[commentary on Roam post-mortem] I have to say, I still don’t understand the cult of Roam or why people were so impressed by, eg. the [[link]] syntax borrowed from English Wikipedia (which introduced that syntax ~18 years before on what is still the most widely-read & edited wiki software in history), which you remark on repeatedly. Even in 2019 in beta it just seemed like a personal wiki, not much different from, say, PmWiki (2002_22ya) with some more emphasis than usual on the common backlink or ‘reverse citation’ functionality (that so many hypertext systems had supported going back decades in parallel with Project Xanadu ideas). It may be nicer than, say, English Wikipedia’s “What Links Here” (which has been there since before I began using it early in the 2000s), but nothing to create a social-media cult over or sell “courses” about (!).

But if the bubble has burst, it’s not hard to see why: any note-taking, personal knowledge management, or personal wiki system is inherently limited by the fact that they require a lot of work for what is, for most people, little gain. For most people, trying to track all of this stuff is as useful as exact itemized grocery store receipts from 5 years ago.

Most people simply have no need for lots of half-formed ideas, random lists of research papers, and so on. This is what people always miss about “Zettelkasten”: are you writing a book? Are you a historian or Teutonic scholar like Niklas Luhmann? Do you publish a dozen papers a year? Are you the 1% of the 1%? No? Then why do you think you need a Zettelkasten? If you are going to be pulling out a decent chunk of those references for an essay or something, possibly decades from now, then it can be worth the upfront cost of entering references into your system, knowing that you’ll never use most of them and the benefit is mostly from the long tail, and you will, in the natural course of usage, periodically look over them to foster serendipity & creativity; if you aren’t writing all that, then there’s no long tail, no real benefit, no intrinsic review & serendipity, and it’s just a massive time & energy sink. Eventually, the user abandons it… and their life gets better.

Further, these systems are inherently passive, and force people to become secretaries, typists, reference librarians, archivists, & writers simply to keep it from rotting (quite aside from any mere software issue), to keep it up to date, revise tenses or references, fix spelling errors, deal with link rot, and so on. (Surprisingly, most people do not find that enjoyable.)

‘Tools for thought’ winds up being a lie: there’s tools, but not much additional thought. There is no intelligence in such systems, and they don’t do anything. The user still has to do all the thinking—and it adds on a lot of thinking overhead.

So what comes after Roam and other personal systems which force the user to do all the thinking? I should think that would be obvious: systems which can think for the user instead. LLMs and other contemporary AI are wildly underused in the personal system space right now, and can potentially fix a lot of these issues, through approaches like actively surfacing connections instead of passively waiting for the user to make them on their own and manually record them, and can proactively suggest edits & updates & fixes that the user simply approves in batches. (Think of how much easier it is to copyedit a document using a spellcheck as a series of Y/N semi-automatic edits, than to go through it by eye, fixing typos.)

However, like most such paradigm shifts, it will be hard to tack it onto existing systems. You can’t reap the full benefits of LLMs with some tweaks like ’let’s embed all the documents and then add a little k-nearest-neighbor retrieval pane!’.

You need to rethink the entire system and rewrite it from the ground up on the basis of making neural nets do as much as possible, to figure out the new capabilities and design patterns, and what to drop from the old obsolete personal wikis like Roam.

From what it sounds like, the Roam community would never stand for that (because that would be costly for current users) and those who want more have already left, and I have a lot of doubts about whether it makes sense economically to try. It seems like if one wanted to do that, it would be better to start with a clean sheet (and a clean cap table).

See Also: