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

Show HN: Using stylometry to find HN users with alternate accounts

stylometry.net

Post reply on HN
Print view View original

401–410 of 562 posts

Re: Show HN: Using stylometry to find HN users with alternate accounts

#401

Post by gilleain »

ricardobayes wrote: post #363
thesz wrote: post #198

Earlier quoted context omitted.

I keep no alternate accounts, but this tool reports best matches for me that appear to be Slavic or just Russian - and I am Russian. Best match score in my list is just above 0.5. There are some clearly alternate accounts on the list, their match scores with this tool are well above 0.7. It is probable that persons of same cultural origin will have similar writing style and vocabulary. It is also probable that person…

My highest was 0.41 and the person writes nothing like me. I guess I'm a unique snowflake after all.

my second highest hit (ie, third in the list) is gwern at 0.45 who i'm fairly sure is not me.

Re: Show HN: Using stylometry to find HN users with alternate accounts

#403

Post by meta2023 »

lzooz wrote: post #374
wizzwizz4 wrote: post #347

Earlier quoted context omitted.

Why use John? Just run down the list of Hacker News usernames; it'll take less time. (Or, better still, don't; just because the privacy's theoretically compromised doesn't mean we have to exploit that.)

I don't think there's a public list of all HN usernames is there? Found this, it includes 250k usernames, but it's not there. https://www.kaggle.com/datasets/hacker-news/hacker-news-corp...

The username in question isn't in this dataset but maybe it was created in the past 10 days, as the max(timestamp) is Nov 16th, 2022.

https://console.cloud.google.com/marketplace/details/y-combi...

Re: Show HN: Using stylometry to find HN users with alternate accounts

#404

Post by scarmig »

gilleain wrote: post #401
ricardobayes wrote: post #363

Earlier quoted context omitted.

My highest was 0.41 and the person writes nothing like me. I guess I'm a unique snowflake after all.

my second highest hit (ie, third in the list) is gwern at 0.45 who i'm fairly sure is not me.

I was actually just looking at near hits for gwern and found what's almost definitely a defunct alt for him.

Re: Show HN: Using stylometry to find HN users with alternate accounts

#405

Post by bee_rider »

Aachen wrote: post #348
musicale wrote: post #141

> I made this site mostly to show how easy this is and how it can erode online privacy looks like it can indeed > Here are some frequent HN commenters: (EDIT: Removed due to privacy concerns) How surprising that someone might object to being included in a demonstration of the erosion of privacy! Is the site opt-in or opt-out?

I doubt they asked 78k users for permission when there's no standardized way of reaching out if you're not a site admin. It's opt out if anything.

You opt into making your writing publicly available when making posts on this site. I’m not sure what Ycombinator’s user agreement* says about this, but it is pretty obvious that they haven’t done anything to prevent it (and it isn’t clear what they could do).

* and I mean they author of the tool is here making posts, so I guess they have agreed to the TOS, but clearly someone who hasn’t agreed to it could also make this tool and scrape out publicly available posts without agreeing to anything.

Re: Show HN: Using stylometry to find HN users with alternate accounts

#407

Post by marbu »

kaba0 wrote: post #294
tbrownaw wrote: post #159

Earlier quoted context omitted.

> but this functionality turned out way creepier than I thought the moment I tried it Hopefully this raised awareness means that people who actually need anonymity will be more likely to know to take precautions.

Genuinely asking, what way is there to combat this? Is there a tool that takes out stylistic elements of your comment?

One way would be to run such tool before posting and then based on the results, tweak the post and repeat until the similarities are not statistically significant. Or instead of tweaking, start posting under a new throwaway account. But this won't save you when some new way to analyze style appears in the future. Moreover there are other types of meta data which can be taken into account to narrow down the search space a bit such as timestamps. And obviously more you write, harder it is to control these things.

Re: Show HN: Using stylometry to find HN users with alternate accounts

#408

Post by jacooper »

Yeahsureok wrote: post #15

On the how to avoid section: Isn't running comments through a randomised translator a few times then back considered a countermeasure also? Also think it's probably poor form to list users as examples without their permission.

> However, using automated methods like machine translation services do not appear to be a viable method of circumvention.

https://www.whonix.org/wiki/Stylometry

Re: Show HN: Using stylometry to find HN users with alternate accounts

#409

Post by jrumbut »

ricardobayes wrote: post #363
thesz wrote: post #198

Earlier quoted context omitted.

I keep no alternate accounts, but this tool reports best matches for me that appear to be Slavic or just Russian - and I am Russian. Best match score in my list is just above 0.5. There are some clearly alternate accounts on the list, their match scores with this tool are well above 0.7. It is probable that persons of same cultural origin will have similar writing style and vocabulary. It is also probable that person…

My highest was 0.41 and the person writes nothing like me. I guess I'm a unique snowflake after all.

I have a few in the low 0.5's and, honestly, they seem cool and I want to meet them.

Re: Show HN: Using stylometry to find HN users with alternate accounts

#410

Post by sitkack »

JKCalhoun wrote: post #310

The asymmetry is interesting. I have no alts but of course it nonetheless reported accounts similar to mine. Running then the most similar person to my account did not put me in their top 20.

I believe this is the https://en.wikipedia.org/wiki/Friendship_paradox
Post reply on HN