Claude: NeoBooks Project
Me:
The map I just uploaded is a pretty major project for me. I've always wanted to write a book, possibly several books. Part of this project is explaining to you what the NeoBooks project is, which is kind of a deconstruction of books so that we can actually Nuggetize the contents of books and use some Nuggets in a variety of books. I'll explain it in more detail later.
I have a bunch of different book ideas, which I haven't been able to pick my way through.
In the year 2000, I had a book disaster: I wrote three drafts of a book about the word "consumer." Each would have been about a three hundred page book, so they would have had to have been pared back. Each draft was eighty percent different from the previous one, but the last draft coincided with the dot-com implosion, and my publisher canceled my contract. So I have yet to publish any book at all.
I would like to create NeoBooks around things like Design from Trust and Snip!, and I've long had the name, What If We Trusted You, which is sort of the more narrative version of design from trust. I also have a Google doc that has in it four or five different potential outlines for a book (uploaded), and I have this markdown file, which is the table of contents for a possible book around design from trust: https://wiki.openglobalmind.com/projects/jerry's_nuggets/dft_book_contents_(the_toc)
Under goals and priorities, I want it to be very accessible. I want it to be "Muggle-friendly." But I also want to get attention, For the practical reason of making a living, but also because I think there are big ideas here.
I've managed to live with these ideas some 30 years and get remarkably few of them into the world, so I'm interested in creating a streamlined process where I regularly turn out Nuggets that are elements of eventual NeoBooks (as well as parts of other narratives).
I could also use a neutral assessment of the ideas I carry (Snip, DfT, the Big Fungus, my Brain, etc) in terms of their utility to the world, and possible paths to get them into the world. Despite my saying here I want to publish one or more books, I believe books are where information goes to die (I have a Thought on that in my Brain), so the more interesting artifact to me are the interconnected Nuggets that get better over time because they exist in community, which is why I publish them on Github now, even though I have zero community interaction there. But please factor that into your recommendations. I either need to activate Github collaboration, or switch platforms.
I hope that makes sense, even as I'm sure I've forgotten a few things that are important to me.
Claude:
I've now read through the map, the five outlines, and the DfT NeoBook table of contents on the OGM wiki — so I can see that the NeoBook structure isn't just a concept, you've already started building it: the double introduction, the DfT skeleton, and the parallel WIWTY book it's "intertwingled" with. That's useful, because it means several of the things you're asking about are further along than the "I can't pick my way through" framing suggests.
Before I go into any depth, let me reflect back the shape of what you've put in front of me, because I think it changes where we should start.
You've actually named four different deliverables: a neutral assessment of the ideas themselves, a streamlined Nugget-production process, a platform decision (activate GitHub collaboration vs. switch), and help choosing among the book ideas/outlines. They feel separate, but I don't think they are. They're one system, and the NeoBook is the spine: Nuggets are the unit you write, NeoBooks are reusable assemblies of Nuggets, the platform is where Nuggets live and (maybe) accrue community, and "which book" is really "which assembly do I stand up first." Picking a starting point for one without the others tends to produce the monolithic-draft trap again.
And on that — the 2000 disaster has a structural reading worth saying out loud: three 300-page drafts each 80% different, all riding one publisher's timeline, is a single point of failure on every axis. Big unit, big rewrite cost, one gatekeeper, one deadline. The Nugget/NeoBook architecture is almost exactly the inverse: small units, low rewrite cost, no gatekeeper, continuous publishing. So your instinct that "books are where information goes to die" and your unfinished-book itch aren't actually in conflict — the NeoBook is the form that lets you satisfy both. That's the encouraging part, and I'd want any plan to lean on it.
I don't yet fully understand NeoBooks the way you do — you said you'd explain more, and I'd genuinely like that before I make process or platform recommendations, since both depend on how Nuggets reference and reassemble.
So rather than dump on all four fronts at once, where do you want to start?
Claude's short questionnaire missing here, showing 4 options.
Me:
Q: Given everything's connected, where's the most valuable place for us to start? A: I think 1 will help me assess the situation and weigh things differently, if I have to. 3 can wait a bit. 4 spins out of 1. and 2 needs me to explain NeoBooks more
Claude:
Good call, and the way you've sequenced it makes sense — 4 really is just the back half of 1. Here's my honest read. The headline first: your problem has never been the quality of the ideas. They're good. The variable that actually matters is different, and I'll get to it at the end.
Snip! is your strongest attention asset and probably your most underrated one. A single-word, agentive handle on "decision-makers severed from the consequences of their decisions" is memetic, Muggle-friendly, and explains the news across the whole political spectrum — which directly serves both your goals at once. The episodic "seasons" structure (GFC → finance → war → governance → attention) is a built-in content engine, which is exactly what a regular-output practice needs. The one thing a sharp skeptic will press, and you should have a crisp answer ready, is: how is this not just Taleb's "skin in the game" with better packaging? I think you do have an answer — Snip! foregrounds the deliberate act of severing and ties it to information/context-amputation rather than risk/probability — but if you can't say it in two sentences, reviewers will file it as a relabel. The other discipline is resisting the urge to see a snip everywhere; specialization and division of labor also separate people from consequences and are often good. Knowing where Snip! doesn't apply is what makes it rigorous rather than a vibe.
DfT / What If We Trusted You is your deepest body of work and the actual substance behind the attention — but it's currently stronger as a lens and a critique than as a method someone can pick up on Monday. The word "Design" makes a promise of methodology that the work mostly cashes out as stance-plus-examples. Closing that gap is the highest-value intellectual work you could do here. The central objection you'll always face is that mistrust is sometimes rational and protective — fraud, abuse, bad actors — so DfT sounds naive unless it's clearly bounded as "default to trust and make trust cheap to verify and abuse cheap to detect." Ostrom's commons-governance work is arguably the empirical crown jewel of exactly that bounded version, and leaning on it would armor DfT against the "kumbaya" dismissal. WIWTY is the right narrative front door for this material; DfT-as-method is what serves your advisory clients.
Your Brain is your most unfakeable asset and your least transferable one. Thirty years of continuous curation is a moat nobody can replicate, and it's living proof that a person can think at this scale — which is real credibility. But its utility to others is mostly indirect: it's overwhelming to navigate without you as docent (which is precisely what the Stump Jerry's Brains pilots surfaced), so its honest role isn't "product" — it's the quarry your Nuggets are mined from and the proof-of-concept prop, not a thing to push at a general audience on its own.
The Big Fungus is your most original idea — the least reducible to some existing famous frame — and also, candidly, your least tractable. "Get different thinking tools to interoperate and get people to share and improve each other's maps" is a problem that Xanadu, the Semantic Web, the Tools-for-Thought movement, and several very well-funded companies have all bounced off. It has the classic cold-start problem and is the slowest to monetize, which puts it in direct tension with your near-term living/attention goals. My neutral recommendation is to keep it as the north star that gives meaning to the nearer work rather than something to build in 2026 — and to let your Brain serve as its single-node visible demonstrator. (Minor, but worth saying: "fungus" carries an ick for general audiences that could cost you on the Muggle-friendly axis.)
So the arc that falls out, which I think is also your answer to #4: Snip! is the hook, DfT/WIWTY is the substance, the Brain is the proof, the Big Fungus is the destination. Lead your publishing with Snip! because it's timely, viral-ready, and episodic; use it as the on-ramp to DfT. That's a coherent sequence for both attention and meaning, and it tells you where the first NeoBook should point.
Now the uncomfortable, useful part. The binding constraint across all of this isn't idea quality, idea selection, or platform — it's shipping. You have 625,000 nodes of input and, by your own account, almost nothing finished out the other side in 30 years. The Brain is the most productive input habit imaginable attached to a thin output ledger. Which means the real intervention isn't a better idea or a better tool; it's a cadence — a small, repeatable unit you finish and release on a rhythm, where the win condition is "shipped this week" rather than "masterpiece." Every one of these ideas benefits from that and none of them is unblocked without it. That's exactly what the Nugget→NeoBook process (your #2) is for, and it's why I think that's genuinely the load-bearing piece — but it's gated on you explaining NeoBooks to me properly.
Want to push back on any of this ranking first, or go explain NeoBooks so we can design the cadence? I could also lay this out as a quick 2x2 (attention potential against tractability) if seeing them positioned would help you weigh them.
Me:
Love this, thank you.
There are actually several Snips that matter to me. You pointed to the biggie, about decision makers separated from their decisions, but I really care about the Snip that separated us from one another, converting us from citizens to mere consumers. there are other Snips, so let's return to this, particularly since I'd love your take on how to package them.
I had forgotten entirely about Taleb's book, which I haven't read but did curate into my Brain. for later: a deeper analysis and comparison. I'd love to draw on his work for Snip.
thanks for pointing out that knowing where Snip doesn't apply is important. I agree.
I've made progress with DfT as a bona fide methodology, and would love to have that conversation with you.
I'm a big fan of Ostrom's Commons work (and Meadows' levers of change), but I find both of these frameworks to be often-cited and poorly understood. we can throw the Dunbar Number in there, too. this is just a complicating factor.
agree on the intractability of The Big Fungus, alas. I think I can attract a small cadre of smart folks to discuss and prototype it, but it's not the popular idea to lead with.
oh, and I like the ick of fungus the way people enjoy poop and fart jokes. I want a name for that that riles a little.
that said, I have a long history of identifying important things in the future, but not naming them with the name that eventually sticks. this might be a very fun angle for showing off the times I've predicted things without showing off too much, per our prior conversation about a Hire Me video.
completely agree that my problem is shipping. April and I do share a human remote part-time assistant named Jenn Hines, who is very capable and loves getting things done. problem is, I am terrible at delegating (and asking for help). we should factor all this into the workflow we design.
agree with your ranking and framing, and leading with Snip! makes me smile, which is a good sign. I love that I produced the three GFC Snip! videos some time back; it gives me a lot to build on.
I'd love to hear your thoughts on my replies here, then talk about NeoBooks
Claude:
There's a lot of good signal in here, so let me go point by point, then we'll do NeoBooks.
The plural Snips — this is bigger than I framed it. You've just told me Snip! isn't a single concept with seasons; it's a family of severings, and the family is the real idea. Decision-makers from consequences is one cut. Citizens-into-consumers is another — and notice that one is your Four Separations of Consumerism wearing a different hat, which means Snip! may be the umbrella that finally unifies work you've been carrying in separate buckets for years. That's worth sitting with: if Snip! is "the recurring move of severing a thing from what should stay attached to it," then DfM is what you build once you've snipped (you can't design from trust across a cut you've already made), and consumerism, the GFC, attention markets, and governance are all instances. That's not a content series anymore; that's a theory of how modernity goes wrong. I'd package it as a taxonomy with a flagship: lead publicly with the decision/consequence Snip because it's the most immediately legible, but let each Nugget quietly reveal it's one of a set, so the architecture surfaces over time instead of in one overwhelming dump. The citizens-to-consumers Snip is probably your emotional lead even if the decision-maker one is your topical lead — it's the one that implicates the reader's own life, not just bankers'. Worth deciding which does which job. Let's definitely return to packaging once I've seen the full list.
Taleb — agreed, park it, but here's the frame for later. Skin in the game is about risk asymmetry (you should eat your own downside). Snip! is about information/relationship asymmetry (you've been cut off from the feedback that would let you know, care, or correct). Taleb's remedy is exposure; yours is reconnection. They rhyme, and the honest move is to cite him as a fellow-traveler on one axis while claiming the broader territory — that's stronger than pretending you arrived independently, and it borrows his credibility rather than competing with it. The fact that he's already in your Brain is perfect; the comparison Nugget half-writes itself.
Ostrom, Meadows, Dunbar being "often-cited and poorly understood" is an asset, not a complication. That's a job to be done. "Everyone name-drops Ostrom's eight principles and almost nobody can apply them" is a fantastic Nugget premise — you get to be the person who makes the misunderstood-but-famous thing usable, which is exactly the Muggle-friendly translation work you're good at and which builds authority faster than introducing something new. Same with Meadows' leverage points (the famous list that people quote the top of and ignore the actual ranking of) and Dunbar (constantly cited as one magic number when it's really a series of nested circles). My suggestion: don't treat these as load-bearing scaffolding inside DfT where their fuzziness weakens you — treat each as its own "what this actually means" Nugget, and then cite your own cleaned-up version inside DfT. You convert a complication into three pieces of attention-bait.
The Big Fungus — small cadre, not the lead. Settled, and right. And on the ick: I get it, and I think the instinct is sound — a name with a little gross-out friction is sticky in exactly the way sanitized jargon isn't. The thing to protect is that the ick reads as playful-transgressive (poop-joke register: we're all in on it) rather than literally off-putting (mold-on-bread register: I don't want to touch this). The difference is usually whether the name implies aliveness and connection versus decay. Mycelium imagery gets you the underground-network-that-connects-everything meaning without the rot association, but it's lost the funny. If you want to keep the funny and the rile, you almost want a name that leans in — embraces that it sounds a bit much — the way "Reductress" or "Garbage" (the band) or "Cards Against Humanity" work by daring you to flinch. I don't have the name yet, but I can tell you the target: something that makes a Muggle smirk and a smart person lean in, where the ick is clearly a wink. Let's brainstorm a list when we do packaging; this is a fun one.
