Hacker News new | past | comments | ask | show | jobs | submit login
Most-favorited Hacker News posts (observablehq.com)
1266 points by tlarkworthy on Sept 2, 2020 | hide | past | favorite | 259 comments

The OP says "The most favorited articles by the top 10k most active Hacker News members." How was "most active" defined? (Edit: oh I see - the users who post the most comments: https://news.ycombinator.com/item?id=24352688). It's an interesting list, and it never occurred to me that by counting the top favorites of different subsets of users you might get quite different interesting lists.

I got curious about what the global most-favorited would be. Here are the top 50. The first column is the fave count. It's interesting how many are Ask HNs, i.e. text posts with no external URL. Sorry that the item ids aren't clickable:

  836 19087418 Ask HN: What books changed the way you think about almost everything?
  783 16745042 Ask HN: What are the best MOOCs you've taken?
  685 16775744 Ask HN: How to self-learn electronics?
  581 21332072 Ask HN: Successful one-person online businesses?
  554 21581361 Ask HN: What's the most valuable thing you can learn in an hour?
  510 18588727 Ask HN: What are your “brain hacks” that help you manage everyday situations?
  510 20264911 Ask HN: What do you do with your Raspberry Pi?
  506 22786853 Ask HN: What are your favorite low-coding apps / tools as a developer?
  472 15919115 Machine Learning 101 slidedeck: 2 years of headbanging, so you don't have to
  470 23151144 Ask HN: Mind bending books to read and never be the same as before?
  463 20916749 Questions to ask a company during a job interview
  461 22299180 Ask HN: What are some books where the reader learns by building projects?
  454 23092657 Ask HN: Name one idea that changed your life
  448 23904000 Systems Design for Advanced Beginners
  447 22400375 Mathematics for the Adventurous Self-Learner
  444 23588896 Teach Yourself Computer Science
  441 21585235 Basic Social Skills Guide (2012)
  439 17238135 How to be a Manager – A step-by-step guide to leading a team
  439 22105229 Tricks to start working despite not feeling like it
  432 16493489 Machine Learning Crash Course
  425 24351073 Most favorited Hacker News posts of all time
  422 22310813 Gears
  421 20985875 The boring technology behind a one-person Internet company (2018)
  416 19490573 A guide to difficult conversations
  409 24120275 How to stop procrastinating by using the Fogg Behavior Model
  409 21324768 Ask HN: What's a promising area to work on?
  406 23229241 Linux Productivity Tools (2019) [pdf]
  398 21712194 Ask HN: Best book / resources on leadership, especially for tech teams?
  396 12702651 Ask HN: What is your favorite YouTube channel for developers?
  381 18805624 Algorithms, by Jeff Erickson
  374 21536789 Build Your Own React
  372 18104814 Ask HN: What are the best textbooks in your field of expertise?
  369 23170881 Ask HN: What's your quarantine side project?
  366 22226380 The missing semester of CS education
  365 23053981 Medium-hard SQL interview questions
  364 17163251 The Importance of Deep Work and the 30-Hour Method for Learning a New Skill
  363 22276184 My productivity app for the past 12 years has been a single .txt file
  360 22235279 Ask HN: What Skills to Acquire in 2020?
  360 13660086 Ask HN: What are some books where the reader learns by building one project?
  358 14486657 Ask HN: What language-agnostic programming books should I read?
  358 15602538 Ask HN: Where can I find high-end stock images for a website?
  356 19900955 Ask HN: What overlooked class of tools should a self-taught programmer look into
  356 20044876 Advanced Data Structures (2017)
  355 19264048 Immersive Linear Algebra (2016)
  355 12713056 Ask HN: How to get started with machine learning?
  353 23339830 Tools for Better Thinking
  353 21900498 Ask HN: Best books you read in the past decade?
  352 20254057 Startup idea checklist
  347 17999659 Ask HN: Favorite teachers on YouTube?
  347 23276456 Ask HN: What startup/technology is on your 'to watch' list?

I did normalize by number of favorites too, so each person had 1 vote shared over all their favorites. The idea that a person who only had one favourite was a really strong signal for that single link, whereas people who have lots of favorites are using it differently to what I want to select for.

At some point I had a similar selection of Ask HN dominance. One problem with my data is I only grabbed the first page of favorite results.

EDIT: SHOOT, you exposed a bug, my parser skips things things without URLs! Hence I ended up filtering out most of the Ask HN's!

EDIT: Fixed it

Now my list looks a bit more similar to yours but not quite. Currently "Ask HN: What books changed the way you think about almost everything?", your #1 is my #50. If I turn off normalization, it goes to #13 on my lists, so that has quite a significant impact on the rankings.

sidenote: Lol this post is the 20th most favorited article now.

> sidenote: Lol this post is the 20th most favorited article now.

Would that make this post a meta-favorite?

I'd actually assume one favourite to be a misclick.

But those would be randomly distributed so it would not affect the top ranks much.

> The idea that a person who only had one favourite was a really strong signal for that single link

Perhaps karma or account age should join the equation.

Yeah, more recent articles should be penalised to normalize for the increased quantity of accounts.

I kind of liked the list without Tell/Ask HNs as well (it has the bash shell guide on it) so I made it a toggle.

I guarantee that a lot of favorites are just ‘to read later’. In that model, articles themselves are read at once or go into Pocket-like services, with selective perusal of comment sections. In most cases, comments are sorta predictable, or interesting perspectives are quickly exhausted—and only especially rich threads go into the favorites to later be pored over. However, interesting ‘Ask HNs’ are obviously followed by lots of interesting comments (hopefully), so they fall straight into the category to be sorted out with attention.

This needs links :( a gist perhaps?

