I clicked my first hyperlink in 1998.
It was blue, underlined, and utterly magical. I was reading about dinosaurs on a webpage when I clicked the words “Jurassic Period” and suddenly I was on a completely different page about geological time. Then I clicked “continental drift” and found myself reading about plate tectonics. Within ten minutes, I’d gone from dinosaurs to earthquakes to volcanoes to the formation of the moon.
My mind was blown. Not just by the information, but by the way information could connect. Books had footnotes, but you couldn’t instantly jump to the referenced work. Encyclopedias had “see also” references, but you had to walk to another volume. Hyperlinks made knowledge instantaneously traversable.
That experience planted a seed that took decades to fully understand: hyperlinks weren’t just a navigation mechanism. They were a new way of thinking.
Fast forward to building stashed.in, and I’ve spent years thinking deeply about what hyperlinks really are. Not just technically (URLs pointing to resources), but conceptually. What do they represent? How do they change us? Where are they going?
The answer surprised me: hyperlinks have evolved from simple document connectors into something resembling external memory. And understanding this evolution changes how we should think about organizing the web.
The Original Vision: Vannevar Bush’s Memex#
Before we had hyperlinks, we had dreams of them.
In 1945, Vannevar Bush published an essay called “As We May Think” in The Atlantic. He described a theoretical device called the Memex: a desk-sized machine that could store all your books, records, and communications on microfilm, with the ability to create “associative trails” between documents.
Bush wrote:
“The human mind… operates by association. With one item in its grasp, it snaps instantly to the next that is suggested by the association of thoughts, in accordance with some intricate web of trails carried by the cells of the brain.”
He imagined a machine that worked the same way. You’d create links between documents based on your associations. These “trails” would be personal, representing your unique way of thinking about connections between ideas.
The Memex was never built. But the vision profoundly influenced the people who would eventually create hypertext and the World Wide Web.
Ted Nelson and the Dream of Hypertext#
In the 1960s, Ted Nelson coined the term “hypertext” and spent decades working on Project Xanadu, attempting to create a universal hypertext system.
Nelson’s vision was more ambitious than what we got with the web. He imagined:
Two-way links. When you link to someone’s content, they know about it automatically. The link goes both directions.
Transclusion. You could include parts of other documents in your own, with the original content staying in place. Changes to the source would appear everywhere it was transcluded.
Version tracking. Every version of every document would be preserved. Links would never break because content never disappeared.
Micropayments. When you linked to someone’s content, they’d receive tiny automatic payments. This would fund quality content creation.
Xanadu never launched. It became a legendary example of ambitious vaporware. But Nelson’s ideas influenced everyone thinking about hypertext.
Tim Berners-Lee and the Practical Web#
In 1989, Tim Berners-Lee took a simpler approach. He wanted to solve a practical problem: physicists at CERN needed to share documents and reference each other’s work.
His solution became HTML and the World Wide Web. It had hyperlinks, but they were deliberately simple:
One-way links. You could link to anyone without their permission or knowledge. This made the web possible but lost Nelson’s two-way dream.
Brittle links. If the target document moved or disappeared, your link broke. No version tracking. No permanence guarantees.
No built-in attribution. Links pointed to content but didn’t automatically credit or compensate creators.
Berners-Lee’s simplicity was the web’s superpower. Because linking was so easy, people did it prolifically. The web grew from thousands of pages to billions because the barrier to creating connections was nearly zero.
But we lost something in that simplicity. Nelson’s vision of a rich, bi-directional, permanent hypertext system gave way to a messier, more fragile network.
How Hyperlinks Changed How We Think#
The web’s hyperlinks weren’t just a new technology. They fundamentally altered human cognition in ways we’re still understanding.
From Linear to Networked Thinking#
Books are linear. You read from beginning to end. Even with footnotes and indexes, the primary mode is sequential.
The web is networked. There’s no “beginning” or “end.” Every page is potentially an entry point. Every link is a fork in the road. You can arrive anywhere from anywhere.
