Brain 2.0

I. Why Externalise Cognition

The biological brain has constraints that pre-modern thinkers worked around through writing, libraries, and broader external memory systems. The opportunity is that those constraints have not changed but the externalisation tools have expanded enormously. Most people are operating with cognitive infrastructure weaker than what’s available, often because they don’t recognise the externalisation question as a question they could be asking.

The biological brain stores poorly, recalls inconsistently, fabricates retrospectively, and forgets reliably. None of these are bugs. Rather, they are evolved features that traded off detail retention against energetic efficiency. The hippocampus is a pattern-matching system optimised for survival rather than knowledge work, so don’t treat it like a hard drive.

The compensating moves humans have developed over centuries:

• Writing (the fundamental cognitive extension, dating back ~5000 years)
• Libraries (organised external memory at population scale)
• Filing systems (personal external memory organised for retrieval)
• Note-taking (active engagement with external memory during learning and thinking)
• Indexing (retrieval infrastructure that makes external memory usable)
• Databases (computational external memory with query capabilities)
• Personal Knowledge Management systems (the contemporary integration of these moves into personal practice)

Each step has been an expansion of what humans can think with. The brain itself hasn’t changed; the cognitive infrastructure surrounding it has. The Brain 2.0 framing in this section refers specifically to the contemporary opportunity: using current tools to externalise more of your thinking than previous generations could.

You can now construct a personal external memory that genuinely augments your thinking rather than merely storing information for occasional reference. The integration of capture, organisation, retrieval, and active use produces measurable improvements in what you can think about and how reliably you can build on previous thinking over time.

II. The Limits of Biological Memory

Worth understanding the specific constraints so the externalisation work has an empirical basis.

• Working memory capacity: Approximately 4-7 items at a time, depending on the research framework you adopt. The original Miller (1956) “magical number seven plus or minus two” has been refined downward by subsequent research; current estimates suggest 4±1 chunks for most cognitive tasks. The implication: complex thinking requires holding more than this in some accessible form. The externalisation tools function as working memory extension, letting you operate on more variables than your biological capacity supports.
• Long-term memory storage: Effectively unlimited capacity but unreliable encoding and retrieval. You forget most of what you experience. What you do retain is often distorted, blended with other memories, or reconstructed from gist rather than recalled precisely. The Ebbinghaus forgetting curve (1885) documented that without reinforcement, approximately 50% of new information is lost within hours and ~70% within a day.
• Retrieval failures: Even retained information often cannot be accessed when needed. The “tip of the tongue” phenomenon is the most familiar example; the broader pattern is that you have information that doesn’t surface when relevant. Externalisation tools improve retrieval reliability.
• Reconstructive memory: Memory is reconstructive rather than reproductive. You don’t replay encoded information; you rebuild approximations using the gist plus current context. This means each recall slightly alters the memory. Memories accessed frequently become more solidified but also more shaped by the contexts of repeated recall. The fidelity to the original experience can diminish over time even as confidence in the memory increases.

The implications for thinking work: Trying to do knowledge work entirely within biological memory means:

• Operating with severely constrained working memory
• Forgetting most of what you encounter, including your own previous insights
• Reconstructing previous thinking inaccurately when you try to build on it
• Repeating the same intellectual moves multiple times because you’ve forgotten you made them before
• Missing connections between ideas separated by time
• Producing thinking that doesn’t compound across years the way it could

How much of your thinking work do you want to do with the assistance of external infrastructure?

III. What Brain 2.0 Actually Is

Brain 2.0 is the deliberate construction of a personal external memory system designed to augment biological cognition. The components:

• Capture: Reliable methods for getting information, ideas, observations, and thinking into the external system as you encounter them. Without reliable capture, the system contains only a subset of what you’ve thought about.
• Organisation: Structure that lets you find what you’ve captured when you need it. Without organisation, the captured material accumulates without becoming useful.
• Connection: Links between related items that let ideas combine in ways biological memory often misses. The connection layer is what produces emergent insight from accumulated material.
• Retrieval: Mechanisms for finding specific items or related material when you need them. The retrieval question is what determines whether the system supports your thinking or just stores it.
• Active use: Practices that engage the external system during thinking rather than treating it as a passive archive. The active use is what distinguishes a Brain 2.0 from a sophisticated filing cabinet.

