A second brain is an external store for the things you should not have to keep in your head: references, decisions, half-formed ideas, and the reasons behind them. The metaphor is popular because the underlying complaint is real—working memory is small and the volume of information arriving each day is not. What the metaphor hides is that a second brain, unlike the first one, does not remember anything unless you build the remembering.
Where the idea comes from
The phrase was popularized by Tiago Forte, whose book Building a Second Brain appeared in June 2022 after several years of teaching the method as a course. It arrived into an existing tradition: commonplace books kept by readers since the Renaissance, Luhmann's slip-box of about 90,000 cards, and the digital note-taking wave that followed the launch of Roam Research in 2019 and Obsidian in 2020.
Forte's contribution was less the storage idea than the sequencing: capture first, organize by actionability, distill on re-reading, and produce something. That ordering matters, because most people who describe a failed system describe having done the first step for two years and none of the others.
CODE and PARA, in plain terms
The method has two acronyms, and they answer different questions. CODE is the workflow; PARA is the filing scheme.
- Capture: save what resonates, not everything—the instruction is deliberately subjective, which is both its strength and its ambiguity.
- Organize: file by how actionable an item is, using PARA—Projects, Areas, Resources, Archives.
- Distill: on each revisit, bold or summarize the parts that matter, so a future read costs seconds rather than minutes.
- Express: turn the material into a draft, a decision, or an argument. This is the step that pays for the other three.
What the method gets right
Two of its claims hold up well. The first is that offloading reduces the low-grade cognitive tax of trying not to forget—the mental rehearsal that runs in the background when you are holding something you have not written down. The second is that value comes from use, not accumulation: an archive that never feeds a document, a decision, or a conversation is a hobby.
The insistence on capturing before judging is also correct in a way that is easy to underestimate. Ideas rarely announce their importance at the moment they arrive, and a filter applied at capture time is applied by the version of you least equipped to judge.
Where the method breaks in practice
PARA asks for a placement decision on every item, and placement is exactly what an interrupted person cannot do. A note taken while walking out of a meeting belongs to a project, an area, and a person simultaneously; forcing one location either loses the other two or triggers duplication.
Distillation has the same problem in a worse form: it is a recurring manual task with no deadline, so it loses to everything with one. And the archives layer quietly assumes you will remember which of four buckets a two-year-old note went into—an assumption the passage of time falsifies reliably.
The 2026 version of the same idea
The parts of the method that survive are capture and use. The parts that are increasingly done for you are organization and distillation: automatic topics, transcription of voice, extracted commitments, and retrieval that matches meaning rather than spelling.
That is a genuine narrowing of the workload, and it is not a solved problem. Automatic classification is probabilistic and mislabels items; extracted tasks need review; and an answer assembled from your notes is only trustworthy if it shows which notes it used and lets you open them.
Should you build one?
If your work involves returning to earlier material—writing, research, product decisions, client work—then yes, but build the shallow version. One capture stream, no filing at capture time, and retrieval you have actually tested. Most of the elaborate setups described online were built by people whose job is describing setups.
If your work is mostly execution against a short horizon, the honest answer is that a task list and a calendar will serve you better, and the archive will become a museum. There is no controlled evidence that any note-taking method improves output; what the popular accounts share is survivorship, and the systems that quietly collapsed do not get written up.
Build the smallest version that answers questions: capture without filing for a month, then ask five things about that month you genuinely cannot recall. The number you can answer from your own archive is the only measure of whether the system is real.