1. Ask HN: What books changed the way you think about almost everything? - https://news.ycombinator.com/item?id=19087418

2. Ask HN: What are the best MOOCs you've taken? - https://news.ycombinator.com/item?id=16745042

3. Ask HN: How to self-learn electronics? - https://news.ycombinator.com/item?id=16775744

4. Ask HN: Successful one-person online businesses? - https://news.ycombinator.com/item?id=21332072

5. Ask HN: What's the most valuable thing you can learn in an hour? - https://news.ycombinator.com/item?id=21581361

6. Ask HN: What are your “brain hacks” that help you manage everyday situations? - https://news.ycombinator.com/item?id=18588727

7. Ask HN: What do you do with your Raspberry Pi? - https://news.ycombinator.com/item?id=20264911

8. Ask HN: What are your favorite low-coding apps / tools as a developer? - https://news.ycombinator.com/item?id=22786853

9. Machine Learning 101 slidedeck: 2 years of headbanging, so you don't have to - https://news.ycombinator.com/item?id=15919115

10. Ask HN: Mind bending books to read and never be the same as before? - https://news.ycombinator.com/item?id=23151144

11. Questions to ask a company during a job interview - https://news.ycombinator.com/item?id=20916749

12. Ask HN: What are some books where the reader learns by building projects? - https://news.ycombinator.com/item?id=22299180

13. Ask HN: Name one idea that changed your life - https://news.ycombinator.com/item?id=23092657

14. Systems Design for Advanced Beginners - https://news.ycombinator.com/item?id=23904000

15. Mathematics for the Adventurous Self-Learner - https://news.ycombinator.com/item?id=22400375

16. Teach Yourself Computer Science - https://news.ycombinator.com/item?id=23588896

17. Basic Social Skills Guide (2012) - https://news.ycombinator.com/item?id=21585235

18. How to be a Manager – A step-by-step guide to leading a team - https://news.ycombinator.com/item?id=17238135

19. Tricks to start working despite not feeling like it - https://news.ycombinator.com/item?id=22105229

20. Machine Learning Crash Course - https://news.ycombinator.com/item?id=16493489

21. Most favorited Hacker News posts of all time - https://news.ycombinator.com/item?id=24351073

22. Gears - https://news.ycombinator.com/item?id=22310813

23. The boring technology behind a one-person Internet company (2018) - https://news.ycombinator.com/item?id=20985875

24. A guide to difficult conversations - https://news.ycombinator.com/item?id=19490573

25. How to stop procrastinating by using the Fogg Behavior Model - https://news.ycombinator.com/item?id=24120275

26. Ask HN: What's a promising area to work on? - https://news.ycombinator.com/item?id=21324768

27. Linux Productivity Tools (2019) [pdf] - https://news.ycombinator.com/item?id=23229241

28. Ask HN: Best book / resources on leadership, especially for tech teams? - https://news.ycombinator.com/item?id=21712194

29. Ask HN: What is your favorite YouTube channel for developers? - https://news.ycombinator.com/item?id=12702651

30. Algorithms, by Jeff Erickson - https://news.ycombinator.com/item?id=18805624

31. Build Your Own React - https://news.ycombinator.com/item?id=21536789

32. Ask HN: What are the best textbooks in your field of expertise? - https://news.ycombinator.com/item?id=18104814

33. Ask HN: What's your quarantine side project? - https://news.ycombinator.com/item?id=23170881

34. The missing semester of CS education - https://news.ycombinator.com/item?id=22226380

35. Medium-hard SQL interview questions - https://news.ycombinator.com/item?id=23053981

36. The Importance of Deep Work and the 30-Hour Method for Learning a New Skill - https://news.ycombinator.com/item?id=17163251

37. My productivity app for the past 12 years has been a single .txt file - https://news.ycombinator.com/item?id=22276184

38. Ask HN: What Skills to Acquire in 2020? - https://news.ycombinator.com/item?id=22235279

39. Ask HN: What are some books where the reader learns by building one project? - https://news.ycombinator.com/item?id=13660086

40. Ask HN: What language-agnostic programming books should I read? - https://news.ycombinator.com/item?id=14486657

41. Ask HN: Where can I find high-end stock images for a website? - https://news.ycombinator.com/item?id=15602538

42. Ask HN: What overlooked class of tools should a self-taught programmer look into - https://news.ycombinator.com/item?id=19900955

43. Advanced Data Structures (2017) - https://news.ycombinator.com/item?id=20044876

44. Immersive Linear Algebra (2016) - https://news.ycombinator.com/item?id=19264048

45. Ask HN: How to get started with machine learning? - https://news.ycombinator.com/item?id=12713056

46. Tools for Better Thinking - https://news.ycombinator.com/item?id=23339830

47. Ask HN: Best books you read in the past decade? - https://news.ycombinator.com/item?id=21900498

48. Startup idea checklist - https://news.ycombinator.com/item?id=20254057

49. Ask HN: Favorite teachers on YouTube? - https://news.ycombinator.com/item?id=17999659

50. Ask HN: What startup/technology is on your 'to watch' list? - https://news.ycombinator.com/item?id=23276456

Thank you! Now I just wonder - did you use some command line magic to do this, or did you compose it manually? (no complaints if you did, but it would be interesting to know how to do that in a line of script). Anyway, thanks again.

You can do this with Notepad++ and Block-Selection in like 1,2 Minutes:

- Paste the original list

- Block-Select (Alt + Pressed Mouse1) a block-line behind (trailing) all entries; I admit this step is hard to describe with words, but currently im unable to provide an image/video

- Paste "https://news.ycombinator.com/item?id="

- Block-Select all ids from original list and paste it at the end

- You may be finished now or you can Regex-Replace " +" with " " to remove the long spaces between the link title and the link itself.

Block-Selection + Copy-Paste is amazingly fast for doing things like that.

Pretty dirty solution, but:

cat list| cut -d ' ' -f 4 | xargs -I{} echo "https://news.ycombinator.com/item?id={}" | paste list - | cut -d ' ' -f 5-

Sure. I just did it quickly in Python, it was all of about three or four lines long. I've gotten a bit rusty in Python so when I see a small task I try to give it a go.

Consider adding this comment on the grandparent. Thank you!

Favorited, thanks!

> 425 24351073 Most favorited Hacker News posts of all time

Interesting that a post about most favorited Hacker News posts is itself one of the most favorited Hacker News posts.

I totally missed that yesterday.

Could you please post this as a “Tell HN” post so that we can all favorite it and make it the most favorited post?

Jokes aside, such an official list posted once in a while (once a year) could be useful.

Much more interesting list, thank you.

Maybe people are favoriting Ask HN items because there are no external links.