Niklas Luhmann, the German sociologist, built his Zettelkasten (slip-box) system in the mid-twentieth century and produced over 70 books and 400 academic papers using it. He attributed significant portions of his productivity to the slip-box rather than his own intellectual capacity. The system did much of the work of holding and connecting ideas; Luhmann did the work of generating and refining them.

Contemporary tools have made similar systems accessible to anyone, not just academics with filing space. The opportunity is to build systems that produce similar compounding benefits.

IV. The PKM Landscape

Personal Knowledge Management has emerged as a popular movement over the past decade. 

The intellectual lineage:

• Vannevar Bush’s Memex (1945): The conceptual precursor. Bush’s essay “As We May Think” imagined a microfilm-based personal memory system with associative links between items. The system was never built but the concept influenced subsequent computer pioneers.
• Ted Nelson’s Hypertext (1965): The conceptual ancestor of the World Wide Web. Nelson’s Project Xanadu imagined a global hypertext system with bidirectional links. The web that emerged is a diminished version of Nelson’s original vision.
• Niklas Luhmann’s Zettelkasten (1950s-1990s): The most-cited practical example. Luhmann built ~90,000 paper slips organised through reference codes and tags, producing extensive academic output that he credited to the system.
• Tim Berners-Lee’s World Wide Web (1989): The contemporary infrastructure that makes networked personal knowledge systems possible.

The contemporary popularisation:

• Tiago Forte’s Building a Second Brain (2022): The current popular synthesis with the BASB methodology (Capture, Organise, Distill, Express). Forte’s online course has been taken by tens of thousands of people; the book is the accessible written version.
• Sönke Ahrens’s How to Take Smart Notes (2017): The contemporary Zettelkasten popularisation. The book translates Luhmann’s method for modern knowledge workers.
• Roam Research (2019-): The bidirectional-linking note tool that introduced many users to networked note-taking. This is what the original Human Operating Manual was drafted on. 
• Obsidian (2020-): The local-file Markdown-based alternative to Roam that has accumulated a substantial user community.
• Notion (2016-): The broader productivity platform that integrates notes, databases, and project management. I now store all my files here. 

The PKM movement: A large part of the PKM community spend more time configuring their tools than thinking with them. The “tool tourism” pattern is well-documented; users repeatedly switch between Roam, Obsidian, Notion, Logseq, Tana, and others, optimising the system instead of using it. The reasonable position: pick a tool that meets your needs, commit to it for at least six months, evaluate based on whether your thinking has improved. The optimisation work past a certain point is procrastination disguised as preparation.

V. The Three Major Methods

Zettelkasten

The slip-box method was developed by Luhmann and popularised by Sönke Ahrens.

The principles:

• Atomic notes: Each note captures one idea, concept, or argument. Not summaries of articles; specific ideas extracted and rewritten in your own words.
• Permanent notes: Notes are written for your future self. They include enough context that they make sense without external reference. They use plain language rather than coded shorthand.
• Linked notes: Each note connects to related notes through explicit links. The connection structure is the value; isolated notes are less useful than connected ones.
• Emergence: The system is designed to produce emergent insight as connections accumulate. Patterns become visible across many notes that wouldn’t be visible in linear reading.
• No predetermined hierarchy: Notes are organised by connection rather than by category. Categories often turn out to be wrong; connections are organic.

The practice:

• When you read or think something interesting, write a fleeting note (rough capture)
• Within a day or two, convert fleeting notes into permanent notes (atomic, well-written, contextualised)
• Link new permanent notes to existing related notes
• Periodically review the slip-box; let connections surface ideas for writing
• When writing, draw from the slip-box rather than from blank page

The strengths: Produces durable thinking that compounds over the years. Luhmann’s output emerged from the slip-box’s accumulated connections. Works particularly well for writing-intensive knowledge work.