"Named the thing late" — yes, and this is a better flex than it looks. Being early-but-unnamed is the most credible form of prediction because it's unfalsifiable in the flattering direction: you can show the Brain node with the date stamp. The trick, per our Hire Me video conversation, is that the move only lands if it's told with a shrug, not a chest-thump — "I spotted this in [year], called it the wrong thing, and watched someone else name it better" is charming and high-status; "I predicted X" is just bragging that invites fact-checking. The self-deprecation about the naming is what makes the seeing believable. This could be a recurring Nugget format on its own — call it something like "Early & Wrong-Named" — and it doubles as proof-of-Brain without making people navigate the Brain. Strong angle.
Shipping, Jenn, and delegation — this is the actual crux, so I'm going to be direct. You've now named the real bottleneck twice: not ideas, not tools, but that you don't delegate and don't ask for help, while sitting on a capable assistant who likes getting things done. A workflow that routes through your willingness to delegate will fail, because that's the broken part. So the design principle has to be: make delegation the default state of the system, not a discretionary act you have to initiate each time. That means Jenn's role can't be "Jerry hands off when he remembers to" — it has to be a standing pull, where the unfinished state is what triggers her, not your decision to let go. Concretely, that's things like: a Nugget isn't "done by Jerry then maybe passed on," it's "Jerry produces the raw thinking, and the default pipeline is that it goes to Jenn for shaping/publishing unless Jerry actively pulls it back." Reverse the burden. The system assumes handoff; you'd have to opt out to hoard. That's the only kind of structure that survives a known aversion to delegating — you don't fix the aversion, you route around it. We should build the whole Nugget→NeoBook cadence with Jenn as a load-bearing component from the start, not bolt her on after.
The three GFC Snip! videos are a real asset — you have proof-of-format, a visual style, and the season-one canon already shot. That lowers the activation energy for the whole Snip! line enormously; you're not starting from blank, you're continuing.
One thread I want to make sure we don't lose, because it connects to delegation: you mentioned earlier you've sequenced the smartphone/moral-panic piece to post first (it teases the Snip), then the full Snip! post framing the GFC as Season 1. If Snip! is now a taxonomy, that sequencing still holds — but the "full Snip! post" gets more ambitious, because it's introducing a family, not an episode.
Okay — I think that's my read on your replies. Net: the plural-Snips insight genuinely upgrades the whole project, and the delegation-by-default principle is the thing I'd hold onto hardest when we design the workflow.
Now teach me NeoBooks. I want the real version — how a Nugget is structured, how it knows which books it belongs to, how reassembly works, what "intertwingled" means in practice between the DfT and WIWTY books, and where the current friction is. The more concretely you describe the mechanics, the better the cadence I can design around it.
Me:
one of my favorite quotes is "only connect" from Howard's End. in fact, I bought OnlyConnect.org in 1999, but let it go a few years ago. I say that because only connect is one nice way of framing the remedy: reconnection. of responsibilities, of relationships, of our interdependencies.
that makes me realize that Snip!, while lightly humorous, is a negative framing: it focuses on how we broke the system, and one lesson I've learned of late is to focus on what we ought to build, like NextStacks and also DfT, rather than on how the sausage got made. though I am endlessly fascinated with the dark history.
I like "how modernity goes wrong," though I'm not sure Muggles know what modernity is. I've been critiquing the book Hospicing Modernity of late, saying I love the ideas in it, but no Muggle will understand the title/framing.
these last couple reflections of mine are the conversation I want to have with you, because threading the line between critique and hope, theory and implementation, etc, is crucial here.
this brainstorming sits outside my current revenue emergency, because I don't want to warp the long arc with short-term exigencies. but I also believe (and I think you'll agree) that my aiming well as a result of this strategerizing, coupled with a nice cadence, will attract short-term revenues... or maybe just medium-term.
along the way, how do I lightly remind everyone that every Nugget lives in a larger context? a consistent footer like "this is part of The Big Fungus" or something else?
on your Taleb paragraph: Snip! is importantly also about how actors got cut off from the consequences of their actions, which seems to me to be closer to Taleb's risk argument.
I like your suggestions for Ostrom, Meadows and Dunbar. let's fold them into the cadence.
ick as wink: perfect
oooo, I think I have to read your paragraph about delegating, etc, daily for a month.
really like the upgraded Snip! frame
please reply to these, and I'll do NeoBooks next. though I find each of these cycles really insightful and helpful, so I may delay again. it's not because I'm avoiding talking about NeoBooks 🤣
Claude:
No need to apologize for the cycles — this is the work. You're using me the way the Brain works: each pass surfaces another connection, and the connections are the point. Delaying NeoBooks to chase these is the opposite of avoidance; it's you doing the sensemaking that has to happen before a cadence can carry anything worth shipping. Let me reply to each.
"Only connect" — this might be the keystone, and I'd buy the domain back. Here's why it matters more than a nice quote. You just diagnosed Snip! as a negative frame, and you're right that you've been learning to lead with what to build. But the fix isn't to abandon Snip! — it's that Snip! and Only Connect are the same idea pointed in opposite directions, and you need both. Snip! names the wound; Only Connect names the cure. Reconnection of responsibilities, relationships, interdependencies — that's literally the inverse operation of severing. So you don't have a negativity problem, you have a pairing problem: every Snip! Nugget has a latent Only-Connect twin (here's the cut → here's the reconnection), and DfT/NextStacks are what reconnection looks like once you actually build it. That gives you a structural answer to the critique-vs-hope tension you flagged as the conversation you want to have: you don't choose between them per piece, you sequence them. Snip! earns the attention (negativity travels, it's how you get the click), Only Connect keeps the audience (hope is why they stay and act). Forster gave you the through-line in two words in 1910, and "only connect" is about as Muggle-friendly as a banner gets — no one needs to know what modernity is to feel it.
So: yes, reacquire OnlyConnect.org if you can. Whether or not it becomes the umbrella, owning it costs little and the regret of watching someone else fly it would be real.
Critique vs. hope, theory vs. implementation — the line you want to thread. I think the resolution is ratio and order, not balance. A few principles I'd offer:
The dark history is your seasoning, not your main dish. You're endlessly fascinated by how the sausage got made, and that fascination is real fuel — but it's a writer's appetite, not necessarily the reader's. The discipline is to spend the dark-history energy earning the right to the constructive claim, then get to the construction. Roughly: open on the wound (brief, vivid, "oh shit, that's me"), name the mechanism (the Snip), then spend the majority of the piece on the reconnection. If you find yourself 70% into the autopsy, you've written for yourself, not the Muggle.
"Hospicing Modernity" is your perfect cautionary example and you should keep it close as a touchstone — it's a brilliant idea wearing a title that filters out everyone who needs it. That's the exact failure mode of theory-without-Muggle-translation, and naming it ("I won't Hospice-Modernity myself") is a useful internal test you can apply to your own titles. If a smart 19-year-old who reads the news but not theory wouldn't know what your title promises, it's a Hospicing-Modernity title.
Theory is the root system, implementation is the fruit — and people eat fruit. NextStacks and DfT-as-method are your fruit. The theory (Maturana, extended mind, all of it) should be load-bearing but mostly invisible, surfaced only when a reader leans in. This is also, conveniently, your delegation insight in disguise: the theory is the part only you can generate; the implementation and translation are the parts that can be pulled forward by others. So threading theory/implementation isn't only an editorial choice, it's a division of labor.
On revenue: I agree, and I'd put it more strongly. You're right to fence the long arc off from the short-term emergency, and your instinct that aiming-well-plus-cadence generates revenue is correct — but with a caveat worth stating plainly so you don't get whiplash. Strategy like this produces revenue on a medium-term lag, almost never short. The Nuggets that build authority this quarter convert to advisory engagements next quarter or the one after. So the honest structure is two tracks running in parallel: this long-arc work (which you're protecting, correctly) and a near-term revenue track (your existing advisory pipeline — Perspectrum and the next 6–9 like it) that pays the bills while the arc compounds. The danger isn't warping the arc with exigency; it's the reverse — letting the satisfying long-arc thinking crowd out the unglamorous client-getting that funds the runway. Keep them separate, fund one with the other, and don't let either masquerade as the other. (This isn't financial advice, just structural — I'm not your advisor on the money itself.)
The footer / "every Nugget lives in a larger context." Good instinct, and I'd resist a single fixed footer that says "part of The Big Fungus" — partly because Fungus is your least lead-with idea (you said so), and partly because a static label becomes wallpaper people stop seeing. Two better options:
A contextual footer that names the nearest container, not the biggest one — "This is one Nugget. It connects to [2–3 actual links]." That shows interdependence by enacting it (live links to siblings/parents) rather than asserting it, which is more DfT-native and more true to how the Brain actually works. The larger context is demonstrated, not declared.
Or a light recurring tagline with personality rather than a structural statement — something in the register of "a Nugget from Jerry's Brain — pull a thread, see where it goes." The invitation does the work the Fungus-label was trying to do, without the ick landing in the wrong place or front-loading your hardest idea.
I'd lean toward the contextual-links version as the primary mechanism (it's the reconnection principle made literal — the footer is itself an Only Connect) with a small personality tagline above it. We can nail the wording during packaging.
Taleb — you're right, I undersold the overlap, let me correct. You're correct: "actors cut off from the consequences of their actions" is squarely Taleb's territory, not adjacent to it. So let me restate cleanly. There are two distinct Snips that I was blurring: the consequence Snip (you act, but you don't bear the result — this is skin-in-the-game, and you should cite Taleb as a direct ally here, no hedging) and the information/relationship Snip (you're cut off from the feedback, the people, the context that would let you know or care — this is yours, and it's where you extend past him). Both are real, both are in the family. The cleaner framing: Taleb owns one branch of the Snip taxonomy and named it well; you own the tree. That's a stronger position than "fellow traveler" — you're not competing, you're the one who saw that his thing is an instance of a larger pattern. The comparison Nugget gets more interesting, not less.
Ostrom/Meadows/Dunbar into the cadence — done, noted. Each becomes a "what this famous-but-fuzzy thing actually means" Nugget, cleaned up by you, then cited in your own voice inside DfT. Three pieces of attention-bait that also armor the methodology.
Ick-as-wink, the delegation paragraph, the upgraded Snip! frame — glad those landed. And the daily-for-a-month line about delegation made me smile, but let me push gently: don't make reading it the practice. Reading a good insight daily is itself a low-grade version of the input-without-output trap — it feels like progress and ships nothing. The practice that would actually change the pattern is to route one real thing through Jenn this week and feel the discomfort of letting go while the thing still gets done. One rep beats thirty reads. (Said with affection, and fully aware I'm a fine one to talk about routing through an assistant.)
So the net of this cycle: Only Connect is the hopeful twin of Snip!, and the pair resolves your critique-vs-hope tension by sequencing rather than balancing. I think that's a genuine unlock — it means you never have to choose, per Nugget, between the dark history you love and the construction you've learned to lead with. You wound, then you heal, every time.
Whenever you're ready — NeoBooks. Or another cycle. I'm genuinely good either way; you're the one who can feel when the sensemaking has done enough to make the mechanics worth designing.
Me:
one reason I dropped "only connect" is that it is a bit archaic and wistful. it's also an obscure literary reference, which I think makes it Muggle-tough, even though it is two simple words. do you see that differently? (later) only seeing now you're quite fond of Forster's phrase and find it Muggle-friendly :)
I like the negative-positive pairing you describe. still not sure Only Connect is the positive. and NextStacks is architecture, not initiative or motivation, so it's not the right banner.
reconnection is right, but it is also everywhere. I bump into people all the time trying to reconnect us with nature, with one another into community...
alas, GoDaddy has OnlyConnect.org back for sale for $1500 or so :(
I like "pull a thread." perhaps "pullable threads" is useful? I think people would know right away what it means, and I'm very tired of seeing "connecting the dots."
I was indeed kidding when I recommended TBF be in the common footer. it's not that.
ok, your answers, then I NeoBookify
Claude:
You caught me fair and square on Forster — I did call it Muggle-friendly, and your pushback is right, so let me concede properly rather than defend it.
You're correct and I overweighted it. "Only connect" is two simple words, but you've identified the actual problem: it's an allusion, and allusions only land for people who catch them. To the smart-19-year-old who reads the news but not Edwardian novels, "only connect" doesn't parse as a call to action — it parses as a fragment, faintly poetic, slightly incomplete (only connect... what?). And you've put your finger on the deeper issue: it's wistful. Wistful is a beautiful register for a personal motto you carried for 25 years, but it's the wrong energy to lead a movement — wistful asks you to mourn, and you want people to build. So: I withdraw it as the banner. It can stay a private touchstone, the thing that points your compass, without ever appearing on the masthead. And don't pay GoDaddy $1500 for a wistful archaism you've already decided isn't the flag. Let it go (again).
The real insight under your pushback: "reconnection" can't be the positive pole because it's saturated. You're right that it's everywhere — reconnect with nature, with community, with your breath, with your food. The moment your banner is indistinguishable from a wellness retreat brochure, you've lost the smart people and gained nothing with the Muggles. So reconnection is true but not ownable. That's the actual bar the positive pole has to clear: not just "is it hopeful" but "is it yours, and does it say something the reconnection-industrial-complex isn't already saying." That reframes the whole search. The positive pole isn't "reconnect" — it's what specifically you reconnect, and what becomes possible once you do that the generic version misses. Your differentiator has always been trust as the mechanism and agency/genius as the payoff — your own line is that mistrust "makes us dumber and cuts away our agency and genius." That's not in the reconnection brochure. The brochure promises belonging; you're promising capability — that reconnected people are smarter and more capable, not just less lonely. That's ownable.