I know that when there's a good HN discussion of a link, I usually bookmark the link itself, knowing that I can easily use that to look up HN discussions of it later. But if there is no external link, the only option is to bookmark the HN discussion itself.

Ask HN are just so hugely valuable. There are so many smart, experienced people here who are generous enough to post hard-won insights and information.

Yes! Same.

I think it's interesting at how many of these are "Ask HN" type posts. It could be that people are often seeking to gather specific crowd-sourced knowledge from experts, what happened to Quora?

I disabled my Quora account when I discovered it was selling my data.

I got rid of mine when they started paying people to ask questions, and the quality of questions plummeted. Also, the clearly fake "so-and-so would like you to answer a question" email notifications annoyed me.

SEOs happened

This community happened to it. The number of top-voted hacker news blog posts talking about how they acquired their first n users by shitposting on Quora about their startup are too many to enumerate. I think the difference between here and Quora is that the mods here are more opinionated and will kill posts that feel spammy/use voter rings/etc.

External links from Quora have the rel=nofollow attribute, so they don't help with SEO. Most SEOs would understand this.

Ask HN is the best feature of HN bar none and it is unfortunately under exposed as well. As far as I am concerned it could be the homepage.

What about most favorited comments?

140 https://news.ycombinator.com/item?id=4247615

138 https://news.ycombinator.com/item?id=15603013

108 https://news.ycombinator.com/item?id=18442941

93 https://news.ycombinator.com/item?id=13436420

86 https://news.ycombinator.com/item?id=8902739

81 https://news.ycombinator.com/item?id=11042400

81 https://news.ycombinator.com/item?id=14948078

76 https://news.ycombinator.com/item?id=6199544

65 https://news.ycombinator.com/item?id=12901356

63 https://news.ycombinator.com/item?id=35083

60 https://news.ycombinator.com/item?id=7135833

58 https://news.ycombinator.com/item?id=14691212

57 https://news.ycombinator.com/item?id=35079

57 https://news.ycombinator.com/item?id=18536601

55 https://news.ycombinator.com/item?id=9224

55 https://news.ycombinator.com/item?id=21260001

54 https://news.ycombinator.com/item?id=16402387

53 https://news.ycombinator.com/item?id=9282104

53 https://news.ycombinator.com/item?id=23285438

52 https://news.ycombinator.com/item?id=14791601

51 https://news.ycombinator.com/item?id=9440566

51 https://news.ycombinator.com/item?id=22787313

50 https://news.ycombinator.com/item?id=12900448

49 https://news.ycombinator.com/item?id=11341567

47 https://news.ycombinator.com/item?id=19604657

42 https://news.ycombinator.com/item?id=20609978

42 https://news.ycombinator.com/item?id=2439478

40 https://news.ycombinator.com/item?id=14852771

39 https://news.ycombinator.com/item?id=12509533

38 https://news.ycombinator.com/item?id=22808280

38 https://news.ycombinator.com/item?id=16126082

37 https://news.ycombinator.com/item?id=5397797

37 https://news.ycombinator.com/item?id=21151830

37 https://news.ycombinator.com/item?id=19716969

36 https://news.ycombinator.com/item?id=17022563

36 https://news.ycombinator.com/item?id=19775789

35 https://news.ycombinator.com/item?id=11071754

33 https://news.ycombinator.com/item?id=20571219

33 https://news.ycombinator.com/item?id=7260087

33 https://news.ycombinator.com/item?id=17714304

32 https://news.ycombinator.com/item?id=22043088

32 https://news.ycombinator.com/item?id=18003253

30 https://news.ycombinator.com/item?id=341288

29 https://news.ycombinator.com/item?id=7789438

29 https://news.ycombinator.com/item?id=9048947

29 https://news.ycombinator.com/item?id=14162853

28 https://news.ycombinator.com/item?id=20869111

28 https://news.ycombinator.com/item?id=19720160

28 https://news.ycombinator.com/item?id=287767

28 https://news.ycombinator.com/item?id=1055389

Ha, I see three of those are already gray-visited for me, and clicking them now, I see why!

* https://news.ycombinator.com/item?id=18442941 was a great look at Oracle DB development

* https://news.ycombinator.com/item?id=35079 was one of the most hilarious situations of all time on HN

* https://news.ycombinator.com/item?id=9224 was a fun early Dropbox "No wireless, less space than a Nomad, lame" moment.

I'm gonna have to check out some of these others that I haven't seen before! Great list, thanks!

Re 9224, I feel bad that BrandonM has been singled out that way because the comment has been universally misinterpreted. He was commenting on Dropbox's YC application (that's what "app" meant back then) and trying to be helpful.


Top comments dont work so well as there is no curated titles. The content is good but may not be relevant to the reader

I think a prefix of the comment string is in most cases sufficient to disclose what the comment is about. I tried to include that above, but it got too tricky for silly technical reasons so I bailed.

The fact that favorites are a hidden feature and hard to use made me think the quality of the curation signal would be even better. Also the fact the HackerNews API, Angolia don't expose this feature made it even more interesting to aggregate.

Hope you like it and as it was a pain gathering the data, I put it in an observable notebook so hopefully we don't need to gather that dataset again for a while.

The learning resources were the most useful thing to me. I have seen most of those links whoosh by me when reading HN, but this list has made me revisit some of them that were lost to the sands of time. In particular the bash shell resource [1] is something I have been trying to re-find for ages!

[1] https://news.ycombinator.com/item?id=17057596

When I emailed mods to argue that the 'web' button should be restored[1], I was told by dang that he was actually considering removing 'hide' and 'favorite' too, because people rarely use those features. (I routinely use 'hide' but hardly ever use 'favorite'.)

Btw, 'favorite' may be lesser-known but it's certainly not hard to use.

[1] https://news.ycombinator.com/item?id=23762479

Favorite is one of my most used features. Please don't remove it, dang! :(

I frequently go back to these stories, whereas I don't often care to find things I upvoted again. If the two features become conflated, I'll probably stop upvoting so that I can use upvoted stories and posts as a bookmarking system.

Please keep favorite stories and posts around.

If you're worried about feature utilization, maybe name it "save" for better discoverability. That's essentially what I use the feature for.

I don't trust or use browser bookmarks much anymore. HN favorites are high quality bookmarks that work across all my browsers and devices.

