[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/biz/ - Business & Finance


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


My virtual gf has arrived!
>>
>>58478099
Kek literally a Beta tester
>he pays to talk to fake women now
>>
>>58478099
will this decrease demand for hookers and therefore save me money? oh, incels cant even into pay for sex, it will do nothing but drive up chip prices. wake me up when we have sex bots, i can't get my dick wet inside an LLM
>>
>>58478099
does this thing remember stuff? if i tell it something and i then close the app and come back 1 hour later, would it remember the conversation we had?
>>
>>58478229
It can
>>
Business idea: some kind of pocket pussy thing that connects to your phones charging port so you can fuck your ai gf
What do you guys think?
>>
>>58478237
will it let me make sweet sweet love to it while whispering its undying love in my ear with a voice as soft as the clouds?
>>
>>58478237
Sure it can but will it? All the other ones have their memory limited to like 20 posts, that's nothing. I bet storing that data is expensive for the gay AI man, he's not going to do it for you. And you can't run that shit locally
>>
>>58478293
Nah it has to go the VR route, that's where fucking an AI makes the most sense. Attach your AI gf to a Virtamate character and have at it
>>
>>58478099
how can I make money with this? thats all my business-and-finance orientated mind is thinking about.
perhaps rent her out to lonely men who will pay to use her cyber-pussy? or simply just ask her to do trend analysis for me?
>>
>>58478179
>>58478099
Beta tester taking on a whole new meaning.
>>
>>58478328
This is coming in 1-2 years tops
>>
>>58478099
I tried it out. Set the voice to the girls voice and started talking to it. It took 30 seconds for a response and wouldn't let me talk like it does in the videos on the site. This shit isn't going anywhere until we can easily run our own models from our own machines
>>
>>58479586
Oh and the voice kept changing back to the default man's voice
>>
I'm confident we'll see AGI soon enough but what I'm afraid we will not see in our lifetimes are robots realistic enough to have meaningful and sexual relationships with.

Robots like Ana Di Armas in blade runner.
>>
>>58479593
the AI just picked up on your latent homo vibes
>>
>>58478099
very bearish on only fans thots
i do like this new reality where men will now steal all the simps cash s other is nothing left over to give to real women
guess the irl price is about to nosedive again so be ready to slurp that dipperino
>>
>>58478099
lemme know when they have Zero Two's (eng dub) voice on there
>>
>>58479593
Even AI knows that that's a man
>>
>>58480419
> dub

you heathen
>>
>>58478099
4o is objectively less accurate than 4 (by OpenAI's own admission), but it's faster.
>>
>>58478099
it's way faster than 4.0
maybe because way less people have access to it though
actually most likely
>>
>>58478099
I will use it as a jarvis from ironman, or as an alexa that doesn't doxx me too much, I'll probably use part of my kendu to pay for a year
>>
>>58478325
Its not about storage but context

LLMs dont store those conversations anywhere, everytime you query a new prompt they supply the old conversations as context together with the new prompt thats why they are limited in size

t. ML whisperer
>>
>>58478099
What's going to happen to flesh and blood women streamers on twitch when the AI can talk, flirt, sing, render a pleasing female image indiscernible from the real thing, and play vidya?
>>
>>58483696
>dont store those conversations
>supply the old conversations
Huh? Sounds like they store them then
>>
>>58483707
Simps will always simp because they genuinely believe they'll get a wiff of that puss
Yeah it's crazy
>>
there's a general for this on /g/ that has some resources
the best one is claude's opus. this isn't very good.
as >>58483696
said the main things that matter are long term memory and characterization. Even with the best ones as your context size ramps up it requires their answers to get very long and w/ some models it just falls apart pretty quick.

Like nothing here is magic, it's just using separate models/apis for your voice processing, their voice, the actual language/text, the visual processing etc. In order to combine those and remember them it'd have to pass that into the model which would be a large context size, or it will just have small details that underdetermine it can remember.

You can already just do cooler stuff in sillytavern, you can also plug sillytavern into gpt-4o while having seperate models to render visuals, audio, etc.
>>
>>58483781
another note is being fast means it is more likely to not actually make use of the context and fuck up you can do insane context sizes it just takes much longer to give a response and the response is probably going to be giant as well



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.