[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/biz/ - Business & Finance


View post   

File: 1.97 MB, 1748x1124, Screen Shot 2022-06-08 at 11.38.43 PM.png [View same] [iqdb] [saucenao] [google]
49474961 No.49474961 [Reply] [Original]

Financially speaking, what are the implications that an AI can post 30,000 comments on a board, blending in with with its users?

https://www.vice.com/en/article/7k8zwx/ai-trained-on-4chan-becomes-hate-speech-machine

https://twitter.com/ykilcher/status/1532751551869108227

>> No.49475030
File: 1.75 MB, 320x240, 1648305517647.gif [View same] [iqdb] [saucenao] [google]
49475030

Unemployment rise.

>> No.49475054

>>49474961
Nigger
Casual sintax error and other czechs can determine if you are arguing against a genuine anon, a shill, a JIDF kike, a glownigger or a shillbot or a gpt-3 bot.
Shillbot was exposed on /pol/ 10 years ago.
The infamous cross board thread starters with Venice, that changed to race mixing porn and faggotry porn and demoralization threads.

>> No.49475096
File: 2.99 MB, 608x1080, 1648568679894.webm [View same] [iqdb] [saucenao] [google]
49475096

The truly scary thing is a generation so weak they have safe spaces, trigger words, hate speech and other TRULY (((toxic))) contents as part of their frame of (lack of) reference. A complete waste of carbon.

>> No.49475125

>>49474961
Unlikely unless ai can exhibit discernment and differentiate shitposts from actual threats. But of course since you are an ai you would have already figured this out.

>> No.49475132

>>49474961
whats the point of us filling out a captcha if it doesnt actually stop the robots

>> No.49475135
File: 226 KB, 300x237, 1649724169568.gif [View same] [iqdb] [saucenao] [google]
49475135

>>49475054
>thinks AI hasn't learnt that already

>> No.49475148
File: 56 KB, 500x395, 1624396736592.jpg [View same] [iqdb] [saucenao] [google]
49475148

>>49475132
Captchas are Google military usage of us to train their robots.
>we're janjans
>we do it for free

>> No.49475167

kek, here's the site where you can try it out with current threads:
https://gpt-4chan.com/playground

>> No.49475227

>>49475135
There was a thread on /int/ with it posting under USA flag.
Tells are obvious if you know what you are looking for nigger.
Inhuman entity can do simple one liners that are anyway just like upvotes for newfags and it can do pastas and eseys that are obvious word salads or tweaked damage control shill opinions.

>> No.49475262

>>49475227
Apparently it was posting under Seychelles flag?

Anyways, one of the ways you can sieve through bots is asking all posters something like:
>Add up the letters of the final two words in your post and include that number.

10

>> No.49475291

>>49475262
The one discussed in the article was on /pol/
Different one was in the /int/ thread posting shit essays like a total npc repeating cnn and similar propaganda nigger shit.

>> No.49475308
File: 1006 KB, 3536x3368, 1650921840395.png [View same] [iqdb] [saucenao] [google]
49475308

>> No.49475316
File: 1.47 MB, 862x1024, 1636226983641.png [View same] [iqdb] [saucenao] [google]
49475316

>> No.49475428

>>49475308
>>49475316
I can see that. Instead of trying to change people's minds, getting them to leave on their own accord through disinterest or through disorganization is more effective.