The limits: Requires commitment to writing notes well. Half-effort Zettelkasten produces less value than other systems with similar effort. The connection work is genuinely demanding; you can’t shortcut it.

PARA

Tiago Forte’s organisational framework. Stands for Projects, Areas, Resources, Archive.

The principles:

• Projects: Time-bounded efforts with specific outcomes. “Write blog post on habits” is a project; “be a better writer” is not.
• Areas: Ongoing responsibilities and standards. Health, finances, relationships, professional development. No completion date.
• Resources: Topics of interest that may inform projects or areas. Information stored for future reference.
• Archive: Completed projects, inactive areas, no-longer-relevant resources. Stored for reference but not active.

The practice:

• Every piece of information goes into one of the four categories based on actionability
• Projects get the most attention because they have specific outcomes
• Areas get sustained attention because they represent ongoing standards
• Resources get attention only when they’re relevant to a current project or area
• Archive gets attention only when something there becomes relevant again
• The system is reviewed periodically and items move between categories as their relevance changes

The strengths: Action-oriented. The categories are designed around what you actually do with information rather than around how the information is structured. Works well for people whose knowledge work is project-driven.

The limits: Less emphasis on connection between ideas than Zettelkasten. The atomic-note discipline that produces Zettelkasten’s emergence isn’t core to PARA. Works less well for sustained intellectual work where the connections themselves are the substantial output.

Building a Second Brain (BASB)

Forte’s broader methodology incorporates PARA but extends it.

The CODE method:

• Capture: Save anything that resonates from any source. Don’t filter heavily at capture; let the filtering happen later.
• Organise: Place captured items into the PARA structure based on actionability.
• Distill: Progressively summarise items so the essential ideas surface when you revisit them.
• Express: Use captured and distilled material to produce output (writing, decisions, creative work).

The practice:

• Capture interesting material in a quick-access tool throughout the day
• Process captured material into PARA structure regularly (weekly or so)
• Progressively summarise important items (bolding key passages, then italicising the essential parts of bolded sections, then writing a brief summary at the top)
• Use the distilled material when producing output, drawing on what you’ve captured rather than starting from blank page

The strengths: Practical and accessible. Lower barrier to entry than Zettelkasten. The expression emphasis is genuinely useful; many PKM users get stuck at the capture stage and never produce output.

The limits: The Forte approach has accumulated a commercial ecosystem (courses, certifications, premium memberships) that warrants calibration. The methodology itself is sound; the marketing has produced some overselling. The core principles can be adopted without purchasing additional courses.

VI. The Capture-Organise-Distill-Express Practice

Capture

The reliable habit of getting interesting material into the system. Worth getting right because everything downstream depends on capture quality.

What to capture:

• Ideas that strike you during reading
• Observations from daily life that seem significant
• Things you said or someone said that captured something important
• Questions that recur in your thinking
• Connections between previously-unconnected ideas
• Excerpts from books, articles, conversations
• Decisions and their reasoning
• Specific examples that illustrate broader principles

The capture tool: Something low-friction enough that you actually use it. The fanciest system you don’t use is worse than the simple system you do. For many people, the right capture tool is the notes app on their phone. Sophisticated users may use dedicated capture tools; the dedication usually doesn’t matter as much as the consistency.

The capture discipline: Capture is the moment when the cost is highest (you have to interrupt what you’re doing) and the value is least visible (the idea seems obvious now, you’ll surely remember it). The discipline is capturing anyway, because the value compounds over time, but only if the capture happens.

Organise

The cadence: Daily quick review of capture, weekly broader processing into the PARA or Zettelkasten structure. Less often than daily produces backlogs that become demoralising; more often than weekly tends to be over-engineering.

The decisions during organisation:

• Is this still relevant?
• Where does it belong in my system?
• What does it connect to?
• Does it need to be reformulated for future-me to make sense of it?
• Should it become an atomic note (Zettelkasten) or stay as a captured excerpt (PARA/BASB)?

The purge discipline: Some captured material doesn’t survive organisation. Recognising what isn’t actually useful and removing it (or archiving it) keeps the system functional. The accumulation pattern (capture everything, organise nothing) produces systems that become unusable.

