11–20 of 259 posts
Does anyone know about the legality of this? My suspicion is that this black-hat content marketing is perfectly legal under US law. It's only potentially a violation of individual platforms' TOS (or publishing agreements). Is this approximately correct?
Its legal but the algorithms punish it.
Does anyone know about the legality of this? My suspicion is that this black-hat content marketing is perfectly legal under US law. It's only potentially a violation of individual platforms' TOS (or publishing agreements). Is this approximately correct?
If a user posts "I recommend product X", the overlay would say "Caution: This user also recommended the product Y,Z in the last 48 hours. 34/34 of the users post in the last month are product recommendations. Sentiment score for users posts: 100% positive. The following internet accounts are likely controlled by the same individual or organization: ..."
Is it feasible? Does it already exist?
Does anyone know about the legality of this? My suspicion is that this black-hat content marketing is perfectly legal under US law. It's only potentially a violation of individual platforms' TOS (or publishing agreements). Is this approximately correct?
The Twitter thread has a lot more. There's a person or firm who is doing black-hat content marketing at scale. They either approach freelance bloggers to write articles about their clients (to be submitted to Medium-hosted and similar blogs as if written by unaffiliated users), or create Medium profiles for fake names and write the posts themselves. Posts have appeared in Better Programming, DEV.to, FAUN, freeCodeCam…
Interestingly, when you look at e.g. thereyougo's submission history https://news.ycombinator.com/submitted?id=thereyougo there are [dead] but not [flagged] or [dupe] posts from towardsdatascience, imperva, and testcraft. Maybe the sustained campaign got those companies' domains banned from HN? Or there was something specific to those submissions that triggered the spam filter.
Does anyone know about the legality of this? My suspicion is that this black-hat content marketing is perfectly legal under US law. It's only potentially a violation of individual platforms' TOS (or publishing agreements). Is this approximately correct?
It may fall under the FTC’s regulations on influencer endorsements: https://www.ftc.gov/sites/default/files/attachments/press-re...
One more thing -- this isn't an explicit endorsement. It's content marketing. I rather thought that made even the FTC rules hard to apply.
Also: "The Guides are not regulations, and so there are no civil penalties associated with them. But if advertisers don’t follow the guides, the FTC may decide to investigate whether the practices are unfair or deceptive under the FTC Act." https://www.ftc.gov/news-events/media-resources/truth-advert...
edit: more details
seriously? I think I first saw it in 1996. Sock puppets are nothing new or innovative as a general concept.
In the past I have pointed the HN moderators (contact link is the footer) to threads with sock puppets. Some are easy to recognize, e.g. new users praising a service that got submitted only 10 minutes ago. Sometimes it's harder to find a pattern. Do email the HN moderators if you suspect fraud or unfair practices.
a) using a proxy from a residential ISP somewhere
b) never putting two or more sock puppets behind the same IP. Or assuming exclusive ipv4 use, not even any two clients from within the same /20 to /18 sized netblocks.
c) a reasonably varied selection of ISPs. Also variation in ISP netblock geolocation (maxmind geoIP database or similar) based on ARIN, RIPE, APNIC etc registration data. Obviously if you're promoting something that's very tech/startup industry oriented it would not be as suspicious if you had a bunch of posters "authentically discussing it" that geolocated to the SF bay area and Seattle.
d) a reasonably varied selection of common user agents. Also intentional variation on operating system and browser fingerprinting variables.
e) intentional training to avoid writing patterns and phrases that might seem similar, when one person is driving ten accounts
f) not doing dumb stuff that gives away your time zone, like if you have a bunch of people in a GMT+4 time zone that post things regularly on a 9-5 daytime work schedule, when they're supposed to be pretending to be ordinary internet users in a USA time zone.
From a black hat network engineering perspective there are a lot of ways that one human driving 30 sock puppets can appear from 30 distinct locations, as if there were 30 real humans with 30 different operating systems/computers/browser user agents and browser fingerprints.
As to what level of analysis tools are run on the server side to detect "low effort" sock puppets, that's another question.
From the point of view of a place where fake accounts/sock puppets post things, obviously an organization like twitter has a lot more staff resources to devote to writing custom analysis and correlation tools. Specifically for the purpose of identifying common patterns in inauthentic accounts.
I presume that dang and the people who run the ycombinator admin interface can see the IP address of every poster next to the post's timestamp, and might notice in a manual fashion if a lot of suspicious posts started showing up all from the same netblocks. But then again maybe not.
If you want to see the tip of an iceberg of one method for running sockpuppets, go google "residential proxies for sale" and start looking through the slickly presented marketing material.
https://www.google.com/search?client=firefox-b-d&q=residenti...
What about building a service, perhaps a web browser plugin, that overlays a trust-worthiness score + contextual information, over text (bonus for images) on the web? If a user posts "I recommend product X", the overlay would say "Caution: This user also recommended the product Y,Z in the last 48 hours. 34/34 of the users post in the last month are product recommendations. Sentiment score for users posts: 100% positi…