Thread 105963972

My 4chan posts don't get enough replies. How do I make an ML model to predict how many replies I will get for a given post? I think a standard model won't work well because reply rate depends on the cultural context at the time.

>>105963972 (OP)
You fucking retard, you either rage bait or you just post and don't give a shit if it doesn't get "traction" this site is ephemeral, nothing matters

>>105964081
the archives counter your point
the archives are the reason why we have so many schizos

Does Jinx top or Silco?

saar listen to me
pls kindly listen to me
you must make good content

>>105964081
>rage bait
go back zoom zoom!

>>105964081
I WANT (YOU)S

>>105964087
Jinx tard-tops, fucks while drawing pink monkeys and letting off fireworks. Until Silco gets pissed and tard wrangles her, then she becomes Powder and submissive

>>105964172
Jinx only becomes Powder when Vi is around, Silco has her cured before Vi shows up to ruin everything (again).

>>105964200
Silco mindbroke powder into jinx. He can turn him back.

>>105964224
Powder mindbroke herself by murdering her entire family. Silco put her back together by helping her separate herself from her past actions. All she needed was time to heal without her deadbeat excon sister showing up to fuck everything up.

>>105964279
Silco go back to bed

>>105964553
All you have to do is watch the show without pulling out your phone to realize that Silco unironically did nothing wrong.

>>105964679
If he didn't make Jinx, he didn't do anything right.

>>105964081
>rage bait
kill yourself zoomard

>>105964679
>Tries to have Vi killed
>Never tells this to jinx

>>105963972 (OP)
sex with jinx

>>105964122
Nigger I've been here for 12 years

>>105967185
>GamerGate tourist
kys

>>105964081
>>105964122
>>105965161
Hmm, curious. Looks like he baited you instead

>>105963972 (OP)
you need to be able to turn words into tokens
so you could start with straight up a dictionary
have a layer that accounts for typos, you got a couple options here
idk, tree list of acceptable spellings + distance factor bw words
theres a couple ways you could do that too

and then once the words are numbers
you reformulate the problem as "which numbers elicit what other numbers in response"
you throw a whole lot of maths at the problem and see what sticks

you need to define success to see what correlates with it
so you might have to label a whole lot of data
which isnt a problem bc you can define what a successful thread is using the reply count and automate the whole process

>>105963972 (OP)
the thing is most of 4chan is just artificial discussions made by a bot, the responses and ops by the same bot

ofc they will ignore original posts

>>105965815
He was in an untenable position. What would you do if your mentally ill daughter's deadbeat fuckup of a sister magically returned from the dead one day to destroy your daughter's fragile but slowly recovering psyche?

>>105967214
fr fr unc he deadass ragebaited bro ong :skull: :rose:

>>105963972 (OP)
Is this zoomer coombrain shit of a show worth it?

passionate sex with jinx

>>105963972 (OP)
scrape the archives, use some image embeddings model to produce a vector out of the OP image, perform linear regression on your vectors.
it will be a shit model but better than chance.
if you can scrape thousands of OPs you can train a neural model and get better results.
good luck anon!

>>105963972 (OP)
Go on /r9k/, mention moids
Go on /pol/ and make a thread praising jews, indians or other thirdworlders, you can also just make a bait thread with indian words inserted in it
Make a wayland vs xlibre thread on /g/
Post a xigger screenshot with controversial opinion
that's it nigger