Distill

The Forte technique:

• First pass: highlight the parts that resonate
• Second pass: bold the parts of the highlighted material that are most essential
• Third pass: italicise the parts of the bolded material that capture the core idea
• Final pass: write a brief summary at the top in your own words

The progressive narrowing means you can scan any note and immediately access the essential idea while preserving the full context for when you need it.

The Zettelkasten alternative: Rather than distilling existing material, write atomic notes that already contain the essential idea in compressed form. The distillation happens at the writing stage rather than as a separate process.

The principle: Material you’ve captured but never distilled tends not to be used. The distillation work is what makes the captured material accessible during thinking.

Express

The practice that distinguishes Brain 2.0 from elaborate filing. Using the captured and distilled material to produce output.

The forms:

• Writing that draws on accumulated material rather than starting from a blank page
• Decisions informed by previous thinking on similar questions
• Conversations enriched by relevant captured material
• Creative work that combines previously-disparate elements
• Sustained thinking on questions that benefit from accumulated material

The discipline: Expression should be the goal of the system, not capture or organisation. The PKM communities that spend extensive time configuring their systems without producing output have lost the plot. The system exists to support thinking and output; if it isn’t, something has gone wrong.

The integration with writing: For people doing writing work, the Brain 2.0 system becomes the source material. New writing starts from what’s already in the system rather than from external research alone. This is what produced Luhmann’s productivity; it remains the value of the system for writers.

VII. The Current Tool Landscape

The tool ecosystem as of the current moment. Tools come and go; the underlying practice is more durable than any specific tool.

Networked Notes Tools

The category that has emerged in the past several years around bidirectional linking and graph views.

• Roam Research: The original bidirectional-linking note tool that catalysed the contemporary PKM movement. Subscription-based, cloud-hosted. Strong community of “Roam thoughtpartners.” The original tool but has been criticised for slower development and platform lock-in.
• Obsidian: Local-file Markdown-based alternative. Free for personal use; paid sync and publishing services. Plugin ecosystem. Strong user community. The tool many people who left Roam migrated to.
• Logseq: Open-source alternative with outliner structure. Free, local-file. Smaller community than Obsidian but actively developed.
• Tana: Newer entrant with more structured database features. Subscription-based. Some users find it more powerful; others find it more complex than needed.
• RemNote: Note tool with spaced repetition built in. Useful for people doing learning work.

Broader Productivity Platforms

Tools that include note-taking among broader features.

• Notion: The dominant broader productivity platform. Notes, databases, project management, team collaboration. Massive feature set. Some users build their entire Brain 2.0 in Notion; others find it too heavy for note-taking specifically.
• Craft: Apple-ecosystem-focused alternative to Notion with cleaner interface.
• Coda: Document-database hybrid with automation features.

Traditional Note Tools

The category that predates the networked-notes movement.

• Evernote: The dominant note app for ~15 years. Recently sold and has accumulated mixed reviews. Workable but less fashionable than current alternatives.
• OneNote: Microsoft’s note app. Free for most users. Some find it powerful; others find the interface less inviting than alternatives. I used this during my university years uuntil I hit the free space limitations.
• Apple Notes: The default Apple notes app. Improved in recent years. Increasingly capable for basic Brain 2.0 use.
• Google Keep. Google’s lightweight note app. Useful for quick capture; less useful for knowledge work.

Project Management Tools

Sometimes adapted for Brain 2.0 use.

• Trello: Kanban-style boards. Useful for project tracking; less useful for note-taking specifically.
• Todoist, Things, OmniFocus: Task management tools. Useful for the project layer of PARA but not for the broader knowledge work.

File-Based Systems

Sometimes overlooked but functional.

• Google Drive: Documents, spreadsheets, slides plus storage. Many people’s de facto Brain 2.0 even when they haven’t thought about it as such. A large portion of knowledge work happens here regardless of dedicated PKM tools.
• Dropbox: File storage with third-party tool integration.
• Plain text and Markdown: The most portable option. Text files in folders, organised through naming conventions. Works in any tool; never locked into a platform. Some serious knowledge workers stick with this for portability reasons.

The Practical Recommendation

For most people starting:

• Pick a tool that meets your needs without being more complex than you need
• For basic capture-and-retrieve work, Apple Notes or Google Keep is enough
• For networked thinking work, Obsidian or Roam
• For broader productivity integration, Notion
• For maximum portability, plain text files

Commit to the choice for at least six months. Re-evaluate based on whether your thinking has improved, not based on whether someone else’s setup looks more elaborate.

VIII. The AI Extension Question

The development: Large language models (ChatGPT, Claude, Gemini, and others) have rapidly become useful cognitive tools. They can summarise, expand, restructure, critique, and generate text in ways that genuinely extend what individuals can do. The question is what role they should play in personal knowledge work.

LLMs are imperfect tools that produce useful output when used appropriately and misleading output when used inappropriately. They are not substitutes for thinking; they are extensions of certain thinking operations. The skill is knowing which.

Any idiot can see that I used LLMs to help write the HOM. However, this project began by writing out the framework and bulk content and years of research. I then used this information to inform the writing process on Claude before editing it here. A useful tip for writing with the aid of an LLM is to leave intentional roadblocks for you to edit. LLMs are notoriously bad at adding repetitive words and predictable sentence structures. If you’re smart, you’ll pick up on this and rewrite it in your own way, helping to consolidate your knowledge. If you’re lazy (and most people are by our very nature), you’ll post it publicly with hubris and unearned pride. Similar to that same feeling you get when you find a dollar on the ground and get a free prize in a potato chip bag. The HOM is a solo project that will take the rest of my life to complete, so if you notice particularly hallucinatory content in any section, please reach out and let me know. It may save me another year of editing.  

Useful applications:

• Summarisation: Compressing long texts into shorter forms. Particularly useful for processing captured material.
• Restructuring: Reorganising existing thinking in different forms (turning notes into outlines, outlines into prose, prose into bullet points).
• Expansion: Taking compressed ideas and expanding them with examples, applications, or related considerations.
• Critique: Engaging your existing thinking with counter-arguments, alternative perspectives, or specific weaknesses.
• Connection: Identifying potential connections between ideas you might not have seen.
• Translation: Moving between technical and accessible language, between disciplines, between formats.

Limitations:

• Hallucination: LLMs confidently generate false information. Any factual claim from an LLM warrants verification before use.
• Sycophancy: LLMs tend to agree with users in ways that aren’t always accurate. Pushback on your ideas often requires explicit prompting to override the default agreement.
• Surface-level pattern-matching: LLMs often produce output that sounds substantive but lacks the underlying understanding the words imply. Distinguishing genuinely useful output from articulate-sounding nonsense requires judgement.
• Training-data limits: LLMs know what’s in their training data and don’t know what isn’t. For specialised fields or recent developments, the LLM may not have relevant information.
• Style homogenisation: Extensive use of LLM-assisted writing tends to produce homogenised prose. Distinctive voice requires deliberate work to maintain.

The “AI as second brain” framing often substitutes the LLM for actual thinking rather than augmenting it. The pattern of using LLMs to write content that the user hasn’t actually thought through is widespread and produces low-quality output that the user often cannot recognise as low-quality.

This entire page could be garbage for all I know. It looks lovely, though, doesn’t it?

LLMs are useful tools for specific operations within a broader thinking practice. They don’t replace the thinking; they extend specific moves within it. The skill is integrating them appropriately, neither dismissing them nor over-relying on them. Remember, a hallmark of learning is the suffering experienced while solving a problem. 

Practical recommendations:

• Use LLMs for tasks where verification is easy (summarisation, restructuring, critique)
• Don’t use LLMs for tasks where verification is hard (novel factual claims, expert judgement in domains you don’t understand)
• Always read what the LLM produces before using it; don’t pipe output to other places without review
• Maintain your own thinking as the primary thread; LLMs are tools you reach for, not the substrate everything runs through
• Notice when LLM use is degrading your own capacity; some people find sustained LLM use reduces their independent thinking capacity over time

The LLM integration question warrants its own ongoing development as the tools mature. The current state of practice is less settled than it will be in five years; treat your own approach as provisional and update based on experience.

