[ home / board list / faq / random / create / bans / search / manage / irc ] [ ]

/bmw/ - The Bureau of Memetic Warfare

He that controls the memes, controls the world

Catalog

Name
Email
Subject
Comment *
File
* = required field[▶ Show post options & limits]
Confused? See the FAQ.
Flag
Embed
(replaces files and can be used instead)
Options
dicesidesmodifier
Password (For file and post deletion.)

Allowed file types:jpg, jpeg, gif, png, webm, mp4, swf, pdf
Max filesize is 8 MB.
Max image dimensions are 10000 x 10000.
You may upload 5 per post.


File: 1458524172667.png (1.37 MB, 1024x1024, 1:1, graphh.png)

 No.5498

Hello. I've been lurking this board for a while and decided it's time to take some action. I believe we are lacking some tools to be really effective.

Our memetics are currently an art, not a science. Stuff like UI design used to be similar: an artist makes a design, changing a few things until his boss thinks it's good. Nowadays businesses constantly iterate on their websites using techniques like A/B testing to optimize some clearly defined performance indicators (e.g. the number of users who buy a product). No anecdotal evidence - only numbers.

Of course memes are a bit different. A bad meme will naturally die and a strong meme will survive. It's evolution at work! And if this is evolution, then memetics are genetic engineering. We don't wait until a good meme evolves: we artificially create it. But how can we know if the memes are any good? If it is our memes affecting the society or maybe someone else is? And where are the other memes originating?

So how do we do that? This is what this thread is for.

It's best to starting with something as simple as possible. I would like to have a way to track how fast an image is moving through the internet, something like a timeline for each image with info about the websites it appeared on. If the image got popular (e.g. appeared on 100+ sites in a day) we analyze the most common words used to see how different groups are reacting.

The biggest problem with that is scale. Crawling the whole internet is impossible. We can restrict to "only" common social media websites, online media and chans. That's still way too big without a decent budget for a proper cluster (and some pain in the ass because twitter sure doesn't want to be data mined). Perhaps we could cheat a bit and crawl only what google images or tineye gives us for a specific image.

What do you think about this idea anons? Any different things you would find useful.

 No.5501

File: 1458536065185.gif (20.77 KB, 250x180, 25:18, Gospers_glider_gun.gif)

Not gonna lie you sound like alphabet agency and/or applied memetics.

Good idea, in theory, but your proposal is… clinical. Regimented. Like you're stuck in a prescribed form of writing. That and your focus on the mainstream is odd.

Assuming good faith though. The mainstream is easily influenced, but has little influence itself. The pockets are where one must look, where the ideas may be picked up and resonate. Those cysts of inclusiveness are where ideas perpetuate and pop back into the mainstream after it forgets. We want people to be affecting for us, not just be effected.

An idea. Start using the word robust in your language, but spell it robost. It isn't a commonly used adjective, and a seemingly innocuous misspelling might be taken for the proper. Consider how though is becoming tho. It seems to me the more a word is repeated the more minds are reminded of it, and the more likely those minds are to go to it to describe similar situations.

Robost coffee, robost campaign, robost speech.


 No.5506


 No.5507

Some more research:

http://truthy.indiana.edu/

http://truthy.indiana.edu/highlights/ - topics like "Social Bot Detection", "Meme Competition & Virality"

http://www.nature.com/articles/srep00335?message-global=remove&_ga=1.156162876.647407355.1396608935 - "Competition among memes in a world with limited attention"

http://arxiv.org/abs/1603.01511 - "Hoaxy: A Platform for Tracking Online Misinformation"

https://www.aaai.org/ocs/index.php/ICWSM/ICWSM14/paper/view/8081 - "Predicting Meme Virality in Social Networks using Network and Community Structure."

This is only the stuff that is public… Obviously companies like Facebook and Google must have a big interest in this topic too. Who knows what are they working on.


 No.5528

File: 1458740797703.jpg (39.38 KB, 634x357, 634:357, effect of sunscreen.jpg)

>>5506

>Stanford is light years ahead of BMW

Tracking memes is for sterile cuckademics, that like to sit in the corner and watch.

What have Stanford meme department done, that is comparable to cuck and meme magic?


 No.5576

Obvious but, underused.

https://www.google.com/trends/

https://books.google.com/ngrams

For experimentation, an ideas is to create memes with archaic or original words attached. Periodically you'll have to record the results and the spike in their usage. See what kind of memes work and what doesn't.


 No.5605

File: 1459278796616-0.jpg (72.3 KB, 960x960, 1:1, 1458921609382.jpg)

File: 1459278796646-1.jpg (93.07 KB, 544x544, 1:1, 1458956704849.jpg)

File: 1459278796646-2.png (323.59 KB, 800x800, 1:1, 1458925724881.png)

File: 1459278796663-3.png (147.57 KB, 800x800, 1:1, 1458925317191.png)

File: 1459278796663-4.png (548.54 KB, 800x824, 100:103, real meme team.png)

>>5576

>Obvious but, underused.

Hits on google last 24 hours/week/month, is what I have used.

It is often not as important to know everything about how a meme spreads, as just getting feedback by look at how and where it is used.

If you take meme magic as an example, seeing how it is used on both pols, and influencing the direction the meme evolves there, might be the only thing that is needed, to pull it in the direction you want.


 No.5618

This is offtopic but where the hell is that one "General propaganda thread" or whatever it was called? I can't find it.

PROPER TITLES FOR YOUR THREADS THANKS

I'm not saying this to you, OP.




[Return][Go to top][Catalog][Post a Reply]
Delete Post [ ]
[]
[ home / board list / faq / random / create / bans / search / manage / irc ] [ ]