Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Saltiest Users on HN

hackersalt.com

Post reply on HN
Print view View original

61–70 of 92 posts

Re: The Saltiest Users on HN

#61

Post by Kassius509 »

_bxg1 wrote: post #46
logfromblammo wrote: post #37

Earlier quoted context omitted.

Agreed. It looks like they went through a dictionary and scored the words on a negativity/positivity axis, and then just took the mean of all the scoring words in a post. I have written posts very much saltier than the ones scored as saltiest by this ranking algorithm, possibly because I didn't use inherently negative vocabulary to express a highly negative sentiment. It's a fun party trick, but its usefulness is lim…

Honestly I think the whole thing would be more meaningful if it just used total downvotes and/or ratio of downvotes to upvotes

I agree. Would have been much easier too. Unfortunately, HN doesn't have downvotes. The current model does incorporate upvotes, but we're seeing a lot more success training with like-kind labeled datasets + BERT fine tuning.

Thank you for trying the app and hopefully V2 will leave you feeling like the system is more precise.

Re: The Saltiest Users on HN

#62

Post by Kassius509 »

ocdtrekkie wrote: post #9
gnode wrote: post #8

This webpage is very inaccessible. The links aren't proper links, so can't be easily opened in a new tab.

Yeah, I noticed that trying to use the back button sets you back to the front page, not the actual last user you looked at. Similarly, modifying the URL to hackersalt.com/usernametolookfor doesn't actually work, despite being what appears when you get a search result.

Oh that's strange! Yep. Need to fix that! Good catch, thanks!

Re: The Saltiest Users on HN

#63

Post by _bxg1 »

Kassius509 wrote: post #61
_bxg1 wrote: post #46

Earlier quoted context omitted.

Honestly I think the whole thing would be more meaningful if it just used total downvotes and/or ratio of downvotes to upvotes

I agree. Would have been much easier too. Unfortunately, HN doesn't have downvotes. The current model does incorporate upvotes, but we're seeing a lot more success training with like-kind labeled datasets + BERT fine tuning. Thank you for trying the app and hopefully V2 will leave you feeling like the system is more precise.

"HN doesn't have downvotes"

It certainly does; although it's possible they aren't stored independently and simply "cancel out" an upvote, so maybe that's what you meant.

The interface and the graphs are really nice; even though basing the data on votes alone would be less interesting in one sense, I think the rest of the site would still provide value even with that simpler metric.

Re: The Saltiest Users on HN

#64

Post by etrautmann »

aetherson wrote: post #27

It is my feeling that sentiment analysis has a ways to go. Here are a few of the comments I've made that the system has described as among my saltiest: Oh, sorry, "nm" means "nanometer" to me, but of course nautical miles. (Score: -0.5. Comment is entirely taking responsibility for misinterpreting someone) Well, if he was trying 1M combinations every 40 seconds, for $7 per hour, and he didn't need to use hundreds of…

It seems like this is heavily biased towards the most active users: Dang, tptacek, jaquesm, TeMPOraL, etc...

Re: The Saltiest Users on HN

#65

Post by reificator »

mrob wrote: post #33
reificator wrote: post #10

I hope it's not based on the contents of the quoted posts alone. > This was explored (all the way to its grim dystopian conclusion) in tv series Dollhouse. Might want to check it out. That doesn't seem very salty to me, and I could see myself writing that same comment.

It seems to be based on the individual words, not the meaning. "Grim" and "dystopian" are probably considered "salty" words. My lowest scoring comment was about dissonance in music, and presumably "dissonance" is a "salty" word too.

> My lowest scoring comment was about dissonance in music, and presumably "dissonance" is a "salty" word too.

Alright, I'll avoid posting the lyrics to Tool's Schism...

I figured it was something along those lines, it's just very jarring to see harmless comments rank as my saltiest when I know I've posted worse.