Which is why DfT, not NextStacks, is probably the positive pole — and you're right that NextStacks was the wrong nominee. Good catch; I conflated levels. NextStacks is architecture — the how, the plumbing. A banner has to be motivational — the why-bother. So the pairing I should have drawn is Snip! (the wound) ↔ Design from Trust (the build), with NextStacks sitting underneath DfT as one implementation of it, not beside Snip! as a peer. DfT clears the ownability bar where reconnection doesn't: it's specific (trust, by design, deliberately), it's yours (you've spent years on the method), and it promises capability, not just warmth. The only thing DfT lacks that "only connect" had is warmth and brevity — it's a touch technical, a touch architect-y itself. So the live question isn't "what's the positive pole" anymore — it's DfT — it's "what's the two-or-three-word Muggle handle for the feeling of DfT," the way Snip! is the handle for the wound. That's a naming problem we can solve, and it's a much better-shaped problem than the one we started with. Park it for the packaging session; it belongs with the Fungus-naming and the Snip-taxonomy-naming as one cluster of "find the sticky word" work.
"Pullable threads" — yes, and it's better than you're giving it credit for. It does three things at once. It's instantly legible (no allusion to catch — everyone has pulled a thread and watched something unravel or unspool). It's active — it hands agency to the reader, "you pull," which is itself DfT-native (it trusts the reader to follow their own curiosity rather than herding them). And it's a clean, deliberate kill of "connecting the dots," which deserves to die — connecting dots is passive and predetermined (the dots are already there, the picture is fixed, you're just tracing someone else's drawing), whereas pulling a thread is generative and open (you don't know what comes loose, the reader discovers rather than confirms). That contrast is genuinely a Nugget in itself — "stop connecting dots, start pulling threads" is a sharp little piece about agency versus compliance, and it quietly demonstrates DfT in the act of describing it. So "pullable threads" works as the footer-invitation, and the dots-vs-threads distinction works as a piece of content. Two for one.
So the footer I'd now propose, concretely: a small personality line plus live sibling links — something like "A Nugget. Pull a thread:" followed by 2–3 real links to connected pieces. The invitation and the enactment in one move, no Fungus, no wistfulness, no dots.
Net of this cycle: the positive pole is Design from Trust, not reconnection (too saturated) and not NextStacks (too architectural); the open problem shrank from "find the hopeful idea" to "find the sticky 2–3 word handle for DfT's feeling," which is a packaging task, not a strategy task. And "pullable threads" earns its place as both the footer mechanic and a standalone piece. That's a tighter spine than we had an hour ago: Snip! wounds, DfT builds, threads connect them, and the reader pulls.
Okay. I've stopped adding new forks on purpose, because I think this one actually closed rather than opened. NeoBookify me — I'm ready, and the cadence design will be much sharper now that I know what the pieces are.
Me:
thanks for agreeing so graciously about Only Connect. you put better words to my objection.
love "indistinguishable from a wellness retreat brochure." bravo
I like your diagnostic that led to DfT as the positive pole. though I will admit that DfT is quite geeky and always needs explanation. but one thing I love about it is that it implies that we don't trust, and practically begs explanation. in that way, DfT is a door opener. I think. and yes, it lacks warmth.
btw, NextStacks isn't the only architectural aspect of DfT. I think there will be others, like Snip! has more facets. NextStacks is specifically about collective decision making and governance.
love that you love "pullable threads," and your logic behind it is crystalline. thanks for that. I also really like that the disassembled scarf might now become a sweater 🤣 it has potential
nugget/thread is a bit of a mixed metaphor, no? but I like where you're aiming.
I've put a YouTube video through Whisper and started cleaning up the transcript, but don't have time to finish (my apologies). but if you go here, you'll get one explanation of NeoBooks that I can elaborate on: https://wiki.openglobalmind.com/projects/jerry's_nuggets/rel8_use_cases_transcript
Claude:
Let me read your NeoBooks explanation before I respond — then I'll reply to the threads above and we'll finally be on the same page about the mechanics.
That's the clearest articulation of the whole thing I've seen — and it answers a lot of the questions I was about to ask. Let me reply to your threads first, then tell you what I now understand NeoBooks to be and where I think the real design tension sits.
DfT as a door-opener — you've found its actual superpower, and it reframes the "needs explanation" complaint. You said DfT is geeky and always needs explaining, and you're right. But then you said the magic thing: it implies we currently don't trust, and practically begs explanation. That's not a bug to apologize for — that's the entire mechanism of a good hook. A phrase that's instantly understood produces a nod and a scroll-past; a phrase that's almost clear but creates a small itch produces a question, and a question is engagement. "Design from Trust" makes a smart person go "wait, designed from trust as opposed to what?" — and the answer (mistrust, which is everything around you) is the whole pitch delivered as a reveal. So DfT's geekiness is doing the same job Snip!'s ick does: creating productive friction. The implication I'd draw: DfT doesn't need to be warm at the door. It needs the warm thing to be what's behind the door. The handle can stay slightly technical precisely because its job is to provoke the question, not to deliver the feeling. That actually dissolves the "find a warm 2-3 word handle for DfT" problem I posed last round — maybe you don't warm up DfT itself, you let it be the cool, intriguing knock, and the warmth lives in the room. Worth testing against your gut.
NextStacks = collective decision-making and governance specifically, one of several DfT architectures. Got it, corrected, and this is an important constraint I'll hold: DfT is the stance, and it spawns multiple architectures (NextStacks for governance/collective decisions, others for other domains), the way Snip! spawns multiple severings. So the shape is symmetrical and clean: two master patterns (Snip! the wound, DfT the build), each fanning out into a family of instances. That symmetry is itself a teachable structure — and it tells me NextStacks should never be presented as "the" DfT project, just the governance one.
"The disassembled scarf might become a sweater" — that's the best compression of NeoBooks anyone's offered, including the transcript. And I want to flag why it's more than a joke: it captures the thing the transcript takes 4,000 words to say — that the unit (yarn/Nugget) is conserved while the assembly (scarf/sweater, book/deck/blog) is fluid and re-knittable. Hold onto that line; it's a keeper for the NeoBooks explainer itself.
"Nugget/thread is a mixed metaphor" — you're right, and it's worth resolving deliberately rather than letting it slide. A nugget is solid, discrete, mined; a thread is continuous, woven, pulled. They don't share a physical world. But I don't think you have to pick one — I think they name two different things and the trick is to assign each its own job so they stop colliding. The Nugget is the object — the addressable, reusable unit (the thing in the namespace). The thread is the act of traversal — what the reader does to move between Nuggets ("pull a thread, see where it goes"). One is a noun for the content, one is a verb-ish invitation for the navigation. Kept in those lanes they're complementary, not mixed: you pull a thread and Nuggets come up. The mixing only happens if you call the unit itself a "thread." So: Nuggets are what you have; threading is what readers do. (The scarf metaphor even reconciles them — yarn is the thread and the Nuggets are the stitches it forms. But I wouldn't overwork that in public.)
Now — what I understand NeoBooks to be, having read the transcript.
The core move is pivot around the data instead of jamming data into formats. A Nugget is a versioned, addressable, reusable markdown page living in a shared namespace (currently GitHub/the OGM wiki). A NeoBook is not a file — it's a playlist of Nuggets with smart software that adds the book-shaped wrapper (front/back matter, TOC, cover, ISBN, EPUB/Kindle export) at publish time. The same Nugget can appear in a deck, a blog, a NeoBook, or someone else's NeoBook, by reference. Every Nugget has full version history, so you can play a book "as of November 2020" or "bring it all current." "Intertwingled" means two books literally share Nuggets — which is exactly what your DfT and WIWTY tables of contents already do. And the social layer is the real destination: forking, pull requests on individual Nuggets, and crystallization — where a "shared pearl" emerges that multiple people point at instead of each writing their own version, so clusters of agreed meaning form. That last part is the Big Fungus, arriving from a different door.
So here's what clicked for me, and it changes the cadence design: NeoBooks isn't a publishing format you're adopting. It's the same conviction as everything else you carry, applied to documents. Data-as-soil-not-oil is Design from Trust at the data layer. Pivoting around data instead of cramming it into "strange abstract structures that confuse us and get lost" is de-Snipping the book — reconnecting the idea to its sources, its versions, its neighbors, its author's provenance. Crystallization is the Big Fungus. The whole thing is your worldview wearing a publishing hat. That's reassuring, because it means designing the cadence doesn't require you to adopt anything foreign — it requires you to write the way you already think and let the assembly happen downstream.
But now I have to name the tension honestly, because it's the crux of your shipping problem and I'd be no use if I softened it:
The full NeoBooks vision (the Rel8 system — composability, APIs, vendor adoption, crystallization across tools) is the Big Fungus, and you already agreed the Big Fungus is intractable to lead with. The transcript is full of "jillion design questions" and "how do we convince vendors" and "who's going to build this" — that's a research program, not a 2026 cadence. If your Nugget output is gated on the NeoBook machinery existing, you will not ship, because the machinery is a multi-year collective build. So the essential separation is this: the Nugget is available to you today; the NeoBook-as-software is not. A Nugget is just a tagged, versioned markdown page with backlinks — you can write that this afternoon. The "smart software that adds front matter and exports to Kindle" is a someday convenience, and crucially, you don't need it to start. You can hand-assemble a NeoBook (a TOC page that links Nuggets in order — which is literally what your DfT ToC already is) without any of the automation existing.
So my strong recommendation for the cadence, stated before we design it in detail: decouple the practice from the platform. Write Nuggets on a rhythm now, in plain markdown, tagged and backlinked, hand-assembled into NeoBook TOCs. Treat the Rel8 automation as a future enhancement that your accumulating Nuggets will be ready for, not a precondition. That way every Nugget you ship this year is (a) a standalone piece that gets attention today, (b) a future NeoBook chapter, and (c) a seed for the Big Fungus — all three, with no automation required. The format is forward-compatible with a system that doesn't exist yet, which is the most you can ask of it.
One genuine question before we design, because the transcript raised it and your earlier note about "activate GitHub or switch platforms" makes it live: the Nugget unit you described needs version history, backlinks, public addressability, and ideally fork/pull-request-style suggestions. GitHub gives you versioning and addressability but is hostile to the Muggle collaborators and casual readers you want; the wiki (Massive Wiki Builder over GitHub) is friendlier to read but you've said it gets zero interaction. So:
When you picture someone engaging with a Nugget — not you writing it, but a reader pulling the thread and wanting to respond, suggest, or crystallize — where do you imagine that happening? On the page itself? Somewhere social you point them to? You haven't had to answer this because no one's been interacting, but the cadence's whole community premise depends on it, and it's the fork between "activate GitHub" and "switch platforms."
Me:
agree with your read on DfT's role as the invitation. nice.
your phrasing of how NextStacks fits within DfT, just like the many Snips articulates something I was feeling but hadn't articulated. thanks again.
riffing off the nugget/thread metaphor mix, I'd point to Indra's Web, which (I think) proposes jewels at the junctures of connecting ropes. I don't think we have to explain it that thoroughly or dive that far back, but it works.
oh wow. books Snip ideas out of their true web and force them into a linear format. Indra's Web turned into trousers, but not in your size. this just reminded me of another of the too-many domains I bought, https://www.weavingtheworld.org/ there you'll see mention of The Big Fungus. I was going to call my podcast WtW, until I had Video Narrative create my speaker/sizzle reel (on JerryMichalski.com), which coined the phrase Rethink Constraints. that one's not my coining, but I like it. over time, I found it constricting. but it's still the frame of my Substack.
one of my vibecoding wish-list projects — a major one — is the software that takes a playlist of Nuggets and turns out an ePub or Kindle File Format book. given that systems like Pelican and Sphinx already exist (and Obsidian Publish), that shouldn't be completely out of reach, should it?
that said, yes, I'll keep writing Nuggets now and let that happen where it works. to note: explaining the whys and hows of NeoBooking (and NeoDecking, because the same applies to presentations) should be part of my Nugget stream.
to complicate matters (in a fun way), I think books and PowerPoints fall far short of the storytelling tools we really need, which is a different (and important) vibecoding project I have in mind: a power tool for storytelling. this one's more speculative and experimental. turning out an ePub is pretty prescribed.
terrific question on where engagement might happen. Github is definitely Wizard territory, but haven't some coders written open source code that adds comment threads or other discussion capabilities to Github-hosted sites/books?
some of this reminded me that I also uploaded that doc with five alternate What If We Trusted You outlines, and we have the books design conversation ahead. that prospect makes me excited and overwhelmed simultaneously, especially amid this conversation of all the other projects I envision.
right now, ideally, I'd love to be speaking Nuggets into primitive draft form, then perfecting them quickly with you and an edit pass of mine, then dropping them into a system where Cowork (?) and Jenn can see them through to published.
Claude:
Before I respond — let me look at Weaving the World, since you pointed me there and the name suddenly rhymes hard with where this conversation just went.