Yes, like you I use favourite routinely. Please don't remove it.

Favourite is great. I use it sparingly for obvious reasons - I only want truly important items in my favourites list.

Save is a better name. OP has updated list of favourite stories and many of the Ask HN posts make the list because they often yield a lot of useful information in the comments. I use SQL on a daily basis so most of my favourites are SQL stories. I use favourite to bookmark post and comments.

Removing "hide" is the wrong call. People, especially in work environments, need to be able to hide links for all sorts of reasons. For example, if you work at Apple and there is an inflammatory article or if you've recently lost a family member to suicide and there is an article on a famous figure in the tech community committing suicide like Aaron Swartz, you probably don't want to see that on the homepage of HN for the day. It's one of the best features of Reddit and HN, imo.

Hide is super useful for threads you want to read but have the typical knee-jerk flamebait war responses at the top that turn into massive comment trees that you want to skip.

A lot of times the good stuff is at the bottom.

So much this.

The population of commenters seems to be dominated by those whose goal is to argue and win, rather than to understand and learn. So many times I see comments that make no attempt to take context into account, and instead conflate all sorts of things, making it impossible to comment productively. (Basic example: Top comment has not replied to the thesis of the article, but to something else.) When one of these types of comments is at the top of the comments list I just hide the whole story, because there’s literally nothing to see but people talking past each other.

Hide is important to me too. I don't want to be worked up by the threads that I know will br charged that I'm tempted to click on.

Seconded. I hide threads I know I won’t be able to have a productive conversation in.

This may have been referring to the "hide" link on the HN homepage, rather than collapsing threads.

I wish you could hide stories by keyword or by domain.

I feel like that should be doable with a ublock rule

> was told by dang that he was actually considering removing 'hide' and 'favorite' too, because people rarely use those features.

That's something you do with a corporate site where you're constantly A/B testing and aggressively removing features that don't generate revenue. Hacker News isn't (or shouldn't be) run that way. It's perfectly fine to have features only a few people use.

Removing features to decrease maintenance overhead is also a pretty legit reason, and you don't have to be a corporate site for that to apply to you (I would say it applies even more to a hobby site).

That said, I really hope favorites isn't removed because even though I only discovered the feature a few years ago, I really cherish the comments I've favorited.

"decreasing maintenance overhead" could apply equally to every feature on the site, though. How often do people use polls?

there are polls?

> there are polls?

...Exactly. It's been a feature forever, it complicates the codebase, and no one ever uses it.

Two examples[0,1].




Hardly ever see it used in any capacity though.

Re gp: removing buttons shown on every page is probably more about simplifying interface and reclaiming space than maintenance overhead.

there's no space to reclaim, honestly. If removing a 12pt line of text saves space, please buy a phone that supports more than just WAP.

Also, “it’s not used frequently” is a really bad reason to remove a feature. Frequency of use is not a proxy for usefulness. I don’t use my backup software’s “restore” feature often but that doesn’t mean it should be removed.

What's the maintenance cost of the web, save and hide buttons now that they've been implemented and working for years?

Standard MO of software engineers is to rewrite code into the latest and greatest framework/language/style every couple of years to keep the resume fresh.

That hardly applies to HN, which still runs on Arc like it did 13 years ago.

Yikes, what a horror: either no new features, or complexity increases forever. Fortunately we don't live in that prison. And what a dystopian reason you give for removing things!

The humane reason, of course, for removing features, is so you can add new ones while staying simple. I have always found this to be a deep part of, let's call it, healthy software development, at every level. If you remove things as well as add them—code, features, complexity–then the system can breathe over time. When I say "the system" I include the programmers and the community as well as the code—the whole thing. If you're only allowed to add, that's no longer a creative process, just agglutination.

>Yikes, what a horror: either no new features, or complexity increases forever. Fortunately we don't live in that prison.

And fortunately, I didn't suggest that we do, or that features should never be removed... rather that removing a feature that people use because not enough of them use it doesn't make sense for a site like this.

If you want more people to use new features, I think you need to put more effort than you do into making those features more visible. Feature discovery is a problem - there's plenty of anecdotal evidence here to attest to that.

I can think of a couple of possible solutions, such as having new features be bold for a while after being first introduced, so they're more easily noticed, or pinning a thread announcing new site features, or having something like a devlog listing possible and new changes and comments related to development, to better engage the community in the process (or at least raise their awareness that the site is still being actively worked on,) etc. Or all of those.

Totally agree that feature discovery is a problem.

I'd personally find removing favorite rather frustrating as it's a nice way of differentiating posts than just upvoting so that I can easily find them again in the future...

Oh I hope he doesn't remove favorite. I use it all the time.

I really really hope they don't remove favorite. I have a lot of comments and posts saved and return to read them from time to time.

I don’t use favourite often but that’s the point. Favourites are 100% signal

I didn't even know there was a 'hide' option. Just never looked at anything below the title, really! Been here for a while too

Favorite gets indirectly used when an up vote cannot be given to a post and a thread is what I've noticed, which would beg the question what would happen if those buttons got removed? I think it would be a shame if hide and favorite were removed, it might be a minor feature but I have used it in the past to make it easier to ignore certain posts.

I've been missing "web" quite a bit and was just considering emailing him to ask if it was now okay to flag links to inaccessible content!

Why wouldn't it be? Unless you're trying to flag stuff a year after it's been posted, I would expect at least that a posted link would be accessible to read.

Unless what you mean is "inaccessible" in the sense of "not accessible to people with disabilities," in which case, good luck flagging a large percent of everything that gets posted. :/

Unfortunately, links often go to paywalled content. They each waste my time and make HN worse.

I have hundreds of favorite comments.

The "favorites" feature is amazing and just rightly inconspicuous. On my website, https://mihir.ch/#hn, I've been collecting and displaying my favorited posts and comments from an API I created for this purpose only.

The API: https://hn-faves.now.sh

Source: https://github.com/plibither8/hn-faves-api

Thanks for sharing your source too! I was curious how you got a web scraper on a small now.sh server (45mb)

Usually I'd use puppeteer, but that would make it like 400mb. I've never heard of [1] x-ray, it looks a great package for lightweight containers!

