, or a content calendar in Notion with relative ease. When Notion says anyone can build their own software, what it really means is arranging and reusing the same structured data in different ways.
From the results alone, Notion has truly delivered on its positioning as "LEGO for Software." At least in the current market, I do not think there is another product that has done it at this level.
The Strengths and Weaknesses of Obsidian
What makes Obsidian comfortable is that a note can begin as just a note. Recording can come before categorization, and ownership can come before systems. It does not ask you to decide from the beginning whether a piece of information is a Task, a Project, a reference, or an outcome.
A simple folder structure is enough to support most of the narrative around classification. The cost of creating forward links and backlinks is low. A rough location plus a few tags is enough to bring something into your system, which is why Obsidian is better suited to long-term reading, idea incubation, and research writing.
Ideas and inspiration need uncertainty. They need ambiguity, repeated naming, and the possibility of being reinterpreted later. That is the atmosphere Obsidian creates. It also guarantees that your text is not locked inside its own app, which means you can always expect that the same text may take on new forms in the future.
But that is also where Obsidian becomes demanding. A system built on Markdown files can only rely on conventions inside text to represent or simulate structured relationships. Once you start using Properties, building database-like views with Bases, or changing interaction logic through custom themes, Obsidian enters a different kind of maintenance spiral.
The introduction of community plugins makes complexity grow in a far less orderly way. The platform hands over the right and responsibility of organizing information to individuals and the ecosystem. Even plugin security can never be fully reviewed. That is both freedom and cost. You may think you fully own your local files, while still failing to fully grasp the complex relationships that grow around the system.
Obsidian, in the end, is also stacking more and more layers on top of simple Markdown. Incompatible syntax becomes part of the experience too.
At this point, the "atmosphere" created by a product still matters. It shapes whether each input you make leans toward structured expression or toward a freer act of pouring thoughts onto a blank page.
That is why continuing to compare which product can build a more powerful system, which has better relations, which has a richer ecosystem, which is easier to migrate from, or which can call stronger models may not be that meaningful. They serve different people, and both have strengths while borrowing from each other. Over-focusing on one feature tends to produce a shallow conclusion.
When AI Starts Acting for You
AI will gradually flatten the difference between the two products in surface-level capabilities such as writing, search, summarization, and documentation. But once AI stops being just a question-and-answer interface and starts editing, creating, and updating for you, the bigger difference comes from how each system defines operational objects and how each AI action is bounded.
A Task object with a clear status, priority, owner, and due date can be filtered, updated, or reassigned precisely by an agent. A plain text block that only contains a #task tag gives the agent much less structure. In that case, what it can do is closer to search and fuzzy reasoning, and the result of automation will not be the same.
That is why I think we can reach a slightly ambiguous conclusion here. Notion's AI is better at achieving outcomes. Obsidian's AI is better at resonating with the way you think. At least for now.
So when AI becomes the main force behind productivity, do the note systems we once carefully designed to improve retrieval efficiency still matter as much? When you need an idea again in the future, the quality of a note regenerated by AI may not be much worse than a Markdown file you wrote ten years ago.
The freedom of will we used to value, the real thinking we wrote into notes, and the sense that writing helps us know ourselves better may all lose part of their weight in the face of emergent intelligence. I am honestly pessimistic about that.
If AI becomes strong enough, the long-term readability that Obsidian treasures and the structured reusability that Notion pursues will both have to be re-evaluated.
The Answer Still Matters Right Now
Fortunately, at least today, AI is still not strong enough to make information architecture completely unnecessary. Your note system is still the infrastructure for working with AI, which is why the question of which tool to choose still deserves serious thought.
Under a pace where AI models see a small update every two weeks and a larger one every couple of months, you do not need a decision that will remain valid for ten years. It is enough to think clearly about the problem you want to solve right now, or within the next year.
In other words, if a piece of information needs to move quickly into an actual workflow and collaboration context, Notion has a stronger advantage. AI can also produce more precise execution results there. The risk is premature definition. You may classify, add fields, and build processes before the idea has matured, and then end up maintaining the system for the sake of maintenance. A poorly built Notion system only creates deeper paths and more polished confusion.
If a piece of information matters more as something to preserve, revisit, incubate, and rediscover later, Obsidian fits that need better. AI may even give you more unexpected moments of insight. The risk is delayed organization. Structured information still has to be constrained through text and third-party plugins, and the freedom you gain early may return later as naming chaos, loose links, and broken plugins.
With AI context windows still mostly sitting in the 200K to 1M range, you still cannot expect AI to fully understand your system.
At least for now, the more you want AI to participate in execution, the more you need to define identities, states, relations, and permissions in advance. It is similar to how OpenClaw depends on carefully maintained SOUL, AGENT, USER, and MEMORY documents. Longer context still cannot replace information structure itself.
The Divide and the Choice
You can imagine two kinds of people here. One is a Project-driven worker who is juggling three to five things at once, maybe a freelancer, a small business owner, or someone leading a team.
Every week, this person moves back and forth between client needs, content schedules, payment tracking, asset libraries, and team communication. Their biggest fear is that something gets missed, a status is not updated, or a commitment is no longer tracked. What they need is for information to gain a clear origin and destination quickly, so their attention can be freed from the burden of remembering everything.
The other person is a creator writing a book or doing long-term research. They read dozens of pages every day and come up with dozens of associations. Most of them are not immediately useful, but some may resonate with another piece of material late at night two years later. What they fear most is that an idea still taking shape gets classified too early and then disappears for good.
In my view, the former is better suited to Notion, and the latter to Obsidian.
Notion believes that information gains value through what can be coordinated and executed after it enters a system. That is why it asks to name and classify information as early as possible.
Obsidian believes that information gains value through how it can continue to exist even outside a specific system. That is why it prefers to let information remain as text first.
So when you choose between Notion and Obsidian, the real question is this: with AI now in the picture, do you want knowledge to become part of an action system first, or remain as text that can be reorganized and reinterpreted later? One day you may no longer need to choose between the two, but for now you still need to decide at which stage you are willing to pay the cost of organizing information.
Are you willing to take on the pressure of defining information early so that you get clearer reuse, collaboration, and execution later? Or are you willing to accept the delayed confusion of naming, organizing, and maintaining so that your early experience remains lighter and more associative?
What you lack more right now, freedom with less early organization, or order that pushes ideas into action more quickly?
The worlds both tools can build are too large, and the imagination they offer is too attractive. That is why it is hard to "have it all." If you really have to choose one, your answer may already be hidden in your instinctive response to the question above.
That answer probably will not remain valid for long. Still, that is fine. There is no need to think too far ahead. You can even stop thinking about it entirely. Give AI a little more time, and perhaps the question itself will stop needing an answer.
📘 FLO.W System
If this article leaves you leaning toward Notion, you can also take a look at the FLO.W system I have refined over time. It will not think for you, but it can save you from rebuilding the same structure again and again.
