Neural Network based on 'The Lounge'

I see nothing wrong here.

What happens in The Lounge stays in The Lounge. Honestly analyzing or doing anything else with it is even more horrible than the fact people actually participate in it. It needs to burn. It needs to be nuked. Please do not subject yourself OP to anything more than glimpsing at The Lounge once in awhile. My God …

https://forum.level1techs.com/raw/128802/20

1 Like

Something like this?

Just out of curiosity, which language do you plan on wrangling the data/making the neural net in?

Huh, totally not what you were getting at, but does anyone know of the military’s OpFor?
How about an intentionally bad AI, sandboxed, so we find out the worst that can happen, and regulate public AI systems?
Of course there would never be a risk of it escaping…

I only think of this because The Lounge is a place I fear to tread… (and/or it moves way too fast for me to keep up)

I can only assume if the lounge becomes sentient, it would want to gas everything questionable.

I’ll allow it.

2 Likes

It’d talk about politics and guns before US bed time while also dialing to extreme toxic shit, then it’d go briefly quiet, which is followed by cars once US is waking up

That said lounge character needs 24 hour behavior clock

Let’s call it @Disconaut

2 Likes

And it would reply to itself and use shoulder shrug emojis :man_shrugging:

2 Likes

simplest would probably be python with TF, Chainer, or PyTorch

Variety of NN implementations available and you have scrapy and other tools for data acquisition/sorting baked in if you want them

IIRC though most pre-written speech chain nets are for torch7 in lua

Dumbass chiming in. The fuck is a neural network and why would you want the lounge to be one?

The only good reason I could come up with is a shower thought I had:
Maybe the neural network (type of machine learning) could be used to replace what an absent lounger would like or react with. Alternativley, we could feed post count data to predict the post count of the lounge per month.

Or go sinister with it and predict “evil posts” and hide them before they are postedimage

I plan on using the NLP functions of the fast.ai library which sits on top of the Pytorch library. So Python.

AI is not even close to sentient they are just big math functions that calculate probability using matrices and calculus. Have a look at this video for more information.

I’m just going to see what I can do with fast.ai’s library first since that’s the one I know the best.

See the video above. It’s not very complex, all it will be doing is looking at a string and and generating a algorithm to predict the outcome of the next character appearing,

For example say the first character that’s randomly generated is a 'i'
This could be the start of the work 'it's' so it might finish printing the word 'it's'

Since it’s impossible to tell whether a post is sinister due to the lack of a dislike button, we can only tell if a post is popular or not, with varying degrees depending on how many likes it receives.

That did not keep other sites from shadowbanning users in the past.

Human’s can yes, but it’s impossible to give the AI context, since it’s just calculating probability.