IX. Common Failure Modes

• The tool tourism pattern: Constantly switching between PKM tools, optimising the system instead of using it. Cure: commit to one tool for at least six months. Evaluate based on output, not based on whether someone else’s setup looks better.
• The capture-without-use pattern: Capturing extensive material that you never return to. Cure: schedule regular processing time. Aim to use captured material in output rather than just accumulating it.
• The over-engineering pattern: Building elaborate organisational structures that you spend more time maintaining than benefiting from. Cure: simpler is usually better. Start minimal; add complexity only when specific gaps appear.
• The “perfect system before starting” pattern: Spending months designing the ideal Brain 2.0 before actually using one. Cure: start with whatever you have. Build the system through use rather than by anticipating use.
• The “private journal” pattern: Using the system only for emotional processing rather than knowledge work. Cure: a journal is a legitimate tool but distinct from a Brain 2.0. If you want both, keep them separate.
• The “all input, no output” pattern: Heavy capture and organisation without producing anything from the system. Cure: build expression into the practice. Even private writing for yourself counts; pure accumulation without use isn’t the work.
• The “forgetting the system exists” pattern: Building a Brain 2.0 then drifting back to operating from biological memory only. Cure: integrate the system into daily workflow. Make it your first stop for relevant questions rather than an afterthought.
• The “fragility” pattern: Building the system in a tool that disappears or becomes unusable. Cure: maintain backups. Prefer tools with open formats (plain text, Markdown) over proprietary formats. Don’t put years of work into a tool you cannot extract from.
• The “anxiety about completeness” pattern: Feeling you need to capture everything important or the system fails. Cure: capture what you actually care about. Missing things is fine. The system gets better through use, not through completeness.

X. Selecting Your System

How do you actually work? What kind of thinking do you do? What outputs are you producing or want to produce? The right system depends on these answers more than on general best practices.

• For writing-heavy knowledge work: Zettelkasten in Obsidian or Roam. The atomic-note discipline produces durable thinking; the bidirectional linking produces emergent connections.
• For project-driven work: PARA in Notion or whatever broader productivity tool you use. The action-orientation matches the work pattern.
• For learning new material: Spaced repetition tools (RemNote, Anki) plus a capture system for the broader notes. The spaced repetition handles retention; the capture system handles connection.
• For research-heavy work: Zotero (citation management) plus notes in whatever tool you prefer. Reference manager handles the citations; notes handle your thinking.
• For broader life management: Apple Notes or Google Keep plus whatever calendar and task tool you use. Don’t over-engineer the system for a use case that doesn’t require it.
• For maximum portability: Plain text or Markdown in folders. Works in any tool; never locked in.
• The starting principle: Start simpler than you think you need. Add complexity only when specific gaps appear. The system grows through use; you cannot design the optimal system in advance.

XI. Long-Term Maintenance

• Periodic review: Monthly review of what’s accumulated. Quarterly review of broader structure. Annual review of what’s working and what isn’t.
• The pruning practice: Some captured material doesn’t survive long enough to be worth keeping. Periodic pruning prevents the system from accumulating dead weight.
• The migration practice: Tools change. The active practice is more durable than any specific tool. Migrate when needed; don’t migrate just because something new appeared.
• The integration practice: The system should be integrated into daily work, not separate from it. If you have to remember to use it, you’ll forget. If it’s part of the workflow, it stays alive.
• The compounding observation: Brain 2.0 systems compound over years. Year one feels like work for modest return. Year five looks substantially different; the accumulated material starts producing connections and insights that wouldn’t be possible without the system. The compounding is what makes the sustained investment worthwhile.

The patient practice over the years produces results that brief experimentation cannot. The Brain 2.0 work is more like building a habit than learning a skill; the value accumulates through sustained use.

XII. Cross-Links

• Mental Models for the section overview
• Heuristics Basics for the System 1 territory
• Mental Model Basics for the analytical mental models
• Mental Model Cheatsheet for the practical catalogue
• The Mental Model Rabbit Hole for the deeper material
• Resources for the reading list