Research on hypertext reading shows this changes how we process information. We scan. We jump. We follow interesting threads. We rarely read deeply or sequentially.
This isn’t necessarily bad. It’s different. We’ve adapted to navigating information spaces rather than consuming linear narratives. Our thinking has become more associative, more exploratory, more networked.
The Externalization of Memory#
Before the web, knowledge lived primarily in our heads and in physical books. You remembered what you could remember. You referenced books when you needed more.
Hyperlinks enabled something new: externalized, networked memory.
When you know where information lives (which website, which page), you don’t need to remember the information itself. You remember the path to it. Or you remember that it exists and search for it when needed.
This is what psychologists call “transactive memory”, a concept usually applied to how couples and teams share memory responsibilities. Instead of everyone remembering everything, different people remember different things and know whom to ask about what.
The web made this massively scalable. You don’t need to remember facts. You need to remember where facts can be found. The web becomes shared memory for humanity.
The Collapse of Context#
But networked memory has a problem: context collapse.
When you read a book, you understand where you are. Chapter 3 comes after Chapter 2. The argument builds. Context accumulates.
On the web, you can land anywhere. That page you found via search? You’re missing the context that came before. The assumptions the author made. The argument they were building.
Links make this worse. When you follow a link into the middle of something, you’re perpetually arriving without context. You adapt by skimming for the relevant part, often missing nuance.
This is why “link rot” isn’t just an inconvenience. When links break, we lose more than just a connection. We lose context, attribution, and the ability to verify information.
Discovery Through Serendipity#
But links also enabled magical serendipity.
When information was locked in books, discovery required knowing which book to open. The library classification system helped, but you needed to know roughly where to look.
On the web, links create unexpected pathways. That article about web design links to an essay about typography, which mentions a book about the history of printing, which references medieval manuscripts. You started researching buttons and ended up fascinated by illuminated texts.
This serendipitous discovery is a feature, not a bug. Some of the best learning happens when you stumble onto something you didn’t know you needed to know.
The Social Layer: Links as Conversation#
By the 2000s, links had evolved beyond simple navigation. They became social signals.
Blogrolls and Link Culture#
Early blogs had “blogrolls”—lists of other blogs the author read and recommended. These weren’t just navigation. They were endorsements, community building, and taste signaling.
Linking to someone’s post wasn’t just citing them. It was starting a conversation. Blog comments often began “I saw you linked to my post about X…” Links were social gestures.
This created a virtuous cycle. Bloggers linked to interesting content. Those links brought readers. Those readers sometimes linked back. The web grew through mutual discovery and recommendation.
PageRank and the Link Economy#
Then Google changed everything with PageRank. The number and quality of links pointing to a page became a measure of authority. Links weren’t just navigation or social signals. They were currency.
This had good and bad effects:
Good: High-quality content got discovered because others linked to it. The best resources naturally rose to prominence.
Bad: People started gaming links. Link farms, reciprocal linking schemes, paid links. The link economy became corrupted.
Google kept updating its algorithm to counter manipulation, but the fundamental shift happened: links became a metric, not just a tool.
Social Media and the Death of Linking Culture#
Then social media arrived and largely killed linking culture.
On Twitter, character limits made rich discussion hard. On Facebook, the algorithm prioritized native content over external links. On Instagram, links barely worked at all.
People stopped writing blog posts responding to other blog posts. They started writing tweet threads that couldn’t link conversationally. The web became more siloed, less interconnected.
We still have links, but the culture of linking, responding, and building on each other’s ideas has diminished. Social platforms didn’t want you leaving. They wanted you scrolling forever.
The Personal Knowledge Revolution#
But recently, something interesting happened. While social media killed public linking culture, personal knowledge management exploded.
Roam Research and Bi-Directional Links#
In 2019, Roam Research popularized bi-directional links for personal notes. Finally, decades after Ted Nelson’s vision, we had working two-way links.