[1] https://www.npmjs.com/package/x-ray

How did you get the list of HN users to scrape favorites from?

BigQuery public dataset https://news.ycombinator.com/item?id=19304326

Most active was defined by comment count.

How many of the top 10k most active members had favourites?

I didn't know about the favorite feature until this post. I may end up using it now.

it's nice that HN at least acknowledges that whether you like something or not is something that only matters for you, and no one else, rather than pretending there is value in centering a surfacing algorithm around it.

Hidden? They're in plain sight at the top of every article page.

I've used this website since 2010 and just realized it was a feature. Maybe I'm just unobservant...

It's relatively new. Before we used upvotes. Actually I still don't use it and keep using upvotes to favourite stories. I'm sorry for that and even worse: sometimes I upvote a story without following the link, just using the upvote as a "read later" bookmark because I found the title intriguing. Not so bad is when I upvote because there's a very interesting comment, even though the article itself is crappy.

Time ago someone posted a script that collects upvoted stories, but I can't find it now, it's in one of those upvoted stories O:) I guess I could make a conversion... but there are probably more than a thousand upvoted stories, I don't know.

Actually I still don't use it and keep using upvotes to favourite stories. I'm sorry for that

So do I, but I'm not going to apologize for it. Upvoting is the fastest and more convenient way to save a story as "read later", so it's what I do. If they would add the "favorite" link to the front-page items, I'd use it instead of upvoting.

Not so bad is when I upvote because there's a very interesting comment

If you didn’t know, you can “favorite” comments, too.

That has become my favorite comment :)

It's not the most used function because browsers already have a 'favorite' function, bookmarks. I used a scrapbook extension in FireFox for years so even though I know it existed I've never used it. But this post actually serves to shine some light on it and I'll give it a shot for a few weeks to see how well it works for me. HN has some amazing threads every now and then and it is really worth remembering those.

Same here. Been here since the first week of HN, just learned about favorites.

HN's minimalist UI designed to make elements unobtrusive and sometimes difficult to see, and new features being mentioned only in a single thread means most users will just not notice. It also doesn't help that HN is a conservative culture generally antagonistic towards any change to the layout, and people wouldn't want to notice those changes to begin with.

I've seen people not notice thread folding or the hide feature despite those also being "in plain sight." People don't know threads can span multiple pages - dang has to go out of his way to point that out because despite the UI being exactly the same as on the list pages, a "more" link at the bottom of the page, no one ever reads past the first thirty links on those pages to begin with.

Those are not mutually exclusive facts. You're zero words away from the phrase "hidden in plain sight".

To favorite a comment, you need to click on the time to be shown the favorite option.

See other comments. Maybe higher friction would be a better phrase

The favorite feature is hidden in plain sight.

What does Angolia have to do with HN?

Edit: found it, it's Algolia. Thought s/he was talking about Angola.

I have never managed to query by its name ever. I default to googling for hn search when I look for it.

I personally find the comments far more interesting than the stories. Does anyone know some HN users with interesting favorited comments?

If anyone thinks the list of posts here is very heavy on programming things, feel free to browse my favorited comments. I have quite the mix of all topics. The usual disclaimer applies - these are just my HN favorites, I do not endorse them. Many favorited comments are completely wrong, they are not medical, legal or financial advice, I am not licensed to sell insurance etc etc.


Dan Luu has a nice curated list: https://danluu.com/hn-comments/


I favorite comments often. I like to scan back through. I generate a lot of ideas and make a lot of connections doing so.

Same caveats apply to my favorites as yours.

Ditto! I use it for a) things that are worth researching later and b) unusually well-made points and explanations.


You have a nice list, thank you! Hopefully other people post theirs as well, I'm very curious what people favorite.

Is it possible to favorite a comment on HN?

Yes, click the timestamp on the comment and you'll see the full range of actions.

Yes, click the timestamp on the comment and you'll see the full range of actions.

Very interesting, after the very active recent discussion about discoverability in user interfaces.

It's like HN hired someone from the iOS design team to come up with the notion of "timestamp" = "favorite."

Another way to look at it is that "timestamp" = "permalink", which is a fairly common pattern (I'm not trying to imply that it is intuitive, just that it is common). Once you are at the permalink you have a range of options available to you.

Same, as a permalink. I'll bookmark the permalink then run it through archive.org using a simple addons of mine, in case HN goes down I will have an archive backup.

If it were iOS design, the favorite button would only appear when you rotated your device.

Putting more stuff under the permalink page linked from a timestamp is a web tradition that's older than HN. I don't know if it's good, but I do notice when I can't click a timestamp. Even Twitter and Facebook abide it.

I didn't know that. This was very helpful. Been here for a while too (a few years passive read prior to making an account).

My god, after a decade. Thank you.

Ah, nice. Thanks.

Yes. I just discovered that I have done it by mistake, twice during the past six years. I'll blame glitches on the trackpad on my previous laptop. I wasn't even aware of the feature.

Only once you click a comment's date to access the permalink.

Completely agree, I almost always screen comments before I get to the article.

Honestly, I’m kind of surprised more people do not use the favorite feature to bookmark interesting links / posts. 96 favorites on the most favorited seems really low.

Those numbers are so low because the user list was truncated, and then restricted only to 30 per user.

The globally most-favorited is https://news.ycombinator.com/item?id=19087418 ("Ask HN: What books changed the way you think about almost everything?") with 836 faves. You can see the global top 50 here: https://news.ycombinator.com/item?id=24355617

I honestly never even noticed it was a feature until you just mentioned it here.

I'm in the same boat. Even knowing it's there now, I have to look hard for it; there's no visual anchor to draw my eye there.

I've seen it there for years but have no idea what it does. I don't really care-I bookmark things I like.

Ditto, I didn't ever notice it and I've been on here for years.

If you click on the time bit '14 minutes ago' in a users comment you go to a page where you can favourite individual comments too.


I'm on 10k+ karma and I literally didn't know it existed.

20k+ karma here and same. I haven't favorited so much as a single article. Didn't realize it was an option.

So the top favorited list is probably biased in some way towards the type of people more likely to discover and use this feature. I just don't know exactly what that entails.

I've noticed these and that's probably only because I trained myself to look carefully at stuff (that was a lot of work too). Small, light grey font on a beige background, well, what did HN expect.