Re: The Saltiest Users on HN

#66

Post by asveikau »

wsdfsayy wrote: post #12

"salty" is a great euphemism for "nihilistic asshat"

Well done, this is a great comment! And I really disagree with it.

I actually think "salty" is a very deep misunderstanding of why people are writing such comments, and the attempt to politely call it negative by labeling it "salty" is ironically itself very negative and aggressive.

By making that timid, low-level aggression very blunt and calling it "nihilistic asshat", you are making a very good point. But I don't think they are actually nihilistic asshats. And I think even a true nihilist asshat is worthy of some respect.

Re: The Saltiest Users on HN

#67

Post by asveikau »

jamescostian wrote: post #41
wsdfsayy wrote: post #12

"salty" is a great euphemism for "nihilistic asshat"

This comment I'm writing, as well as the comment I'm replying to, are both "salty" https://www.urbandictionary.com/define.php?term=salty

It's funny because this one sounds much more charitable:

https://www.dictionary.com/browse/salty

Re: The Saltiest Users on HN

#68

Post by Kassius509 »

logfromblammo wrote: post #37
_bxg1 wrote: post #22

The scoring heuristic could use some work; I've already encountered multiple "salty" comments along the lines of "That sounds awful", with a sympathetic tone, probably tagged because of the word "awful".

Agreed. It looks like they went through a dictionary and scored the words on a negativity/positivity axis, and then just took the mean of all the scoring words in a post. I have written posts very much saltier than the ones scored as saltiest by this ranking algorithm, possibly because I didn't use inherently negative vocabulary to express a highly negative sentiment. It's a fun party trick, but its usefulness is lim…

You're spot on. This model is based off a dictionary with scored words!

For the curious you can see the dictionary here: https://github.com/sloria/TextBlob/blob/eb08c120d364e9086467...

The package used is a pretty popular one called TextBlob. It is nifty for working with unlabeled data like we have with the HackerNews dataset.

We really focused our definition of saltiness around being a combination of (subjective + negative) comments.

We reduced the impact of (objective + negative) as we feel that criticism, while at times painful, if presented objectively isn't necessarily salty.

We built this model fast (1 week) and have since iterated this week into developing a Fine Tuned BERT model that we are training over a much broader set of toxicity, demographic, and polarity features. The training set is much larger and higher quality so we are expecting a large jump in precision upon deployment.

I hope the app gave you some good chuckles as you went around though. It's hard to explain how excited I felt when I saw pg_is_a_butt at the top of my pandas data frame the first time I processed the data.

It's doing a little bit right. :)

Re: The Saltiest Users on HN

#69

Post by Kassius509 »

msluyter wrote: post #11

I love this! I'm sort of proud to note that my salt score is -0.08, with my saltiest comment being: The iPad and iPhone are especially dangerous when it comes to accidental downvoting. Separating or enlarging the arrows would help those of us with fat fingers. However, (and this will apparently add to my salt score ;)), I'm curious how a comments like these get rated as expressing a negative sentiment?: Out of curios…

We used TextBlob for this one and penalized to help avoid false-positive classifications for saltiness (subjectivity + negativity).

Re: The Saltiest Users on HN

#70

Post by Kassius509 »

reificator wrote: post #51
weberc2 wrote: post #19

Earlier quoted context omitted.

I like the idea, but I'm very skeptical of the implementation. My saltiest comments per the tool are not remotely salty, and I'm sure I have many saltier comments. Turns out NLP is hard. FWIW, my salt score is -0.09 ;)

> I like the idea, but I'm very skeptical of the implementation. My saltiest comments per the tool are not remotely salty, and I'm sure I have many saltier comments. Turns out NLP is hard. > FWIW, my salt score is -0.09 ;) Same salt score, same opinion on the results. My top salty comment has a score of -1.00 > > > Unfortunately it's not a bottle. It's just a plastic cup with a lid and a straw. A really big plastic c…

Yeah, it really thinks code is salty for some reason....

The Muffin Tops part is making me lol.

Post reply on HN