When you linked note A to note B, both notes knew about the connection. You could see everything that linked to any given note. This created graph visualizations showing how your ideas connected.
This wasn’t new technology. Wikis had backlinks for years. But Roam made them central to the experience, and the personal knowledge management (PKM) community exploded.
People started building “second brains” and “personal knowledge graphs” where ideas linked to each other in networks mirroring thought itself.
Obsidian, Notion, and the PKM Explosion#
Following Roam, dozens of tools emerged: Obsidian, Logseq, Notion (with databases), Reflect, and more. Each offered slightly different takes on networked note-taking.
The common thread: links between notes were as important as the notes themselves. Your knowledge base was a network, not a hierarchy.
This movement recognized something important: the future of personal knowledge isn’t just capturing information. It’s capturing the connections between pieces of information.
Where External Links Fit#
But most PKM tools focus on internal links between your own notes. External links to web content get less attention.
This always felt backward to me. Most valuable knowledge already exists on the web, written by people more expert than you. Why duplicate it into notes when you could just organize the links to it?
This realization was central to building stashed.in. Instead of copying web content into notes, what if you organized the links themselves into a network? Add minimal context (why you saved it, how it connects to other things), but let the authoritative content live where its creators maintain it.
Your stashes become more like Nelson’s Memex: curated trails through existing knowledge, organized by your associations, with enough context to remember why things matter.
The AI Age: Hyperlinks as Training Data#
Now we’re in yet another transition. AI is reshaping what links mean.
When AI Summarizes, Do Links Matter?#
If you ask an AI about a topic and it gives you a comprehensive answer, do you need the original links? The AI synthesized from thousands of sources. It’s already done the linking for you, internally.
This creates a crisis for the open web. If people stop clicking links because AI answers their questions, content creators lose traffic. If creators lose traffic, they create less content. If there’s less content, future AI has less to train on.
This isn’t hypothetical. It’s already happening. Search traffic to content sites is declining as AI-generated summaries reduce the need to click through.
Links as Attribution#
One partial solution: AI systems providing links to sources they synthesized from. This preserves attribution and allows verification.
But in practice, when AI gives you an answer, how often do you click through to verify its sources? Probably rarely. The answer feels sufficient.
This is dangerous. AI can be confident and wrong. Without checking sources, we’re trusting synthesized information without verification. Links become vestigial—present but unused.
Personal AI and Your Link Collections#
But here’s where it gets interesting: personal AI that learns from your curated links.
Imagine an AI that knows your stashed.in collections. When you ask it about API design, it can reference the specific articles you’ve saved about that topic, understanding that these are sources you trust and have contextualized.
This is different from AI trained on the entire web. It’s AI trained on your curated slice of the web, organized by your thinking. Your link collections become training data for AI that helps you think.
This could be the next evolution: from hyperlinks as navigation, to hyperlinks as social signals, to hyperlinks as the scaffolding for personal AI assistants.
Hyper-Memory: What Links Are Becoming#
Here’s my thesis: hyperlinks are evolving from connection mechanisms into something more profound: externalized, networked memory systems.
Links as Cognitive Scaffolding#
Your organized link collections aren’t just bookmarks. They’re scaffolding for your thinking.
When you’ve curated 50 resources about system design and organized them with context about why each matters, you’ve externalized your understanding of that domain. The links aren’t just pointers to information. They’re a map of your knowledge.
This is different from notes. Notes capture your interpretation. Links point to authoritative sources while adding your layer of context and connection.
When you return to your collection months later, you’re not just finding resources. You’re reconnecting with your past thinking about that topic. The links plus your context notes create a form of memory.
The Social Dimension of Link Memory#
When you make your link collections public (or share them with password-protected access), something interesting happens: your memory becomes social.
Others can follow your trails through a topic. They see what you found valuable and why. They contribute their own finds. The collection becomes collaborative memory.