Yeah, that's a separate but related issue: HN's UI is too low contrast to the point of posing an accessibility issue. It kind of tells you everything you need to know about YCombinator (and maybe SV VC in general) that its biggest watering hole doesn't make even this basic concession towards accessibility.

I agree but HN does work rather well without JS enabled, so credit where credit's due.

I use it pretty often. There are stories, show HN apps, or just random coding things that I feel I may want to have access to in the future, but are in a domain don't have enough of an interest in to bookmark. I guess if I had a generic "misc." bookmark folder that I didn't care to organize it would go here but I like keeping browser bookmarks minimal.

One thing I wish were possible is to search through upvoted content and favorited content by title and even comments in either. I have upvoted so much stuff a lot of it becomes harder to find if I cant narrow it down by what is relevant to me. Its the one killer feature HN is missing for me. Outside of that one thing I think HN is perfect thus far.

Mostly the favorites feature is used by me when I cant upvote something since thats what HN defaults to. I would use it more often if I could filter by my own saved content.

Oh, sorry, I didn't read / comprehend your comment properly earlier.

Firstly, try contacting hn@ycombinator.com detailing what you'd like to build and see if they're interested / able to collaborate or provide you the necessary access / APIs.

I'm not a programmer by any stretch of the imagination, but have contacted the mods from time to time and found them eminently approachable.

Secondly, if you can't see the up or down vote buttons, try clicking on a comment or submissions time stamp immediately after the username. For reasons I don't know sometimes the vote buttons aren't displayed in the regular view.

No worries! I was just clarifying, I know sometimes I don't express myself as transparently as I'd hope. I might have to follow up on that suggestion, I didn't even think about that, I've wanted this feature for so long!

Unless I'm mistaken, HN exposes that data so anyone can build those features as either a browser add-on or as a separate site built on the data?

One of the things I super appreciate about HN is the (almost) complete abcense of feature creep.

I don't think I've ever seen this on any third party search tools for HN. If it's an actual feature it is highly under documented. The filtering that HN does is on all comments / stories (by post title) but you don't get to filter down by what a user has upvoted since that data is private to the user, only favorited data is public.

I don't think this is a feature creep, I think it's useful. Otherwise I'm going to have to code something to scrape through HN for me and provide that additional context, and even then I still would be missing a way to filter down on HN comment data vs my own local db of upvoted materials.

That's because I already have bookmarks in my browser, marking articles as favorite on HN would give me no value.

HN Comments is where the money is, not the stories themselves.

Comments have links that can be bookmarked too, under that timestamp, here's yours:


I usually favorite a whole story so that I have access to all comments, not just one.

I didn't know that comments can be bookmarked, that's a very unusual and neat feature

Unusual? Reddit and Twitter have it, GitHub/GitLab have anchors too, which platform lacks comment links?

I switch computers frequently (desktop, chromebook, laptop, phone, etc.) so having some kind of centralized 'favorites' is nice. I've not used the bookmark sync with Firefox or such.

I've noticed that I rarely go back and look at my favorites though.

I download the articles, so I can even read them in the event they are taken offline.

I use Pocket which then gets IFTTT'd to Pinboard which archives them for me.

Better imo to use a bookmarking service that works across the entire web. I used to save facebook links and bookmark tweets too because the features were in those platforms, but eventually I just started using Pocket like I use for everything else on the web.

Because "stories" (i.e. top-level posts) are easily searched and retrieved. Individual comments however might have more value being favorited, and while the numbers are likely similarly low, they would make for a better blog post.

On the site since 2012. Never noticed favorite until right now. Was surprised to see I had 3 favorites saved already which all must have been inadvertent clicks.

Seems pretty silly that you can only favourite the comments view - if you click directly into the article from HN homepage you never see the favourite link at all.

I don't use it and use bookmarks instead, but if you just look for points https://hn.algolia.com/ works pretty well.

Spoiler: Steven Hawking has died.

Why would I let my bookmarks get trapped inside the site owner’s database?

I use the mobile app and its horrible. I can't often save comments, i can't see my comment threads, ...

There are probably many HN features nobody knows about because the UI doesn't expose them.

Why are you not using the website? It works really well on mobile and loads almost instantly. I’ve never seen a reason to use app for HN.

I only ever use private browsing and don’t want to have to log in every time I want to upvote someone.

The content is saved unless I explicitly reload or navigate away from the content. So I can close the app and come back to exactly where I was rather than the browser that will not save that state and then try to reload even if I don’t have an internet connection.

Navigation is easier in the app, because every action is clicking anywhere on the post and selecting an action from a large pop up menu. Except for collapsing/expanding which is also easier because you click the post header not just a tiny dash.

Automatic dark/light mode.

Etc. Etc. The app I use isn’t perfect but it’s preferable for lots of reasons to visiting the site directly.

I also only use private browsing mode, but the login annoyance is not enough to make me use an app. Most apps in fact track like crazy (potentially even third-party apps for sites like HN), and the whole point of private browsing is to not be tracked.

Instead I just lurk most of the time since I'm too lazy to log in, which means I'm not really contributing with votes. Maybe what we need to do is reduce the friction in logging in. I've always turned off the browser's password saving features and use my own password manager; maybe I should reconsider that policy.

[edit] Nevermind, apparently Firefox can't remember passwords when in permanent private browsing mode. So much for that idea :/

My only issue with the website is the little minimize "-" sign on comments is too small and difficult to reliably hit without zooming in. However as you say everything else works much more smoothly than any of the apps I've tried

It works fine, but the font is teeny tiny.

I can reliably zoom in the text here on my Pinephone. No left/right scrolling is ever required.

Which on are you using? There are many.

I use this. It is very good imo -https://play.google.com/store/apps/details?id=io.github.hidr...

I'm using one called Hacker News (YC) from Dharmesh Patel. I thought that was the official app.

> "The mobile app"

Which mobile app do you mean? HN has no app. But there are many third-party ones.

This is actually leading to some unanticipated problems. I wrote about this yesterday: https://news.ycombinator.com/item?id=24335289.

I use Materialistic. A HN app I found on Fdroid, and I'm very happy

Didnt know there was an app. THe website renders fine for me on mobile.

