[nostrademons]

I've heard that rather than deleting your FB account, it's better to turn it into a piece of obviously-fictitious performance art. Don't try to remove your data from the Internet (which is futile; you know Facebook still has it on some archive somewhere anyway). Instead, replace it with nonsense. Set your interests to your competitors so that you get notifications everytime they release a new feature, and full information about their advertising campaigns. De-friend your actual friends, and friend random celebrities. Make your political views insane. Post status updates generated by Markov-chain or recursive neural net; if you're feeling particularly bold, you could try to reverse-engineer the weights on Facebook's classifiers by generating data and seeing what sort of ads you get.

Chances are, doing this will get you kicked off Facebook and deleted pretty quickly. And if it doesn't, you've just fucked with not just Facebook, but all their advertisers. GIGO.

[thr0waway1239]

Here is an idea. Get a bunch of cute pictures of cats or babies from the internet. Add text with anti-FB messages and a link to these HN threads. Then write a script to upload the photos to your account, preferably on an endless loop.

If FB does not block/delete your account, it will be non-stop hilarity for your friends, who will probably share the pics too and maybe some of them will actually pause and think. If they do delete your account, you are now mostly free from their clutches without unnecessarily defacing your own profile.

Someone please write this script and put it on GitHub. :-)

[Grangar]

I have a dormant facebook account that I can't be arsed to delete, I need this now.

[taejo]

> which is futile; you know Facebook still has it on some archive somewhere anyway

If Facebook does this with the data of an EU resident, they are breaking the law. The Irish Data Protection Commissioner has audited Facebook and found that their deletion of user data was in compliance.

[mst]

Sounds like you could achieve this easily by feeding your facebook account's updates from a markov chain bot set to train on the @realDonaldTrump twitter feed.

[steve19]

Then you get fired/don't-get-that job when someone takes offense at something "you" have written. Going random is risky.

See Microsoft's Tay: http://gizmodo.com/here-are-the-microsoft-twitter-bot-s-craz...

[mst]

I was expanding on how to achieve "Make your political views insane. Post status updates generated by Markov-chain or recursive neural net". Did you mean to reply to the grandparent rather than me?

[drusepth]

You can use this API for that: http://www.retort.us/markov/create?identifier=realDonaldTrum... :)

[voidz]

Why not just post a nipple?

[andai]

Simple and beautiful.

[CamperBob2]

I've thought about posting some really over-the-top shock/porn imagery to my Facebook wall, basically making it look like it was hacked. Maybe then they'll cancel my account, like I've asked them to do several times.

[nostrademons]

Random evil thought: could you sell your Facebook account to hackers? You've built up a multi-year reputation as a "known good" account; that'd be incredibly valuable to someone trying to generate cross-correlations between your interests and their product. After all, the whole reason Facebook is worth $300B is because they have all this data on their users; since it's your data to begin with, shouldn't you be able monetize it just the way Facebook does?

[kristianc]

I'll start the bidding at 33 cents.

[rm_-rf_slash]

It could sour your relationships with actual friends. But if you're ok with them being deceptively advertised to, maybe they weren't such close friends to begin with...

[dkersten]

Unfriend the ones you care about a few weeks beforehand.

[taneq]

I don't think "hackers" in this context is the right term for people who game the advertising/metadata industry. Maybe "independent advertisers"?

[harry8]

More likely the hacker would rather use it to steal your identity and f. you over very, very hard. Much more lucrative than spoofing with a "known-good" account, right? So you'd really, really better trust those hackers you're selling to.

[harto]

It seems that would be confusing (at best) for your contacts that remain on Facebook.

[nostrademons]

Presumably you would tell them via backchannels (or even Facebook) that you were doing it.

[cwkoss]

Someone needs to make a good market for this.

[Natanael_L]

Already exists for Twitter and reddit accounts. I'm sure spammers will pay for good fb accounts too.

[rhizome]

I would lean more towards setting up a script that will post whatever the 400th most popular story on Google News, thus destroying your demographic membership. Turn yourself into noise.

[behnamoh]

Sorry, I'm new to Markov Chains, but recently I've been thinking about a way to "beat" ML algorithms of GOOG, FB, etc. So that the information stored about me is NOT true. For example, if someone keeps searching for "LGBT" while he/she is not one, GOOG will assume he IS one, or at least, interested in the topic. My question is, is there a general method for beating ML/AI algorithms and stay hidden from them? Thank you.

[thr0waway1239]

If you watch the following video from Google (for about 5 minutes from the marked location), it looks like these algorithms have a pretty big blind spot: the unknown.

https://www.youtube.com/watch?v=sphFCJE1HkI#t=7m50s

So you can use Google itself to find words which are obviously not well known (just construct a dictionary filled with random characters and see how many results come back for each word). Take a passage of clean text, and randomly replace the nouns in the text with these garbage words (capitalize the first character), and post it to your social network. The surrounding text will provide legitimacy, but the garbage words will probably throw off the algorithms. Do it often enough, and someone at Google will probably investigate the issue :-)

Or maybe go and get public domain text which has a lot of words which are not in use today and fill up your FB/Google/Twitter feeds with such data. My view is that such data will very likely throw off the existing algorithms, and if people do it at sufficient scale, we may discover that ML algorithms are only as smart as the training data set.

[nostrademons]

Markov chains aren't really relevant here - they're a way to generate random but "seemingly plausible" data - but yes, there is a way to fool the type of convolutional neural networks used in image recognition. Paper here:

http://www.evolvingai.org/fooling

[jasonkostempski]

I thought about trying something like that to get my HN account deleted.