This is what I see happening on stashed.in. People create stashes about topics they care about. Others discover those stashes, follow them, add to them. The collections become living documents of collective knowledge about specific domains.
It’s not quite Ted Nelson’s two-way links, but it’s a form of networked, social memory built on hyperlinks.
Visual Organization as Memory Aid#
There’s a reason Pinterest is so popular: visual organization matches how memory works.
When I designed stashed.in with visual collections (stashes with header images, like Pinterest boards), it was specifically to tap into visual memory. You don’t just remember “my design inspiration collection.” You remember the aesthetic of that collection, its header image, the overall vibe.
This makes browsing your collections feel more like exploring a space than searching a database. And spatial memory is one of the strongest types of human memory.
Your link collections become places you visit, not just lists you search.
Building Your Hyper-Memory System#
Understanding this evolution changes how you should think about organizing links. You’re not just bookmarking. You’re building external memory.
Save With Context, Always#
Never save a bare URL. Always add at least one sentence about why it matters.
Not what it’s about (the title shows that). Why past-you thought it was valuable. What problem it solves. How it connects to your thinking.
This context is the difference between a list of links and a memory system. It captures the associative reason you saved it, which helps you remember and rediscover it later.
Organize by Association, Not Hierarchy#
Don’t force links into rigid folder structures. Create flexible collections based on themes, projects, or associations.
A single link can belong to multiple collections. That article about remote work might live in both “Management” and “Distributed Teams” and “Async Communication.” This mirrors how your brain actually connects concepts.
Let the organization emerge from use rather than planning it perfectly upfront.
Review and Wander Regularly#
Hyper-memory only works if you revisit it. Schedule time to browse your collections, not just search them.
Weekly or monthly, wander through what you’ve saved. Rediscover things. Notice new connections. Add notes about how your thinking has evolved.
This spaced revisiting reinforces memory and reveals patterns you missed initially.
Share Selectively#
Make some collections public to contribute to collective knowledge. Keep others private for personal exploration. Use password protection for team or group resources.
The act of sharing forces you to curate more thoughtfully, which strengthens the memory value.
Connect Across Collections#
When you notice links from different collections relating to each other, make that connection explicit. Add notes, tags, or references that link them.
These cross-connections are where novel insights emerge. Your AI collection connects to your education collection connects to your business collection. The intersections are often the most interesting part.
What’s Next for Hyperlinks#
Looking ahead, I see several possible futures:
Semantic links. Instead of just “click here,” links could carry more semantic information about the relationship. “Agrees with,” “contradicts,” “extends,” “exemplifies.” This would make the web’s knowledge graph more explicit.
Persistent links. Technologies like IPFS and blockchain could create links that never break, even if the original site goes offline. Content would be permanent and addressable.
Personal link graphs. AI that understands your link collections and helps you navigate them, discover connections, and synthesize insights. Your hyper-memory becomes queryable intelligence.
Collaborative link networks. Tools that make it easy to build shared link collections with communities, creating curated trails through knowledge domains maintained by many people.
Attribution systems. Automated micropayments or credit systems that reward creators when their content gets linked and used. This could fund quality content creation sustainably.
Whatever happens, hyperlinks will continue evolving. They started as simple document connections. They became social signals. They enabled networked memory. What’s next is limited only by imagination.
Your Hyper-Memory Starts Now#
You don’t need to wait for future technologies. The tools to build hyper-memory systems exist today.
Start by shifting how you think about the links you save. You’re not just bookmarking. You’re building external memory. You’re creating trails through knowledge. You’re scaffolding your thinking.
Every link you save with thoughtful context is a node in your personal knowledge network. Over time, these nodes connect. Patterns emerge. Your understanding deepens, not just through reading but through organized rediscovery.
The web is humanity’s shared memory. Your curated link collections are your personal memory within that larger system.
Build it deliberately. Tend it regularly. Share it generously.
Your future self will thank you for the trails you’re leaving through the infinite web.
Start building your hyper-memory today.