A lot of people have pointed out alternatives, but the killer feature of eg bookmarks (or maybe pocket?) is the categorization.

I have bookmarks that I've planned to review in 3 months. I'll have bookmarks for a tool I plan to try out on a future side project

I don't really care about the "how" (I take notes for various things -- books, movies, apps -- all over the place). But bookmarks are convenient and allow for the simplest of categorization, which I value

Maybe people tend to favorite different things. I favorite things from time to time, but don't recall favoriting anything in this list.

Not really surprised anymore because this has been discussed in the past but still perplexed why this is the case. I use the feature intensively and I find it useful. UX could be improved though, especially I'd like to be able to favorite a comment with a single click/tap, very much like voting up or down.

You may use it intensively, but as far as I can tell, you have only 27 favorites at the moment (https://news.ycombinator.com/favorites?id=weinzierl)

I guess there also are a lot more stories (over 24 million, it seems) than regular users. Even taking your 27 as the average per user, I would expect the average number of time an article gets favored to be significantly below 1.

Indeed I do have hundreds of favorites, going back till 2012. What you are looking at are just my favorited stories, but I have a lot more favorite comments. Funnily your comment kind of supports my point that the UX could be improved;-)

Also a disclaimer for those looking at my favorites: I use this feature much like a bookmarking service and for me a favorite means an endorsement only in the sense that I found the contribution interesting and valuable enough to be remembered. I don't necessarily agree with what is said in the postings I favorite.

Aha! Thanks. The UI can be improved, yes.

Another hidden feature is 'top links'(https://news.ycombinator.com/best). It is not visible to me (may be because of low karma), but if I directly type it in the url it works.

It's hidden behind "Lists" in the site footer along with some other pages.

Thank You. I wasn't aware of that section.

more people do not use the favorite feature to bookmark

Why would you, when upvoting achieves the same effect and is available with one less click? If you could "favorite" an itme from the front page then I expect the feature would get more use.

On that note, if there's a feature that I would like to see implemented on HN is that favoriting an article shouldn't redirect to the favorites page, instead just working with AJAX (do people still use that term?)

Wow. Abismal number. I dreamt a side project of using that info to estimate “taste distance” between users and see what personalized recommendations come out of it, but I dont think it would work with those numbers.

Besides being tied to your account, the only plus I see of HN bookmarks over browser bookmarks is that your HN bookmarks are shared with the public. Some people may want to keep their favorites private, though.

Why would I use this instead of my browser's bookmarks, which have tags, folders, and search, and are less likely to be discontinued by the product storing them?

For me, the answer is that my browser bookmarks are tied to a physical machine and specific browser. I use many machines, and sometimes many different browsers on each.

I have another tool (Pocket) that I use for bookmarking interesting posts for later reading/reference that works independently of which site I happen to be on.

It's lower than it could be since only the most recent 30 favourites of the top 10k members are included.

But still lower than expected.

This is exactly my use case.

Yeah I never noticed it

These days most people don’t use Internet Explorer.

I... never knew “favorite” was a feature on hacker news until I read this post.

Now I see the “favorite” link right above this comment form. I guess I never even noticed the links on the post page are different than when browsing stories on the front page. Now I’m noticing there’s also a “hide” link!

I never knew about the favorite “button” as well. I think the “button” might be more clear if a unicode “heart” was used instead.

And elsewhere in this thread, I also discovered you can favorite comments too (by clicking on the timestamp to get to the permalink page for the comment).

Everyone favorite this post so we'll get an infinite loop.

I've occasionally looked at the favorites of highly active users (e.g. the list of karma leaders), and I've rarely found other users to have used the feature. As for me, I love using it – it has the right amount of friction such that I only end up using it for things that really are my favorite. I never look through my own upvote history. But whenever I need to find something that I remember being great, it's almost always in my favorites list.

My only complaint with the favorite feature is that after favoriting, it takes you to your favorites page, versus an async request leaving you in place with the comment or post favorited.

That's on my list to change. It'll work like voting does.

That's funny, I always thought it was intentional. Because the slight inconvenience of the favoriting action definitely makes me more selective in using it.

I suppose the public nature of the favorites list (as opposed to upvoted list) will still have that self-filtering effect.

It was intentional, but the intention was feature discovery. We did it that way in the hope that people would get to know the favorites page and start browsing the favorites of others. It doesn't seem to have had that effect; it's just a pain. I suppose we could put up a confirmation screen first: "Do you want to add 'foo' to your favorites?" But that might just be a pain too.

There's another discovery issue, which is that not everyone knows that favorites are public. Publicness was the primary reason we built the feature: the hope that it would be a new way of finding interesting stories. One can see that happening in this thread to some extent. Overall, though, I was disappointed that it didn't have more of this effect, which is why I considered removing it.

Perhaps we just didn't do enough to promote it. Someone suggested having a quarterly post about the most-favorited stories. That might be a good way. We could have different most-favorited lists based on different subsets of users. We could ask people which accounts have the most interesting favorites. and so on.

It's appreciated.

I wrote some JavaScript and Python tools to search your HN Favorites and export them to CSV or HTML.

The JavaScript runs in your browser like a browser extension. It scrapes the HTML and navigates from page to page.

Setup and usage instructions are in the files.

JavaScript: https://github.com/gabrielsroka/gabrielsroka.github.io/blob/...

Python: https://github.com/gabrielsroka/gabrielsroka.github.io/blob/...

Previous discussion, other tools and more info: https://news.ycombinator.com/item?id=22788236

Why use website-specific favorites instead of browser favorites? Isn’t it inconvenient to have bookmarks siloed across different websites? My browser bookmarks work on any website, are synced across devices and backed up automatically (Safari), and are easily searchable and organized via folders or tags.

I use it as a super-upvote.

I try to only favourite the cream of the crop, in the hopes someone will stumble into my favourites list and enjoy them.

I mostly favourite comments not submissions though.

> in the hopes someone will stumble into my favourites list and enjoy them.

I did. Thank you!

The only benefit is that the HN favorites are public (to HN) so you can get aggregate data like this.

For the individual user there is zero reason.

Interestingly, https://news.ycombinator.com/item?id=21429740 is both on this list and flagged. Apparently people have very strong and diametrically opposed opinions on cancel culture, leading to a lot of flagging and a lot of favoriting of the same article.