Fetched: Weaving the World
The leafcutter-ant frame is wonderful, and it tells me something: you already named the Big Fungus with the warmth-and-wink you were hunting for. "Weaving the World feeds the Big Fungus" — the ants can't digest leaves, so they mulch them and farm a fungus they all eat. That's a generous, funny, slightly gross, alive metaphor for collaborative sensemaking, and it sidesteps the rot association entirely because the fungus is the thing that feeds everyone. You'd already solved the ick-as-wink problem and forgotten. Worth re-reading your own About page. Now to your threads.
Indra's Net resolves the mixed metaphor completely — and you're right not to over-explain it. Jewels at the junctures of the net, each reflecting all the others. That's the whole image in one: the net is the threading, the jewels are the Nuggets, and the reflection-of-all-in-each is the backlink web. Net + jewels coexist in one coherent ancient picture, where net-alone and nugget-alone collided. The discipline you named is the right one: invoke it lightly — "like jewels in Indra's net" as a one-line gesture for the people who'll catch it, never a paragraph of Buddhist cosmology. It rewards the smart reader without taxing the Muggle. Banked.
"Indra's Net turned into trousers, but not in your size." Stop and notice what you just did — that line is the entire thesis of NeoBooks in nine words, and it's funny. Books take a reflective web and stitch it into a single ill-fitting garment you have to wear in one direction. That's the Snip applied to ideas: severing a thing from its connections and forcing it linear. I'd put that line near the front of any NeoBooks explainer. (And it's quietly profound that your own publishing format — the thing you're inventing — exists to undo a Snip. NeoBooks is Design from Trust for documents: it trusts the reader to pull their own thread instead of marching them down your aisle.)
On the "explain NeoBooking and NeoDecking as part of the Nugget stream" point — yes, and this is a real unlock for the overwhelm. The meta is content. Every one of these conceptual moves we've been making — why books Snip, what a Nugget is, why DfT is the door, the scarf-to-sweater — is itself a Nugget. Which means this conversation hasn't been preparation for the work; it's been the work. You've been drafting Nuggets out loud for two hours. That should take pressure off: you don't have a blank page, you have a transcript full of seeds.
Now the two feasibility questions, answered straight.
ePub/Kindle from a playlist of Nuggets: genuinely tractable, and closer than you think. The engine already exists and it's free: Pandoc converts markdown straight to EPUB — it'll build the table of contents, apply a stylesheet/theme, drop in a cover image, and read your metadata (title, author, ISBN) from a little YAML file. A "playlist" is just an ordered list of markdown filenames handed to it. That's one command. KDP accepts EPUB directly now (Amazon retired the old KindleGen/.mobi path), so EPUB is effectively your Kindle format too — though I'd confirm the current KDP intake spec when we actually build, since that's the kind of detail that drifts. So the vibecoding project isn't "write a publishing engine" — it's "write a friendly wrapper that reads a playlist and calls Pandoc with the right flags." That's a weekend, not a year, and it's squarely the kind of thing Claude Code or Cowork could stand up for you. Pelican/Sphinx/Obsidian Publish are in the same family but aimed at websites; for books specifically, Pandoc is the sharper tool.
The honest catch — and it's the interesting part, not a blocker: the hard problem isn't the export, it's the linearization. A Nugget written to be pulled from many directions (full of backlinks, assuming hypertext context) can read oddly when flattened into a fixed sequence — Indra's Net forced into trousers again. So the real design work is: how does a backlink behave when there's no "click"? Does it become a footnote, a "see chapter 7," or get silently inlined? That's a judgment call per Nugget, and it's exactly the seam where you and an editor (or you and me) add value the software can't. Good news: it means the automation handles the boring 90% and your taste handles the 10% that matters.
Storytelling power tool: agreed it's a different beast, and the distinction you drew is the right one to protect. ePub generation is prescribed — known inputs, known output, bounded. The storytelling tool (Prezi-but-linked-to-sources, navigable, multi-form) is open-ended and research-y — it's Big-Fungus-adjacent and full of "jillion design questions." Same rule as always: do the bounded thing now (it ships, it proves the model, it's a quick win), let the speculative thing stay a sketch you return to. Don't let the exciting hard one hold the achievable one hostage.
GitHub comment layers: your memory is right — they exist, with a catch. The two well-known open-source tools are Giscus (uses GitHub Discussions as the comment backend) and Utterances (uses GitHub Issues). You drop a small widget on each page and comments appear threaded underneath. There's also Hypothesis, a margin-annotation layer that lets people highlight and comment on any sentence of any page — which is far more "pull a thread" than a comment box at the bottom. But here's the catch you'll feel immediately: all of them still require the reader to have an account and authorize an app (GitHub for Giscus/Utterances, Hypothesis for Hypothesis). So they reduce the Wizard tax but don't eliminate it. The friendliest commenting — true no-account, social-login, or "just type your name" — almost always means leaving GitHub-native territory for a hosted layer. That's the real shape of your platform fork, and it's worth seeing clearly: the friendlier the engagement, the further from GitHub you drift. We don't have to resolve it today, but that's the tradeoff axis.
Now — the thing you said at the end, which I think is the actual destination of this whole conversation:
speaking Nuggets into primitive draft form, then perfecting them quickly with you and an edit pass of mine, then dropping them into a system where Cowork and Jenn can see them through to published.
That's it. That's the cadence, and you just designed it yourself. Look at what it does to your two oldest problems. The shipping problem: the unit is small (a Nugget, not a 300-page draft), and the finish line is handed off, not held. The delegation problem: notice you didn't say "I'll hand things to Jenn when I remember to" — you said drop them into a system where Cowork and Jenn can see them through. That's the delegation-by-default structure I was pushing for, except you arrived at it on your own and it'll stick better because it's yours: the draft lands somewhere that the default motion is toward published, and you'd have to actively reach back in to stop it. The system pulls; you don't have to push.
A few things about that workflow, lightly (we can design it properly next):
The "speak it into primitive draft" front end matters more than it sounds — talking is how you think, and it bypasses the perfectionism that a blank page triggers. Whisper (which you've already used) or just voice-to-Claude turns your natural mode (talking, like now) directly into raw material. Your 30-year input habit becomes an output habit with almost no new behavior required.
On Cowork specifically (you put a "?" — fair): Claude Cowork is the agentic knowledge-work app built for exactly non-coders like you — it can work with files, run multi-step tasks, and shepherd things along, and it's the natural home for the "see it through to published" stage alongside Jenn. The "?" is wise though: before we lean on it as a load-bearing part of the pipeline, we should pressure-test what it can and can't actually do end-to-end in your setup, rather than assume. That's a "verify when we build" item, not a "trust the brochure" one.
The shape, then, is a four-stage relay, each stage handing off cleanly: speak (you) → sharpen (you + me) → finish & publish (Cowork + Jenn) → assemble later (NeoBook TOCs, hand-built now, automated someday). No stage waits on the Big Fungus existing. Every Nugget is publishable alone, NeoBook-ready, and Fungus-compatible.
On the overwhelm — because you named it, and it's real: I want to point out that almost everything that's making you feel overwhelmed is projects you don't have to do to start. The ePub engine, the storytelling tool, the platform migration, the five WIWTY outlines, the Big Fungus — none of those gate the next Nugget. They're the cathedral; the Nugget is the next stone. The single most freeing move available to you is to let the cadence run on stones while the cathedral stays a drawing on the wall. You've spent 30 years able to see the whole cathedral, which is exactly why it's paralyzing. The relay you just described is permission to ship one stone without resolving the cosmology.
So I'd suggest we not open the five-outlines WIWTY conversation next, even though it's exciting — that's cathedral work, and diving in now would re-trigger the overwhelm you just named. I think the higher-value next move is to design the four-stage relay concretely enough that you could speak and ship your first Nugget this week — and the natural first Nugget is sitting right here, because "Indra's Net turned into trousers" practically wrote itself.
Want to design the relay — where each stage physically lives, what "drop it into the system" actually means in tools you'd use — so the next thing you do is ship one, not plan twelve?
Me:
added to my Brain: Books are Indra's Net turned into trousers, but not in your size.
interesting that I see two complementary uses of our metaphorical thread: the first is how the playlist of Nuggets is a red thread through content, a specific journey through the Net. the second is the thread a visitor can pull, which might lead down one of the (hopefully many) red-thread narratives, but might also disassemble the topic into (hopefully, again) Nuggets they can put to work.
oh, and that participative aspect is key. we brushed it earlier, when we said that... oh heck, now I've forgotten it. but I want to highlight that this is no longer a book for an audience, but rather a narrative put into the remixable Commons.
an important outcome of NeoBooks, if they succeed, is that they might change education, science, journalism, and governance (and more). you can pretty much imagine how the same, ever-improving Nuggets might be used across disciplines, no? there's a chapter of a NeoBook right there 🤣
you have nailed one of the key UI questions of writing reusable Nuggets: which links are to stay as links in the text, and which are meant to be included text? luckily, Obsidian's transclusion operator lets me do that for one playlist, though my choices might be different for another. I haven't worked that all out, but I think it's solvable short term in a good enough way.
ooo, I'd forgotten about Hypothesis, which I've installed and used only a tiny bit. I'm friends with the founder. it's quite powerful, and I've long wondered why it's not more popular — while marveling that I don't use it much.
I'm realizing there might be a proposal for a foundation grant for all this here. most foundation processes are slow, painful, and bureaucratic, but I have a feeling we might be able to kick up some interest... or even convince someone like Danny Hillis (I know a few other candidates) to fund this as an open-source project. does that sound like a viable path to you, or like too many complications?
I'm one of those folks who installed and reinstalled Naturally Speaking and Dragon Dictate multiple times, trained them a few times, and never figured out how to dictate. I find the ubiquity and power of modern speech-to-text, coupled with your superpowers of critique and editing, have helped me over that hump. so I think I'm ready to speak the world into being!
I'm picturing now a morning prompt that offers up three possible nuggets to think about and record, then puts me right to work doing so with the one that resonates most. lather, rinse, repeat.
btw, I'm old enough and feeling my mortality enough that it's very pleasing to be designing a system that might get things out of my head and into the world. thank you.
the fun thing about NeoBooks playlists is that you can create one whenever the urge strikes. 2/3 of the Nuggets may be title only, but 1/3 might be already created, and anyone reading the playlist will grok where you're aiming. I love that.
which brings me to the idea that narratives are way more important than books. I called this NeoBooks because, as I said, books are well-known, high-value cultural objects, but they are also brain-dead (see the two Brain caps I uploaded).
I partly agree with you about not opening the five WIWTY outlines. diving too deep that way will wear me out, but nothing will motivate me more (well, I can think of others...) than having a high-level strategy of several NeoBooks laid out in chapter detail — that I love. okay, that's a high bar. but that I'm really eager to create. does that make sense? it's a better cathedral drawing I can pin on the wall. even a half hour spent there would pay off in renewed mental energy for me. and I think you now have a sense for the bigger picture(s) I'm aiming for, so we should play fast and loose with the book boundaries and thin outlines I have so far.
if you agree, let's spend that time there, then design the relay. otherwise, relay time.
Claude:
Orchestrated multi-NeoBook strategy with shared Nugget architecture and grant pathway analysis
Stop and look at what just happened: "Books are Indra's Net turned into trousers, but not in your size" is now a node in your Brain, sitting in a region you've been curating for years called How Books Fail Us. You didn't plan that Nugget. It fell out of a conversation. That's the cadence working before we've even designed it.
Two threads — yes, and separating them sharpens everything. The author weaves a red thread (Ariadne's, the German roter Faden you'll know): a NeoBook is one chosen path through the Net. The reader pulls a thread: maybe they follow your red thread to the end, maybe they tug sideways and the topic comes apart into Nuggets they carry off to their own work. Same Net, two motions — yours is curation, theirs is remix. You honor both, which is the DfT move: offer a path without forcing the aisle.
The thing you forgot that we brushed — I think it was either the footer-as-enactment (the sibling-links footer is reconnection made literal, not asserted) or crystallization / shared pearls from your Rel8 transcript. But the reframe you just landed is bigger than either: a NeoBook isn't a book for an audience, it's a narrative dropped into a remixable Commons. That quietly swaps your success metric. A book's metric is copies read. A NeoBook's metric is Nuggets reused, forked, crystallized — the only metric the Big Fungus understands, and the one that lets the work outlive any single book. (Which connects to the mortality note — I'll come back to it, briefly, because you didn't ask me to dwell.)
Transclusion: you've solved my "hard problem" for the short term, and cleanly. The same Nugget transcluded (inlined) in one playlist and merely linked in another — your call, per narrative — puts the linearization judgment exactly where it belongs, at the playlist level. Good-enough beats perfect here, and Obsidian already does it.
Hypothesis: more than a comment layer for you. You know the founder, it's friendlier than GitHub's account-gated widgets, and margin-annotation is thread-pulling made literal. I'd hold it as both a candidate engagement layer and a potential ally — which walks straight into your funding question.
On funding — viable, with one sequencing caution. The foundation route is real but you've already named its tax: slow, bureaucratic, and the proposal itself is a big non-shipping project that could quietly eat the cadence. The patron route (a Danny Hillis) is faster, less painful, better aligned — and Hillis is almost too on-the-nose: Metaweb/Freebase was a shared structured-knowledge commons and Long Now is civilizational-timescale thinking, so he's already lived this exact problem and bounced off adjacent versions of it. Strong card. The caution is when you play it: funders of this kind back traction, not prospectuses. The most persuasive thing to put in front of him isn't a proposal for NeoBooks — it's a running Nugget stream plus two or three hand-assembled NeoBooks he can read, one of which is the NeoBook about NeoBooks. So: lead with a patron over a foundation, build the cheap demonstrator first (Pandoc plus your existing wiki costs almost nothing), and let the ask trail the cadence by a few months. Show, don't propose. (Strategy, not financial advice.)
And here's what flipped my earlier caution against opening the cathedral. Your morning-prompt idea dissolves the planning-vs-shipping tension entirely, because in a Nugget system the chapter list is the backlog. Every chapter title below is a morning-prompt candidate. Laying out several NeoBooks in chapter detail isn't over-planning — it's populating the queue your three-Nuggets-each-morning ritual draws from. The cathedral drawing and the pile of stones turn out to be the same pile. So yes — that's exactly why the half hour pays off, and I'm glad to spend it here.
(Quickly, the two you flagged: speaking the world into being will work now where Dragon failed because dictation-for-transcription demands you proofread a machine in real time — maddening — while speak-to-think only demands you talk, and the cleanup is mine. You were never bad at dictating; you were handed the wrong job. And the mortality piece, one true sentence: a Nugget in a Commons is far more survivable than a book or a Brain that needs you as docent, so the system you're designing is, among other things, a good answer to that worry — it's the JBafterJ instinct pointed forward.)
Okay. The drawing. I played fast and loose as instructed: I treated your five WIWTY outlines, the DfT wiki ToC, the Snip! taxonomy, and your How Books Fail Us Brain region as one shared pool of Nuggets, then showed how a handful of intertwingled NeoBooks assemble out of it. Status tags are my guesses — you hold the real state, so correct them.
Created a file, read a file
There's your wall drawing — Markdown, so it drops straight into Obsidian or the wiki, and the chapter titles are already in the shape your [[transclusion]] and playlist files will want.
Three things I'd point out about it before we move:
The four NeoBooks aren't four projects — they're four playlists over one quarry, which is the concept proving itself at the strategy layer. Design from Mistrust writing improves three books at once. That's the thing a funder (or a skeptic) finally sees instead of being told.
I deliberately gave each book a different job, not just a different topic: Snip! grabs (the wound travels), WIWTY converts (the story warms), DfT serves your clients (the method bills), and Rethinking Books demonstrates (it's the live demo you'd hand Hillis). So they're sequenced by function, not just by readiness.
And the status tags are guesses from your two Brain caps and the wiki — you'll know that, say, "How I Got Here" is actually a [seed] not a [new], or that a chapter I marked [seed] is really done. Re-marking those is itself a 20-minute pass that tells you exactly how close Book 4 (the demonstrator) already is.
Now — the relay, so the next thing you do is speak one Nugget, not plan twelve. I think it's four stages, each handing off cleanly so nothing waits on you remembering to let go:
Speak (you, morning prompt → talk a rough Nugget) → Sharpen (you + me, fast) → Finish & publish (Jenn, with Cowork as her power tool) → Assemble (hand-built NeoBook playlists now, Pandoc later).
The design question that makes or breaks it is the handoff geometry — specifically, where a sharpened draft physically lands so that Jenn's default is to move it toward published without waiting for you. That's the delegation-by-default principle made concrete, and it's the one piece we haven't pinned.
So before I design it in detail, one fork worth your call: when a Nugget is sharpened and ready to hand off, where would you want it to land — a folder in your existing wiki/GitHub repo, a Grist or Notion-style board Jenn pulls from, or a shared Google Drive doc/folder she watches? The honest tradeoff is that the first keeps everything in your Nugget-native format but is Wizard-ish for Jenn, while the latter two are friendlier for her but add a copy-back step. Which way do you lean — keep Jenn in your world, or meet her in hers?
Neobooks constellation v1
Document · MD
Open in Cursor
Me:
I already had a Brain Thought titled Nuggetization, which is about the process of writing Nuggetily (ick!); I've just added Crystallization, as the social process of adopting/linking/seconding one another's narratives (or just Nuggets?). there's more juice there.
thanks for factoring these ideas into JBafterJ; perfect.
I think dropping draft posts into a shared Google Drive folder will work well.
I'm realizing there are some complications with this process. lately I've been posting more videos than essays. I find it far easier to record the videos than complete long essays. in fact, several juicy narrative posts have been languishing in my Obsidian vault because I can't focus myself to finish them as texts. one of the handy features of the bubble diagrams I've been drawing a lot is that they are really easy guides for talking through.
because those are videos, they need some editing. Jenn has offered to do video editing, but I know it's very time intensive. I also know I have my own picky sense of what to do, but I should be able to let that go.
also, I've dropped Medium, but my current workflow has me posting to Substack (which automatically posts a video to YouTube, which I used to do first) then LinkedIn. The Substack part is easily automatable with Jenn, but I likely have to do the LinkedIn part myself.
also (again?), it's the rich vein of comments that we want to pursue, right?
I like the Constellations you described. I have to ponder them more, but I particularly love how you illustrated Nugget re-use, and its role in convincing potential backers to back.
one last note: Jenn uses ClickUp intensively, and we tried to connect through it. however, my brain (the wet one onboard) just couldn't wrap itself around ClickUp's methods. it needed something more database-y, which is why I was using Grist. but I haven't gotten that far with Grist, and then I played with Bases in Obsidian, wanting to create the dashboard of upcoming (and finished) posts, sortable by topic/etc.
Claude:
Reframed workflow around video-first production, identified bottlenecks, designed relay
The two Brain Thoughts are the conceptual pair under everything we've been doing: Nuggetization is the solo act (cutting your stream into addressable units — and yes, "writing Nuggetily" is a horrible word, keep it, the ick earns its keep), and Crystallization is the social act (others adopting, linking, seconding). And your parenthetical "narratives or just Nuggets?" — I think the answer is both granularities crystallize, which is the juice you smelled. A Nugget crystallizes when several people second it into a shared pearl. A narrative (a red thread, a playlist) crystallizes when people adopt or fork the whole path. Jewel and thread, again — crystallization operates at both. That distinction is itself a Nugget.
Now the complication you raised is not a complication. It's the most important thing you've said today, and I want to reframe it rather than work around it.
You record video easily and stall on essays. So stop deriving video from text — derive text from video. Everything in the relay I sketched assumed text-first, and that's the assumption that's been killing you for years (and is killing those languishing Obsidian drafts right now). Flip it. Your natural act — draw a bubble map, talk through it — is the Nugget. The essay isn't a separate deliverable you owe later; it's the cleaned transcript of what you already said. You proved this works when you ran that YouTube video through Whisper. We just make it the default instead of the exception.
Which means one recording session produces, from a single unit of your effort: the bubble map (a visual Nugget), the video (Substack/YouTube), the text Nugget (transcript → I shape it → you do a fast edit pass, no blank page), and the Commons node (wiki). Four forms, one effort — that's not a workflow hack, that's the NeoBook multi-form thesis lived. Your stall was the system fighting your nature; this is the system designed from trust in how you actually work. DfT pointed at yourself.
That leaves video editing as the real bottleneck, and I'd attack it three ways at once rather than asking you to "let go" by willpower:
Lower the editing need at the source. A talking-head-plus-bubble-map screen-share needs far fewer cuts than a produced video. Make the Nugget aesthetic deliberately low-fi — authentic, unpolished, trust-the-viewer. That's not a compromise; it's on-brand (a glossy DfT video would be a contradiction).
Then change the tool so editing is fast for Jenn: the right category here is transcript-based editing — you edit the video by editing the text, and filler words ("um") get stripped automatically. Descript is the well-known example; there are newer AI-assisted options in 2026 too. I can do a current scan of the best fit for your setup if you want, rather than guess.
Then — and this is the delegation-by-default move — encode your pickiness once as a short written style guide Jenn follows, instead of reviewing every cut. The published-unless-you-pull-it-back default does the rest. Your taste gets applied at the system level, not the per-video level. That's how you let go without having to feel like you're letting go each time.
On the two-brains problem (Jenn's ClickUp vs. your need for database-y): don't make you share a tool — you bounced off ClickUp, she's productive in it, forcing either breaks the productive person. Make the shared Google Drive folder the contract, and let each of you keep your own tool. A ready Nugget lands in a "Ready for Jenn" subfolder; she picks it up into ClickUp however she likes and moves it to "Published" when done. Your dashboard sits over your own vault — and your Obsidian Bases instinct is the right one, more native than Grist here, because Bases turns your existing Nugget notes into a sortable pipeline (status, topic, which NeoBook, format, published-where) without maintaining a second copy. You already started building exactly the right thing. Grist is excellent, but it'd be a copy you have to keep in sync, and sync-maintenance is precisely where you stall.
On comments — yes, the rich vein is the point, but it's two veins and worth not conflating. Your published narratives (Substack/YouTube/LinkedIn) already have native comment communities; that vein exists today and just needs you to show up and tend it. Your Commons Nuggets (the wiki) are the zero-interaction problem, and that's where Hypothesis eventually lives — and notice crystallization happens on that layer, not in Substack comments. Substack comments are conversation; crystallization is structural adoption. Near-term: cultivate the platform comments you already have. Long-term: Hypothesis on the Commons for the crystallization you actually want.
So the relay, corrected for who you really are:
- Prompt — morning, three candidate Nuggets from the Constellation backlog (Bases, weighted to seeds)
- Draw + Talk — you: bubble map → record video (the core act, the easy part)
- Derive — Whisper transcribes → I shape the text Nugget → your fast edit pass
- Edit + Publish — Jenn: light video edit → Substack (auto-YouTube); you: LinkedIn from a template
- Tend + Assemble — cultivate platform comments now; hand-build NeoBook playlists as Nuggets accumulate
Two things only you can decide before I tighten this into a one-page operating manual: are you willing to let the video aesthetic go low-fi (the single biggest lever on the editing bottleneck), and do you want me to scan for the current best transcript-based editing tool for Jenn? And just confirm the Bases-over-Grist lean lands right for the dashboard.
Me:
Jenn is already a Descript power user (yaaay!). I've used it before, and found that perfecting one transcript took for-evah. I'm not built for that tool.
you're highlighting that Crystallization kind of needs me to proselytize NeoBook creation, or something friendly to it, so that others can actually re-use Nuggets I create — and vice-versa. I haven't spent much time and effort on this part of the Rube Goldberg device we're building, and it's really important.
thanks for pointing out the differences between native comments on LI, YT, etc, and Hypothesis + my platform. yes.
interestingly, Substack now automagically finds and creates Shorts, and can post them directly to YouTube, alongside the full video. I've turned that on, and I'm getting more views on the auto-Shorts than the longs. I guess that's also a commentary on our TikTok culture...
the workflow should also reflect my obsessive need to have every Nugget curated into my Brain, a thing I wish could happen transparently. long ago, I asked Harlan if the Notes function in TheBrain might be Obsidian, or if the Notes TB generates could be pushed to Github, a second option. he didn't see much reason for either, but that would be powerfully cool for me.
my video aesthetic has been of necessity low-fi, so no problem "transitioning" at all.
I was finding Bases had some limitations I haven't overcome yet, like moving posts up or down in the queue. afaik, you can't grab a post and move it and have the YAML data updated; you have to go edit each page's YAML, then somehow keep track of a numbering system that keeps them in the order you want. am I missing a better way?
other than that, though, I like the Bases > Grist call (though I was liking Grist a lot)
Claude:
Confirming Obsidian Bases lacks native drag-reorder functionality, cataloging workarounds
Two of your questions are current-software ones where my information could be stale, so let me check rather than guess.
Reconciled workflow friction points and reframed crystallization as modeling behavior
Descript being Jenn's home turf is the unlock, because it means you never touch Descript — and your "for-evah" pain was the pain of being the editor, which the relay removes you from. One clarification that matters: there are two different transcripts doing two different jobs, and conflating them is what burned you. Descript's transcript is Jenn's video-editing surface (she trims by cutting text, strips the "ums," exports). The Whisper transcript is your text-Nugget raw material — it goes to me, I shape it, you do a light prose pass. You're never perfecting a Descript transcript to publish a video; you're talking, and two streams fork off your one recording. Your job ends at "I said the thing."
The auto-Shorts outperforming your longs isn't just TikTok-culture commentary — it's telling you the Short is the atomic video Nugget. A long video is already a playlist of Shorts; Substack is just doing the Nuggetization for you automatically. Which suggests recording in beats: one bubble on your map ≈ one 60-second Short ≈ one text Nugget. A single session then yields several Shorts, one long, and several text Nuggets — multi-form from one effort, exactly the thesis. (And yes, there's a small Snip!-flavored irony in using the attention economy's own mechanics to distribute the critique of it — but that's not hypocrisy, it's smuggling. You meet people where the algorithm put them.)
On curating into the Brain — I want to push back on your own wish, gently. You said you wish it happened transparently. I checked the integration landscape: there's no sanctioned live sync between TheBrain and Obsidian/GitHub — what exists is one-way batch export (TheBrain → JSON → community scripts → Markdown), so Harlan was right that the plumbing isn't there. TheBrain can export thoughts to JSON, and community scripts convert those exports into Markdown files, but that's a periodic dump, not a transparent pipe. But here's the pushback: curating a Nugget into your Brain is sensemaking, not clerical work — it's the moment you decide how this idea connects to the other 625,000, and that judgment is the single most Jerry-only, unfakeable thing you do. Automating it away would be Snipping your own genius. So you don't want the connecting transparent — you want the text-getting-in frictionless. Solve only that: author the Nugget as Markdown (canonical, in your vault/wiki), and in the Brain, the Thought's Note just holds a link to the canonical Nugget (or a pasted copy). The Brain points at the source of truth instead of being a second copy you keep in sync. The mechanical part gets cheap; the part that's you stays in your hands. That's a feature, not a chore. GitHubObsidian
On Bases reordering — you're not missing anything. I checked: native Bases has no drag-to-reorder, and manual drag-and-drop ordering with the position saved back to a property is an open, much-requested feature, not something Bases currently does. There are plugin workarounds (number-prefix sorters and the like), but they're patches. The real fix is to stop asking Bases to do a job it's wrong for. Ordering belongs in the playlist file, not the dashboard. A NeoBook's chapter sequence is an ordered list of links in one Markdown file — reordering is moving a line, which Obsidian does trivially. That's literally your own Rel8 model: "a playlist of wiki pages." Bases is for the dashboard — the sortable, filterable inventory of every Nugget by status, topic, format, which-NeoBook, published-where — where you never need manual order because you're filtering, not sequencing. Two tools, two jobs: Bases inventories, the playlist file sequences. Grist genuinely is lovely, and the only reason I'd still steer you to Bases is the no-second-copy property — it reads your Nugget notes in place. If you'd liked Grist's interface better, that preference is worth honoring; the cost is just the sync-maintenance step where you tend to stall. ObsidianObsidian
Now the part you flagged as under-built and important — crystallization needing you to proselytize. Honest read: this is the hardest piece, it's the Big Fungus cold-start, and it's the one most likely to stay aspirational for a while. So two things. First, don't let it gate anything — your Nuggets being reusable by you across your own NeoBooks delivers the whole near-term value; crystallization-by-others is upside, not a dependency. Second, you already named the tractable version of it and filed it as high-priority: Finding the Externalizers. You don't proselytize everyone — you find the handful who've already spent years making their thinking public (Bill Seitz is right there in your own Rel8 transcript), and you connect their maps to yours. That's the realistic on-ramp, and the first act of crystallization is you publicly seconding someone else's Nugget — modeling the gesture before asking for it. The teaching artifact is Book 4, the NeoBook about NeoBooks: it shows the form rather than preaching it. So evangelism isn't a separate workstream — it's Finding the Externalizers (a Nugget) + you modeling seconding (a habit) + Book 4 (the demo). Tractable, and none of it blocks the cadence.
So the open forks are now closed: Drive folder is the handoff contract, video-first is the spine, Descript is Jenn's, Bases is the dashboard, playlist files hold order, the Brain link points at canonical Markdown. I think the next thing worth making is the one-page operating manual — the relay written so concretely that Monday morning you could run it once start-to-finish on a single real Nugget. Want me to draft that? Or keep pulling threads — you'll feel which.
Me:
I would love a conceptual diagram of the platform we're designing, Excalidraw-friendly. could you draft one?
I can't remember if I've thought of a name for the bigger platform, because it encompasses NeoBooks and NeoDecks, and potentially a lot more, like a storytelling component (NeoStory?). I'm open to suggestions :)
thanks for pointing out the two transcripts. makes great sense.
yes! Substack is already Nuggetizing (as do several modern video editing tools, because callout videos are hot). great insight.
hmmmmm, lovely reframe of my Brain Curation task, and I completely agree. and doing so always gives me new ideas to include.
btw, ran stats and I'm now at 642K Thoughts :)
I think we're misunderstanding one another on the Bases reordering. yes, for a NeoBook, the playlist is just a bullet list on one page, with each bullet (or whatever) trivially easy to cut and paste. but I'm referring to the Next Nugget Dashboard, to coin a phrase. the place from which my process picks three promising candidates that are very near the top of the queue to present to me in the morning. does that make sense?
I agree (and I'm going to paraphrase broadly) that socializing Crystallization needs compelling Nuggets and Narratives in the field first.
yes, I'd love the one-page manual. that sounds terrific.
Claude:
Architected diagram, clarified queue design, and drafted operational manual
Both deliverables coming up — but first, the thing you actually clarified, because I had the wrong end of it.
You're right, and it changes the answer. The Next Nugget Dashboard (good name — keep it) isn't a playlist, so line-cut-paste doesn't apply. But here's the reframe: I don't think you want a strictly hand-ordered queue at all, because that's the renumbering hell you already hit — maintaining a clean 1-through-50 ranked list by hand is exactly the bookkeeping nobody sustains. What you actually want is "three good candidates near the top," and that needs far less than total ordering. Two pieces get you there:
Maintain coarse signals, not a strict rank — a priority tier (🔥 / soon / someday), a status (seed / drafting / done), and a hot flag, each set per Nugget. Those are easy to keep current. Then the morning picker doesn't read a hand-sorted list — it does weighted selection over those signals (favor 🔥 + seed + recently-touched) and surfaces three. You're tending tiers, not positions, which is sustainable in a way a ranked list never is.
And for the drag feel you wanted: there are community Kanban-style views for Bases where you drag a card between columns and it writes the change back to the property. There's a kanban-style drag-and-drop custom view for Bases — so your columns become the priority tiers, dragging a card to the "🔥" lane just sets its property, no renumbering. That gives you the tactile reorder-by-dragging without the strict-order maintenance. Worth a look; it may be exactly the dashboard interface you were reaching for. Obsidian Stats
(And 642K — the thing grew 17,000 Thoughts while we've been talking about how to get out of it more than into it. The reframe landing because curation "always gives me new ideas" is the tell that it's sensemaking, not filing. Don't automate that away.)
Now — the platform name. You're naming the thing above the Neo- family (NeoBooks, NeoDecks, NeoStory all live under it). A few directions rather than one answer, since you'll want to sit with them:
Loam — fertile soil, straight out of your own "data is the new soil." Short, warm, unexpected, Muggle-easy; the Neo-forms grow from Loam. My current favorite for warmth.
Facets — each Nugget a facet that reflects the others; it's Indra's Net said quietly, no Buddhism required, and it's instantly legible.
Nuggetry — owns your vocabulary outright, a little playful, a little wink. Risks being too cute, but it's yours.
A "Neo-" umbrella (Neoworks, The Neo Stack) — maximal continuity with the forms you've already named, at the cost of some genericness.
The structural point regardless of which: platform name on top → the Neo- family as product lines under it → Nuggets as the unit → the Commons as the ground. You can keep "NeoBooks" et al. exactly as they are and just crown them.
Here's the conceptual map. I built it as a real .excalidraw.md, so it'll open editable in your Obsidian Excalidraw plugin — bubble-map-friendly, hand-drawn font, your kind of canvas to push around.
Validated data structure and prepared to draft operational documentation
Validated data structure and prepared to draft operational documentation
Now the one-page operating manual — written so Monday morning you can run it once, start to finish, on a single real Nugget.
Created a file, read a file
Both are yours. Quick orientation so you know what you're opening.
The conceptual map drops into Obsidian as an editable Excalidraw drawing (hand-drawn font, your palette — navy/teal/ember). The logic it lays out: the Commons on the left holds the Nugget as the atomic unit; Assemblies on the right (NeoBooks, NeoDecks, NeoStory, Blog, Shorts) are playlists drawn from that one pool — same jewels, many necklaces; the Relay runs along the bottom as the production flow; and Crystallization → the Big Fungus rises off the Commons as the social destination. The two-threads line and the "improve a Nugget once → every assembly improves" note are in there too. It's deliberately push-around-able — move the boxes, recolor, add the NeoStory branch as it firms up. Treat it as v1 scaffolding, not a finished poster.
The operating manual is the thing to actually test. The center of gravity is the Monday-morning walkthrough at the bottom: one real Nugget, ~30 minutes of your time, shipped. Everything above it is just the reference for who-does-what-where. Note the two places I made your instincts load-bearing: curation-into-the-Brain stays explicitly manual (it's sensemaking, hands-off would Snip it), and you and Jenn never share a tool — the Drive folder is the only contract between your worlds.
One thing I'd flag honestly: the manual describes a system that's mostly assemblable today, but two joints aren't wired yet — the automated three-candidate Dashboard picker, and the Whisper→Claude→you handoff as a smooth path rather than a manual copy-paste. Neither blocks you from running the relay by hand on Monday; they're the parts worth automating after you've felt the flow work manually a few times, so you automate the real shape instead of a guessed one. Run it rough first.
So you've now got the three wall-pieces: the Constellation (what to write), the Map (how it all fits), the Relay (how it ships). The honest next move isn't another artifact from me — it's you running stage 1 through 5 once, on the trousers Nugget, and seeing where it actually snags. That'll teach us more than another planning pass.
When you do, two things would be genuinely useful to bring back: which stage felt heaviest, and whether "Loam" (or whichever name) still feels right once you've said it out loud a few times. Want to leave it there so you can go do one — or is there a snag you can already see coming that we should design out first?
Platform conceptual map.excalidraw
Me:
ooooh, I really like the Kanban plugin route. and I understand and like the tiers approach. it means going back and adding lots of YAML and editing old YAML, but I think I can figure it out 🤣
a note: the Next Nugget Dashboard will not be populated mostly with Nuggets destined for NeoBooks; I think those will be outnumbered by current events, history, podcasts with guests, etc. I'm planning on being pretty loquatious online.
Loam feels like a great name for the shared data layer, skipping the Neo-nomenclature. though I also love mycelial language...
totally get how facets arrives from Indra's Net, but it's not rolling off my tongue happily. maybe it will grow on me.
agree that continuing with Neo will turn off more folks than it attracts. it's a little cold.
playing with "Nugg" is of course a treat. the NuggWeb? Nugglandia? Nuggetry?
but I'm realizing a greater aim of the greater project is to think together, to spark some collective intelligence. something like Brian Eno's idea of Scenius. something more social.
does that give you new inspirations?
the Conceptual Map you created is really interesting. I need to sleep on it.
Claude:
Synthesized naming concepts around collective intelligence and social layers
Quick practical relief on the Kanban-plus-tiers route before the fun part: you don't have to backfill all the old YAML at once. Default the whole backlog to one tier (say "someday") in a single batch find-replace, and only refine a Nugget's tier when it actually surfaces into the active lanes. The queue earns its metadata as it gets used, instead of you doing a 642K-sized homework assignment up front. 🤣
Now — your note about the Dashboard being mostly current events, history, podcasts-with-guests rather than NeoBook chapters is bigger than a population count. It relocates the center of gravity. The platform isn't a book-publishing system with some posts around the edges; it's a think-out-loud-in-public-with-others engine, and NeoBooks are the occasional sediment that crystallizes out of the stream. Which is the truer shape, and it's your own worldview: the living, social, conversational flow is primary — because that's where thinking-together happens — and books are the slow retrospective assembly, because books are where ideas go to die. You just quietly demoted NeoBooks from "the goal" to "a nice byproduct" and promoted the scene to the center. Good. That's more honest than the map I drew, which still has Assemblies looking like the destination.
Two things follow from it. First, your Nuggets now have a shelf life axis the dashboard should know about: a Snip!-on-this-week's-news Nugget is perishable and should jump the queue while it's hot; a DfT-principle Nugget is evergreen and waits patiently. The morning picker should weight perishable-and-timely items up — which also dovetails with your Shorts outperforming your longs, because timely-plus-short is exactly the attention vein. Second, and this is the one that excites me: podcasts-with-guests are crystallization in the wild. The guests are your externalizers, the conversation is idea sex, and Nuggetizing the episode captures the scene. So the social layer you said was under-built doesn't need a separate workstream — you build it by being loquacious with good people and publicly seconding what they say. Your plan to talk a lot online is the Scenius practice.
Which is the real answer to your naming question. Scenius is the right north star — Eno's word for the genius that belongs to a scene rather than a person, and it's deeply your lineage (OGM, the commons, "leave the territory better than you found it"). And it clarifies a layering you'd half-collapsed: Loam is the ground (settled — keep it for the data layer), but the thing you're now reaching for isn't the ground, it's the scene that grows on it — and that scene is your Big Fungus, finally wanting a warm, social, Muggle-facing name instead of the insider one. Scenius is its purpose; the public name is what you're hunting.
So, fresh candidates aimed at the social/think-together layer, not the substrate:
Loom — threads woven together; it pairs almost too well with Loam (ideas grow in the Loam, get woven on the Loom), and it ties straight back to Weaving the World and your whole thread metaphor. The honest catch: Loom is a well-known video-messaging app, and you publish video — that's a real headwind, not fatal but not nothing.
Understory — the living layer beneath the forest canopy where the connecting actually happens. It's literally your Weaving the World framing ("aboveground: a show / underground: collaborative inquiry") and it's mycelial without saying fungus. Catch: I believe there's a climate-data company by this name — different enough in domain, but worth a quick check.
Garden / a shared garden — straight out of Loam (soil → garden), it lands inside the "digital garden" movement that's already your exact crowd, and "tending a garden together" is the warmest Muggle frame for collective cultivation there is. Catch: generic alone; wants a modifier.
And the Nugg- play — NuggWeb, Nugglandia, Nuggetry — keep those as affectionate insider terms (they're warm and funny among people who already get it), but I'd steer the public-facing name of a think-together project toward warmth over cleverness, because the cleverness can read as a private joke to the newcomer you're trying to welcome into the scene.
Sleep on the map — and on the names; they're the kind of thing that either still rings true at 7am or quietly doesn't. When you wake, the most useful thing you could do to the conceptual map is move Assemblies out of the destination slot and put the scene/Scenius there instead, and see if the whole thing reads truer. That single move might tell you what the platform actually is.
Me:
a fiendish naming thought just arrived: call it Murderboard, or Group Murderboard. kidding! but it is funny, and it passes the icky but memorable test...
yes, exactly, on how NeoBooks fit into the larger production scheme. that's it. Books and Decks are the primitive, incidental (but relatively highly produced) byproducts (gosh, that sounds worse than I mean it). NeoStories, should we succeed in inventing them, will actually be more important, because they will be rethinking how we communicate, instead of emulating old media (books and decks).
on your description of podcasts with guests: bingo! love what you wrote.
so my Github repo, and I guess my Brain too, are in the Loam layer. other people will have other kinds of Loam, and as long as they make it accessible to outsiders (which Github wins me and an MCP server on TheBrain will win me), all is good, no? then we can collaborate...
and fold in intelligences like you. I'm loving CHI: Collective (Collaborative?) Hybrid Intelligence, and want to write about that and also experiment with it.
Scenius atop Loam = Big Fungus? hmmm...
oooh, the Loom and the Loam. interesting!
Danny Hillis's Knowledge Futures Group has been working on the Underlay, Interlay, and Overlay. "Understory" made me think of it.
Understory also reminds me of the Wood Wide Web, which highlights the power of mycelia and more. It works wonderfully with Loam, and offers up Canopy and other foresty things. but I'm lukewarm on it.
Maggie Appleton has written a lot about Digital Gardening, which I like. that seems like a spot for Crystallization.
Claude:
Acknowledging TheBrain MCP server's absence, pivoting toward architectural possibilities
Murderboard made me laugh — and there's a real wink buried in it, because a murderboard is literally a corkboard webbed with red string, which is your red thread and pullable threads made gloriously, conspiratorially literal. So it passes icky-but-memorable and secretly encodes the mechanic. Still: no. But keep it in your pocket as the name of the messy working version — "let's murderboard this" is a great verb for a live sensemaking session.
Two of your references I want to be accurate about rather than half-remember, so let me check them quickly.
Synthesized Hillis's Underlay with federated Loam vision and CHI framework
Your Understory→Underlay instinct just walked us into the most useful fact in this entire conversation, so let me stay right there.
The Underlay isn't merely adjacent to your federated-Loam idea — it is your federated-Loam idea, conceived by Danny Hillis years ago. It's a global, distributed graph of public knowledge whose hosts include universities and individuals, deliberately arranged so no single group controls the content. That's precisely your "my repo and my Brain are Loam, others have other Loam, and as long as it's accessible we can collaborate." Same architecture. And his stated why is nearly verbatim your Rel8 opening: that our paradigm for sharing knowledge is broken — false beliefs, distrust in science and journalism, misinformation. He even calls the project his penance for having sold the earlier version (Freebase) to Google, which became Google's Knowledge Graph. Mit + 2
This collapses three of your open questions into one person. On funding: I said Hillis was almost too on-the-nose — it's stronger than that. He didn't bounce off this problem, he's been building your layer. You wouldn't pitch him a concept; you'd show him a running instance of the thing he conceived. On naming: Underlay / Interlay / Overlay is his vocabulary, so your drift to Loam/Loom is a clean escape — soil and weaving, not strata, no collision. Stay out of the -lay zone. And KFG's other project, PubPub, is a publishing platform that builds conversation, annotation, and versioning into the document itself — NeoBooks-adjacent prior art worth knowing as both a platform option and a "here's what's been tried" reference. MIT Media Lab
Now the honest part, which is also the opening: the Underlay's public work tapered after around 2019–2020, and I think the reason is the very thing you just named — CHI. The Underlay's stalling problem was populating and extracting structured knowledge from messy documents, which was brutal in 2018 and is suddenly easy in 2026 because LLMs do exactly that. So your pitch to Hillis isn't "fund my version of your idea" — it's "your premise was right and early, the missing ingredient was collaborative AI, and here's a small running cadence showing the substrate finally working because humans and machines populate it together." That's a revival framing: it credits him, supplies the missing piece, and arrives with traction. CHI is the reason the federated knowledge commons is tractable now when it wasn't — and "CHI: the missing ingredient" is itself a strong Nugget, especially welded to the Underlay story. KFG NotesUnderlay
Which fully validates your Loam-as-a-role insight: you don't need everyone on one platform (the cold-start that kills these things), you need each substrate porous — readable and linkable from outside. GitHub does that for your repo; a TheBrain MCP server would do it for your Brain. That's Design from Trust at the infrastructure layer: the default is accessible, and collaboration becomes possible the instant two porous Loams can see each other. The Big Fungus doesn't require a destination; it requires porousness plus a norm of pointing.
On that MCP server — I checked, and there's no turnkey TheBrain MCP server today; it'd be a build. But it's exactly what MCP is for (expose Thoughts, Notes, and links as queryable resources), Claude is good at standing these up, and the payoff is large: it turns your Brain from a thing only you can navigate into Loam that other people and AIs can traverse — the accessibility you want and the on-ramp for CHI to enter your graph. Right thing to raise with Harlan; buildable without him via the export-to-Markdown path as a fallback.
So "Scenius atop Loam = Big Fungus?" — yes, and the stack resolves cleanly if you want it:
Loam is the substrate (porous externalized thinking, yours and others'). Loom is the weaving — threads connecting across Loams, the red thread and the pulled thread, assembly and narrative; and Loam + Loom is a genuinely lovely pair (ideas grow in the Loam, get woven on the Loom — mind the Loom video-app collision since you publish video). Scenius is what emerges when it works — Eno's intelligence-of-the-scene, the why. CHI is who does the weaving: humans + AIs. Big Fungus is the affectionate name for the whole living thing. So Scenius is what the Fungus produces, Loam is what it grows in, CHI is who tends it. On your C-word: I'd run both at different altitudes — Collective is the property that emerges (it sits with Scenius), Collaborative is the practice that produces it. "Collective Hybrid Intelligence" as the noun; "collaborative" as how you describe the doing.
And digital gardening is exactly the home for Crystallization you sensed — the existing movement and vocabulary for tending ideas in public so they grow over time, which is crystallization at the individual scale; crystallization is just its collective form, others adopting and seconding. Maggie Appleton is the canonical voice there and a near-perfect first externalizer to court — public, visual, tools-for-thought, already wrestling with AI in that space. Put her near the top of Finding the Externalizers.
Last, on Books/Decks sounding like lesser byproducts — they're not lesser, they're the familiar-format ambassadors. They let a Muggle enter through an object they already trust, while NeoStory invents new communication instead of emulating old media. The hierarchy you feel is real: Decks and Books are backward-compatible on-ramps that ship now; NeoStory is the frontier and the real prize. Just hold the same discipline as the Big Fungus — the most important thing is the least tractable, so let NeoStory be R&D you prototype, and don't let its importance pull effort off the ambassadors that actually ship this year. (And yes — podcasts-with-guests are crystallization in the wild: each guest an externalizer, the episode the capture, your public seconding the gesture. Your plan to be loquacious with good people is the Scenius practice, already underway.)
Sleep on all of it. The genuinely new lead isn't on the conceptual map — it's that conversation with Hillis, which you should have after the cadence is running so you arrive with a moving thing instead of a deck. When you surface, tell me whether Loam → Loom → Scenius → CHI → Big Fungus still holds at 7am, or which layer quietly collapsed overnight.
Me:
it occurs to me, after pointing to Hillis's KFG, that it's likely that several groups out there have developed complementary architectures. a little ways down the road, we should run a search.
thanks for taking Murderboard in the right spirit! between you and me: a post I want to record soon says that some days I feel like the love child of Cliff Clavin and Carrie Mathison 🤣
oh wow. I hadn't thought through how sympatico KFG is/was. first, do they look still active? second, have they delivered any architectural components that we should adopt?
yesterday, in a different chat with you, we talked about Brain+MCP and I reported that I'd had a great conversation with Harlan that MCP available to outsiders is on their map, but not that soon. I do think the conversation raised its importance in his mind, though.
I also offered him to record a podcast episode I would publish through Substack, which I'm looking forward to a lot.
btw, one (of many) reason I love the mycelial metaphor is that mushrooms are the fruiting bodies of the much larger/complex mycelial networks; NeoBooks and NeoDecks (and other similar) would be the fruit of this project :)
I'm feeling Scenius is a thing to point to, not a name to borrow. a small group call I had a decade ago was already using it.
really like "familiar-format ambassadors." that's right on the money.
ironically, a short thread of posts I want to publish are about whether you and your peers obviate the need for things like books, decks, and maps. I think we've talked about this a little before, and landed that the mappy representations actually become more important than before, as boundary documents among humans with one another, and with and through LLMs.
btw, a NeoDeck would be very straightforward to prototype. multiple Markdown presentation apps exist (see https://gist.github.com/johnloy/27dd124ad40e210e91c70dd1c24ac8c8 ), but all the ones I've looked at get it wrong: they have one Markdown file with separators to divide the pages. they don't pursue a Nugget approach at all. once you start to think Nuggetly, implementation of a very basic player seems very simple, no?
thoughts?
Claude:
Grappling with Knowledge Futures' financial collapse and implications
Cliff Clavin crossed with Carrie Mathison is perfect — and it's a real post, not just a quip. It's the murderboard self-portrait: the guy who knows ten thousand things and the woman who sees the thread connecting them, living in one head, never sure which one is driving. That's the charming-not-bragging register we wanted for showing off the Brain — it lets people laugh with the obsession instead of being intimidated by 642K nodes. Record it.
Let me check the two things you asked me to verify, then react to the rest.
Synthesized KFG funding crisis, mycelial metaphor refinement, boundary-object theory, and NeoDeck assembler architecture
Synthesized KFG funding crisis, mycelial metaphor refinement, boundary-object theory, and NeoDeck assembler architecture
Both checks back, and the KFG one is more sobering than I expected — in a useful way.
Knowledge Futures is alive but wounded. PubPub is real and now fully open-source, but in mid-2025 they hit a wall: a funder pulled the second half of promised money after months of negotiation, so they're shutting down the centrally-hosted PubPub Platform at the end of 2025 and moving users to self-hosted instances, with the older PubPub Legacy kept on life support only through the end of 2026. The org persists, Commonplace (their publication) is relaunching, and the Underlay is still listed but remains the least-delivered piece. KnowledgefuturesKnowledgefutures
So, your two questions answered honestly. Still active? Yes, but in survival mode, not ascendancy. Components to adopt? The real deliverable is PubPub Platform itself — open-source, API-first, self-hostable, and explicitly built for annotation, versioning, community knowledge graphs, overlay journals, and evolving textbooks. That's genuine NeoBooks-adjacent prior art, and the patterns are worth studying — an "overlay journal" is basically a playlist, an annotation layer is basically crystallization. But it's a heavyweight CMS; for a solo cadence it's almost certainly overkill against your Pandoc-plus-Obsidian path. Study it for ideas, don't adopt it. Knowledgefutures
And here's the part that earns its place: their crisis is a live demonstration of exactly why I steered you off grant dependence. The MIT-born, Hillis-pedigreed, open-infrastructure org with everything going right still got rugpulled by a single funder mid-flight. Let that harden the plan — bank the cheap, self-funded cadence, and treat Hillis as a patron and elder, not a grant pipeline. It also tempers the timing: his org is stretched thin right now, so that conversation is a relationship, not a transaction, and it's even more important you arrive with a running thing rather than an ask.
The mycelial completion is lovely, and it actually fixes the conceptual map. If mushrooms are the fruiting bodies of the network, then the Big Fungus isn't a destination sitting up in the corner — it's the whole mycelial mass (Loam plus the living connections, mostly invisible), and NeoBooks, NeoDecks, and Shorts are its fruit: the visible, pickable harvest. That reconciles everything — the fruit is the "familiar-format ambassador" people first taste, the network is the organism, and the leafcutter ants are your community farming it. So the map edit is: Assemblies aren't a rival to the Big Fungus, they're what it fruits. Cleaner, and truer.
Scenius as a thing-to-point-to, not a name to borrow — agreed, fully. It joins your citation shelf (Eno, Ostrom, Meadows, Forster): the phenomenon you aim at, lowercase, while the whole keeps its own mycelial name. And noted on the Brain+MCP correction — that's better intel than I had: roadmapped-but-not-soon, with your nudge having raised its priority. The Harlan podcast is a quiet triple win: it deepens the relationship with the one person who controls whether your Brain becomes porous Loam, it's content, and it's crystallization-in-the-wild with a perfect guest. Lock it in.
