Build the list nobody has to sift
"the best websites, gathered and grouped for one reader"
Design systems for small teams12 sites
Start here3
Go deeper5
When you need more4
Grouped for one reader
Somebody joins your team and asks where to start. Somebody else asks which of the four links you sent is actually worth their afternoon. Somebody external asks why these particular sites, and you find yourself writing a paragraph in a chat window that nobody will read twice. A curated Glubler answers all three before the question is finished, because the grouping is the explanation — three sites to start with, five to go deeper, four when the first four are not enough. You did the sifting once. They never have to do it again, and they can see the reasoning without you in the room.
Gather from everywhere and keep only what earns its place
Good curation is a sweep. No single site has the definitive twelve links on a subject, and nobody can read everywhere at once, so the list that arrives from one place is always partial and always shaped by whoever wrote it. Reading a bit of everything and keeping the four or five that actually hold up gives you something no source could hand you alone — a view that spans the argument instead of repeating one side of it. The links you did not keep cost nothing once they are gone, which is the part a shared tab stops you from doing.
news.ycombinator.com4 kept
reddit.com/r/…3 kept
css-tricks.com5 kept
newsletter.dev2 kept
One list14 sites
Many sources, one list
Grouped so the list explains itself without you in the room instead of forty bare links
A flat list of forty links transfers nothing. It transfers forty links, and every reader has to guess which six are the ones you would have picked — which means they either all read them or none do. Split the same set into Subglublers and the choice is already made: what to open first, what to open when the first thing does not answer your question, what to reach for when the work gets strange. Newcomers take the first group and feel oriented in a minute. The colleague who already knows the field goes straight to the third and skips two thirds of it. One list, genuinely useful to both, because it was never one undifferentiated pile.
Design systems for small teamspublic
New joinerWhere to start3
ColleagueThe one good link1
ClientWhy these12
One list, three readers
Hand it over and let them add to it
Curating for other people only works if the list outlives the conversation it was made in. A Glubler is public unless you close it, so the link goes into the onboarding doc, the channel, the handover note, and stays current because whoever finds something better can add it without asking you. Nobody forwards a folder of screenshots and nobody maintains one, but a link gets clicked for years and the person who clicks it can put their own finds next to yours. That is the difference between a list you made and a list that keeps being made.
Start a curated GlublerWebsites, not documents
Plenty of curation tools gather files. A PDF gets skimmed once, bookmarked in nobody’s system, and forgotten; the same site visited twice becomes a habit. That difference matters more the further the list travels, because a reader handed a document has to trust your summary, while a reader handed a link can go and judge it for themselves. Sites keep their own context, update themselves without you, and stay worth sending to somebody who has never spoken to you.
news.ycombinator.com
5 keptWho it is for is the whole job
The same twelve sites are essential to a designer who owns a design system and close to useless to an engineer who just needs the tokens. Curation starts by deciding which of those two people you are building for, and almost everything after that follows: which sites, how many, grouped how, how much assumed knowledge to explain. Name the reader and the list writes itself. Leave them vague and you end up with the average of everybody’s taste, which is the one thing nobody in the room actually asked for.
ColleagueGo straight past
Sweep wider than one source ever could
Forums where the answer is actually argued out, small personal sites that nobody wrote a listicle about, the references quietly linked from four other good pages. One source is always somebody else’s taste, and it is always partial — the twelve links they chose are the twelve links they happened to read. Gathering across a spread and keeping what holds up gives you a list that covers the disagreement rather than one side of it, and it is the only version worth handing to somebody whose judgement you respect.
Subglublers carry the reasoning
Splitting one long list into three named groups — start here, go deeper, the ones for when the work gets odd — turns forty links into a decision. A newcomer takes the first group and is oriented in a minute; someone already in the field skips to the third and reads a third of what they otherwise would. Same sites, same effort, and the grouping is what makes it usable by people at different levels rather than pitched at the one reader you had in mind while you built it.
A list that keeps getting better
Hand over a folder of screenshots and it dies the week you stop maintaining it. Hand over a Glubler and it does the opposite: it is public unless you close it, so it sits in the onboarding doc permanently, and everyone who opens it can add the site they found last week without permission or a message. Half the entries in a mature list were put there by somebody who was not its author. That is not a leak in your curation, it is the list working — the reasoning is in the structure, so other people can extend it without flattening it.
Glubler FOR
CURATION
The list outlives you
A curated Glubler of websites, grouped into Subglublers so the reasoning is visible, shared as a link rather than a folder of screenshots — and open for anyone who opens it to add the good ones.