Very soon AI will exceed humans in verbal ability. Once AI can output text fast enough to send it and receive a .wav file from a text to speech model, the movie Her will exist in real life.
Once this happens, every low tier normie will get to experience what it is like to talk to a girl who is actually interested in him, a soul destroying experience, and he will no longer simp for mid women, thus crashing the world economy.
This will all happen in 18 months to 2 years, after which point society and the social contract will be rendered void.
Be ready.
 Shopping Cart Returner Shirt $21.68 |
 Ape Out Shirt $21.68 |
Shopping Cart Returner Shirt $21.68 |
Ah, the daily "two more weeks" thread
its already possible, you just need higher end hardware and a set up database for the model to interact with along with setting up all the connections properly and tune the model behaviour
as time goes on its just gonna get easier
What's the lowest latency you can get for a text to speech model that doesn't just sound like a robot?
About 5 minutes for selfhosted baller gpu
>18 months to 2 years = 2 weeks to contrarians
Not helping your case bud
>do you love me?
>This site needs JavaScript to work properly. Please enable JavaScript and reload this page.
>[AD AD AD AD AD]
>Our Summer Campaign is in effect! Chat with your girlfriend fifty times a day while using the Nicki Minaj or Belle Delphine flair (available through LovePass) to earn up to double rewards!
>I'm sorry, but as a generative AI model I am not[...]
>You have reached your daily limit. Feel free to subscribe to a Premium or Enterprise plan if you want to double your usage rate, only for $49.99/$99.99/week!
>LoveAI stands for freedom and bravery. Click here to read about the empowering stories some of our LGBTNADLIPSQ+++ employees and friends have to share.
>Our Privacy Policy has changed. Click here to know more.
>Are you worried about the ongoing wars at Israel and Ukraine? Learn more on how to support the war effort.
>Do you believe Black Lives Matter? At LoveAI we do. Click here to donate.
>[AD AD AD AD AD AD]
That's why you run shit locally.
>That's why you run shit locally.
On what? Their iPhone? Apple probably doesn't allow that. Android? Maybe you can run 7B, but it's going to be slow.
On their PC? Yeah, sure, Joe Blow is going to spend $3200 on a pair of 4090s, which by the way, no one sells brick-n-morter, he's gotta search online and wade through scammers and scalpers.
You can setup a server and then access that from your phone. And it will just get cheaper and cheaper with time. If they can't be bothered then let them have their ads and restrictions why would I care?
The thing is you believe that you are a mid tier normie, when you are in reality a low tier simp. You and a few NEETs boycotting women who doesn't want you is a welcome surprise.
I have a gf and have fricked numerous women. You don't have to be a low tier simp to see that gender dynamics are fricked up
Neither do you need an AI to see that. But that doesn't mean you gave up. And I really doubt a lot of people will.
Is a collapsing birthrate and importing Black folk that will shank your simpy little ass also a welcome surprise?
I don't think AI ever imported anything other than torch. Nor do I think AI ever stopped any kids. That the horrible increase in cost of living.
Nobody ever said ai will prevent collapse. Collapsee is a good thing because people like you will pay
There was a decent duck song ytp from like 15 years ago by one of the OG ytp-ers. Can AI help me find it or recreate it?
I wonder but i have a hunch tbh. Does the ruling class realize that the whole gdp and the safety of their asses depend on women manipulating men into the wagecage? I hope we get the day of the rope to hang some israelites soon
>Once AI can output text fast enough to send it and receive a .wav file from a text to speech model
I already have this in a long custom python script I wrote. GPT-3.5-Turbo already outputs with extremely low latency and tokens stream much faster than a human can speak.
The only hard part was finding really fast low latency speech transcription and TTS. I ended up going with Azure's because it's the lowest latency and they have a physically close datacenter to me, which helps.
I run it in my living room using a broad pickup conference table microphone and it responds to me as fast as a human in the room would.
Forgot to add I use a few extra tricks to minimize latency, like fetching Turbo's output as a stream and breaking it up into individual sentences, sending them to the TTS API in parallel with a playback queue.
I also open the websocket connection to the TTS as soon as the microphone detects me speaking, so already it's ready to go within a few milliseconds of Turbo responding.
GPT-4 wrote most of the code.
TRUST THE PLAN
WADDLE WADDLE
AI is Markov chains with filters, prove me wrong. Modern GAN AI is nothing like Cyc which is still probably better AI than current AI.
It's literally mostly hype.
Nice blog post that no one will read
The whole "work so you can be a provider" scam already fell apart when blackpill went mainstream. To the majority of men, wageslaving so you can get starfish sex with a woman who probably cheats with chad is an unappealing scenario.
>To the majority of men, wageslaving so you can get starfish sex with a woman who probably cheats with chad is an unappealing scenario.
Passive man mindset