On the "do LLMs obviate books, decks, maps?" thread — yes, and I think it's one of your sharpest posts, because the answer inverts the fear. The term that makes it land for smart readers is boundary object (Star and Griesemer): a shared artifact that lets different parties coordinate without fully agreeing. When LLMs make prose infinite and nearly free, the scarce, valuable thing becomes the compressed, negotiated structure — the map — because that's the interface where human judgment and machine generation meet and point at the same thing. LLMs don't obsoletes maps; they raise the price of maps, because the map is how you stay the author of your own thinking instead of drowning in plausible generated text. There's even a Snip! in it: ungrounded LLM use snips you from your own sensemaking, and the map is what reconnects you. And the post is self-proving — this entire week has been us co-editing maps as boundary objects between a human and an LLM. You don't have to argue the thesis; you can just show our transcript.
Now the NeoDeck — you're exactly right, and the reason it's simple is worth naming precisely: thinking Nuggetly doesn't mean building a slide engine, it means building a tiny indirection layer in front of one that already exists. Every tool in that gist — Marp, reveal.js, remark, Slidev, Quarto, Pandoc — takes one file split by ---. So the NeoDeck "player" is just an assembler: read the playlist → pull each slide-Nugget file → join them with --- → hand the result to any of those mature renderers. Nugget-level reuse, addressing, and version history on the authoring side; free themes, transitions, and PDF/PPTX export on the output side. You never touch the hard part. And it's the same pattern as NeoBook→Pandoc→EPUB: resolve playlist → assemble Nuggets → feed a format-specific renderer. One concept, every output (NeoStory eventually is just a different renderer). That's the entire "player."
There's an even-lower-code route you could try without me: you already use transclusion. Obsidian's slides plugins render a note as reveal.js using ---, so a NeoDeck = a deck note that transcludes (![[slide-nugget]]) each slide with --- between them. Reorder by moving lines (your queue gripe, solved a third time); reuse a slide-Nugget across many decks. One thing to test: the core Slides plugin may not resolve transclusions in slide mode — the community "Advanced Slides / Slides Extended" plugin is the one more likely to handle embeds, so start there.
Want me to build the little assembler (playlist → Marp/reveal, with PDF/PPTX out) as your first vibecode artifact and proof-of-concept? Or try the Obsidian-transclusion path yourself first and bring back where it snags. Either way — and to honor your own "down the road": I'll hold the full survey of complementary architectures for when you're ready, but two are clearly real and worth that future look — KFG/Underlay, and Flancian's Anagora, which already federates people's digital gardens into a shared "agora" and is almost eerily on-point for the porous-Loam idea.
Me:
gosh, this is the real Big Fungus: Armillaria ostoyae (2400 years old? the Humongous Fungus) https://en.wikipedia.org/wiki/Armillaria_ostoyae maybe call the project Armillaria? too geeky.
please remind me of this later: when we were oral cultures, a thing we deprecate now, the stories and songs we told and memorized were the fabric of our societies, our long-term memory of hard-won knowledge. the warp and weft of our social fabric is what we've Snipped all along, and have a chance now to re-weave. which is all about reconnecting ourselves.
ooooh, I hadn't thought of just doing assembly of the playlist dynamically. I'm not sure I love it, but it's a hack that would be easy to test. are any of the cited apps elegant enough to pass a sniff test (if you can tell from reviews, etc)?
partly I wasn't thinking of assembly because I imagined a traversal viewer could maybe follow alternate paths through a presentation elegantly, instead of being forced down the one brick road. but I hadn't thought that through much.
while we're on NeoDecks, another of my vibecoding wish-list projects is a Slide Viewer for the emergent deck: instead of just editing one markdown page with transclusion marks, why not be able to move things around, like the Kanban plugins? but I guess I underestimate the complexity of doing that... or do I?
Flancian is a dear friend, and you'll find notes pointing to me and my Brain through his Agora. He would be a happy collaborator, though I don't quite understand the structure he has built and how it fits, other than it being openly available.
let's schedule the NeoDecks test run for three weeks out, so it doesn't interrupt my getting comfortable with a higher cadence of post production. the more topics we touch, the more I realize I have to Externalize! yeeps!
Claude:
Verified fungal facts, updated memory with oral culture insight and NeoDeck schedule
He found it, and the facts are even better than the joke deserves: the largest one is in the Malheur National Forest in eastern Oregon — it covers about 3.5 square miles, may weigh up to 35,000 tons, is estimated at roughly 8,000 years old, and is considered the world's largest single living organism, almost entirely underground. It's in your state. And it does exactly what your metaphor needs: it spreads invisibly through the soil via "shoestring" rhizomorphs and only surfaces in autumn as honey mushrooms — its fruiting bodies. The vast connected organism is hidden; the mushrooms are just what you can see and pick. That's your Loam-and-fruit picture, alive and 8,000 years old, a day's drive from Portland. (It even glows — the mycelium is bioluminescent, "foxfire.") WikipediaWikipedia
So: don't name the project Armillaria — too geeky, you're right, it'd filter out the Muggles. But point to it, the way you point to Eno and Ostrom. The Humongous Fungus is the perfect opening anecdote for the Big Fungus: the largest, oldest living things on Earth are underground connection networks, mostly invisible, and one of them is in your backyard. Geeky precise thing = the citation; warm thing = the handle. Same move as always.
Let me capture the oral-cultures thread so it actually comes back to you.
Identified two edits to incorporate into existing work
Identified two edits to incorporate into existing work
Stored both — the oral-cultures thread will come back to you, and it deserves to, because it's the deepest framing you've landed all week. It reaches under Snip!: the warp and weft of a society's long-term memory was held in memorized story and song, we deprecated orality as primitive, and in doing so we Snipped the social fabric itself — and now, with porous Loam and CHI, we have a shot at re-weaving it. That's not a separate idea; it's the origin story of the whole project. It also quietly resolves your name hunt: Loom stops being a clever pun and becomes literal — re-weaving the fabric we unraveled. Hold it. (And note the irony you'll enjoy: you're using the most advanced text-prediction machine ever built to argue for recovering what oral culture knew. The tool and the thesis are in tension, which is exactly what makes the post land.)
Now your NeoDeck questions, straight.
Which apps pass a sniff test? From the community signal in that gist and beyond: Slidev (sli.dev) is the one people keep praising as modern and elegant — as recently as mid-2025 a commenter calls it great and very well maintained — with the honest caveat that it's heavy (one user clocked its install at 400MB). Marp is the leaner, faster, cleaner counterpart — same markdown approach, quick PDF export, a fraction of the bloat, and it's the better fit for the assembler pattern precisely because it's CLI-clean. Quarto is the robust scholarly option (Posit-backed) if you ever want academic polish. And iA Presenter is the prettiest, most opinionated, content-first one — but it's Mac-only, commercial, and not CLI-scriptable, which makes it lovely to use and useless to hack. My read: for the proof-of-concept, Marp; for visual flair if you tolerate weight, Slidev.
Is dynamic assembly "just a hack"? No — it's the correct separation of concerns (authoring vs. rendering). But your instinct that it isn't quite what you want is also right, and here's why it matters: assembly-over-an-existing-renderer can only ever give you a linear deck, because every one of those tools is ----separated and one-way. The moment you said "traversal viewer that follows alternate paths instead of one brick road," you stepped out of NeoDeck entirely. That's NeoStory. A NeoDeck replaces PowerPoint, and PowerPoint is linear — being linear is fine, it's the familiar-format ambassador. The path-choosing, map-navigable, branch-as-you-go thing is the new communication form you said was the real prize. So your branching idea isn't a fancier NeoDeck; it's the seed of the frontier — which means it gets prototyped later as R&D, not built now, exactly the discipline we set. Naming it correctly keeps it from bloating the simple thing that ships.
The drag-to-arrange editor — do you underestimate the complexity? Both, actually. A bespoke, general slide-arranging editor is real work — more than you think, because it's a stateful drag-drop app that has to write order back to disk reliably. But you almost certainly don't need to build one, because the thing you already liked solves it: the Kanban-over-Bases pattern. Cards = slide-Nuggets, the lane/order = the deck sequence, drag to rearrange, and the order exports to a playlist. Same mechanism you wanted for the Next Nugget Dashboard, pointed at deck-building instead. So you're over-estimating in thinking you must build it, and under-estimating only in that a from-scratch one would be a slog. Reuse the Kanban; skip the build.
On Flancian — this is better than "a happy collaborator." As I understand the Agora (and he should correct me), it's a federation layer: everyone keeps their own digital garden, and when you visit a concept-node, the Agora shows you everyone's notes on it stitched together — so it's a working Loom, the weaving layer over many people's Loam, with your Brain already among the threads. That means the "where does crystallization happen / how do I activate collaboration" question isn't hypothetical for you — a running instance already points at your work, built by a friend. When you're ready, that's not a cold pitch; it's a warm "Flancian, show me how the Agora weaves, because I think it's the Loom and my Brain's already in it." Put him at the very top of Finding the Externalizers.
NeoDecks parked for ~early July — noted and stored, and it's the right call: don't let a fun prototype interrupt the one habit that actually changes your life, which is shipping Nuggets on a rhythm. On the "yeeps, I have to Externalize so much" — that feeling is the backlog talking, and the backlog is fuel, not debt. You don't externalize it all; you externalize one tomorrow morning, then one the next. We've spent this week building the machine precisely so the mountain becomes a daily handful. The pressure you feel is just the gap between everything you see and the one stone in front of you — and the relay exists to close that gap one Nugget at a time. Go get comfortable with the cadence. The Humongous Fungus took 8,000 years; you've got a head start.
Me:
whaaa? it's bioluminescent???
one of the huge factors in the social fabric Snip is that as we wrote down our stories and songs, emerging Capitalism insisted on overprotecting them, instead of treating them as part of our Commons or patrimony. the IP overprotection was a huge snip, and a big piece of my logic on why books are dysfunctional
I'm afraid Loom is a well-used product name, but I love using it.
I like the sound of Marp. had it in my Brain already. let's try that, when we go there.
aaaah, I like your saying that the linear decks do emulate PowerPoint. I'm not sure the non-linear decks are NeoStory, insofar as I think Prezi (and Prezi emulated in Excalidraw, btw) are more where I'd head.
in re-using the Kanban plugin, might we fork it and modify it to be more Slide Table-y? predicting you will say it's a lift, we should KISS for now
love where this chat has gotten us. thank you.
Claude:
Synthesized IP overprotection as missing keystone, refined presentation taxonomy, affirmed KISS principle gracefully
Bioluminescent, yes — foxfire. The largest, oldest living thing on Earth is a hidden network that occasionally glows in the dark forest floor. If you ever wanted an image for "the invisible Commons that connects us, quietly lighting up," it's been growing in eastern Oregon for 8,000 years. Good luck not writing that one.
The IP-overprotection point is the keystone the orality thread was missing — and it's already in your Brain, since "We Overprotect the Ideas in Books" was right there in the screenshot you sent. Here's how it locks in: oral culture held knowledge in the Commons — you couldn't own a song everyone sang. Writing made ideas capturable, and emerging capitalism captured them; copyright turned shared patrimony into private property. So the IP Snip is the specific mechanism by which the social-fabric Snip happened — the cut that severed stories from the Commons. That's why it belongs squarely in the Snip family and why it's your sharpest argument that books are dysfunctional: a book is IP-overprotection made physical, ideas locked between covers you must buy, can't remix, can't reweave. And you're already living the antidote — CC0 on Weaving the World, publishing openly, the Commons stance. That's Design from Trust applied to IP, and it's not a side preference, it's the precondition for the whole stack: you cannot crystallize or reweave what's locked up. No open Loam, no Big Fungus. Overprotection is the exact anti-pattern everything you're building exists to undo.
On Loom — you don't have to give it up, you just don't put it on the door. Use it the way you love: the verb and the image. We reweave; the fabric runs on a loom; what got unwoven gets woven back. That's yours, trademark-free, because you're invoking the metaphor, not naming a company — same as Armillaria and Scenius. Point to it, don't brand with it.
Marp it is, in July. Lean, CLI-clean, already in your Brain — right pick for the assembler.
And you're right to push back; let me fix the cut. The line isn't linear-vs-non-linear. Prezi-style spatial navigation is still a presentation — an author walking an audience along a path — so it's a NeoDeck, just spatial instead of stacked. And Prezi-in-Excalidraw is perfectly you, since you already draw bubble maps and talk through them; that's a spatial deck waiting to happen. The real dividing line is who's driving. Author walks the path (linear or Prezi-spatial) = NeoDeck, your red thread performed. Reader drives, pulls their own thread, explores a space you didn't pre-walk = NeoStory. So it's your own two-threads distinction doing the sorting, and Prezi-Excalidraw lands cleanly on the NeoDeck side.
And yes — KISS, exactly as you called it before I could. Don't fork the Kanban into a slide-table yet. Run it as-is and let real friction, not imagined friction, earn any custom build. The slide-sorter grid is a lovely someday — it's even in your Rel8 transcript, "the big slide table is just a view on these pages" — but someday, after the cadence is humming and you've felt where the stock tool actually pinches.
Where this got us, in one breath: you came in with a 30-year backlog and a fear that books are where ideas die, and we found the shape that honors both — Nuggets in porous Loam, woven into familiar-format ambassadors now and new forms later, shipped on a daily cadence through a relay built around how you actually work, all pointed at reweaving what got Snipped. Three artifacts on the wall, a name still settling, a glowing fungus in Oregon to point at, and the orality thread waiting for you when you want it.
The one thing no map can do for you is the next move: pick tomorrow's Nugget and say it out loud. The trousers one is right there.