I quite like the favorite feature. I did not know until reading this post though that you could favorite comments as well - I'll be sure to utilize that in the future.

Others have mentioned that the favorite feature is of little utility because they are using other bookmarking software (browser/browser extensions, etc.). I would say, personally, that I still love the favorite feature on HN because my typical collection of bookmarks contains a whole variety of things - history, politics, music, self-help stuff, etc. The stuff I read and favorite on HN are of an entirely different set of information and curiosity and thus, I really do find it useful to have them all collected in one convenient location.

I was surprised to learn that I wasn't using the favorite feature!

I primarily browse HN through the Android app "Materialistic", which has a prominent "save" button at the top of each story.

However, those saved stories (~1.1/week) aren't favorited. Just saved locally.

Could this be part of why the feature seems so rarely used?

Thanks for the link! I've been looking for such a feature that shows me the most popular posts within a given timeframe. My thought process was that, as much as I enjoy discussions on HN, I don't want to spend too much time on it and, in particular, I want to avoid browsing HN out of FOMO multiple times a day. Instead I'd rather just schedule half an hour to an hour every week to have a look at the most popular posts.

There is a best link https://news.ycombinator.com/best

I don’t fav any posts that I like, but upvote them. I also up upvote the ones I want to bookmark, so that I can later come back to read more comments. I don’t really understand why I would need to fav a post.

Of course now this post will become the most favorited HN post of all time. I certainly marked it and reminded myself to use that feature more often.

It might be embarrassing to admit but I use 'threads' as my bookmark service - if something was engaging/striking/funny enough for me to comment on, then I want to come back to it.

At some point I will write a thingy to go grab all my old comments but for now HN is more useful to me than pocket / delicious etc.

It turns out to be a useful heuristic - the mental process that says "I should store that for later" fails far more often than "that is something I have something vaguely useful to say about".

I find it amusing that the #19 on this is flagged dead.

What's amazed me down the years is where some random user will reveal they are the developer behind some well known software, or the author of a famous book etc. Then when you look at their profile you'll see that they've been a member for years but only have a small amount of karma.

Clicking on the timestamp to get to a comment so that I can favorite it, is not that intuitive either.

The list will be inherently biased for recency (note that “Ursula LeGuin has died” is high on the list) simply because there are more people around to vote.

Doesn’t make the list meaningless, but it may not surface the “best” articles that might interest you.

I posted the global list here: https://news.ycombinator.com/item?id=24355617

I'd like to see collaborative filtering applied to this.


You mean "people who favorited X also favorited Y"? That could be interesting. Not sure there's enough data.

Notice that the 19th most favorited post is also flagged. ("Those People We Tried to Cancel? They’re All Hanging Out Together"). To me, that indicates the flagging mechanism on hn is broken.

Have you actually read that thread? It's a garbage fire.

To me, flagging worked as intended.

Apparently quite a lot of HNers disagree with you, and liked it well enough to save it. That should be enough to indicate the article is appropriate for hn.

And the low quality of discussion it produced and the number of flagged comments should be enough to indicate that it wasn't. Because people don't just flag for the article, they flag for the thread as well.

Then the comments should be flagged, not the article. There wasn't anything wrong with the article, and apparently it was interesting to the hn crowd - which according to the guidelines is exactly what makes it acceptable.

EDIT: to respond to your edit, there was an entire thread discussing this in the article discussion. The article should be flagged if there's a problem with the article, and comments should be flagged if there's a problem with the comments.

This is cool. I think it would be helpful if the list showed the date of each post as well. That’s an important signal for me when deciding which links to explore in a curated list.

Surprised to see my project at the top, considering how old it is. Also surprised by the very low number of favorites overall; I use that feature quite often.

Thanks to the poster and the slidedeck's writer, I have finally convinced and managed to read the Machine Learning 101 slidedeck [1]. I'd recommend you guys to read it as well, very informative reading.


Wow, I've been lurking HN for years and I had never noticed or used the "favourite" button before

I just learned a few things about music and synth and CSS because of this. Also, going to favorite THIS post.

Possible selection bias here - How long has the favourite feature been available in its current form?

That is very good, if only a HN-lore exist, there are some links that are submitted to HN regularly, and every 6 months new people discover them. Great stories people here like sharing, articles, books, etc.. create a ranked list of HN folklore and traditions.

I use the favorites feature as a kind of "read later" list (which I never do).

I made this for scraping favorites a while back. https://github.com/reactual/hacker-news-favorites-api

clicks favourite

Maybe if the favouriteness could be transitive in some way? It could be called “favouriterank” and we could use it to index the web and surface the best pages when people search for things they are interested in?

The original intention of the feature was to try to derive some interesting signal from collective usage (that's why HN favorites are public) so that's an interesting suggestion.

I confess: I didn’t write it in a particularly serious tone (pfff, “favouriterank” indeed — a comparison to the early days of a pre-monetized google.com) but behind it was a serious idea.

I appreciate the charitable response.

When I saw “favorited” I thought that was referring to what I guess is “upvoted” - as others have mentioned, I didn’t even realize that was a feature. Is there a list of most upvoted HN stories of all time?

How many HN members are there?

Is any data available on their locations? Demographics?

It would be nice to have the feature action on the home page itself.

Where can you find the list of most active Hacker News users?

I used BigQuery public dataset. I scored active as the quantity of comments


It depends on how you define "active", but if you click on the Lists link at the bottom of this page, and then on Leaders, you find an ordered list of the users with the most points: https://news.ycombinator.com/leaders.

HN's hall of shame.

I think it's interesting that out of all the news articles about famous people that have died, Ursla le Guin is the one that turns up in this list.

TIL HN has favorite functionality!

Fools. They could've only "favourited" this post taps temple.

I will fav this post, so your list will have a chance to be recursively reachable.

Was just thinking this as I liked it.

It's interesting how few favorites the articles actually have.

Those numbers are artificially low. If you want to compare the global list it's at https://news.ycombinator.com/item?id=24355617.

I did use favorite features often for both submission and comments.

What are the most upvoted HN posts of all time?

What about most favorited comments?

Applications are open for YC Winter 2023

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact