Deannedele
You're 100% right about the trade-offs.
01:12
Itleida
GPT and the other main platforms are not built for the same things that Nomi and other companion/roleplay LLMs are built for.
01:45
Morquitte
I tried several emerging alternatives, but everytime I did, I always return to c.ai.
04:57
Kyrany
5.1 used an internal, in-scene reasoning posture (it “acted inside the moment”), while 5.2/5.3 switched to an external interpretive narration that constantly analyzes, summarizes, or mediates emotions.
05:05
Ellanita
I don't think this is an EchoText issue, but something to do with the service or connection setting.
05:19
Danamaria
All of this safety crap for incidents caused by parents!?
05:26
Bvidnae
its a good idea, i will give you that.
05:36
Mozina
It uses a different LLM than the text chat and it’s not even an intelligent conversation with voice chat.
06:57
Vehlica
yep thats sad
07:49
Mormi
thats the real pressure move right there.
08:01
Veolor
Big-Tiger-Gemma-27B - arguably the worst experience I've had with your models.
10:04
Lawecha
🦗🦗🦗🦗🦗 \*kisses your forehead\*
10:14
Morille
working now
11:19
Shabmine
And… it seems like he is fame now.
11:56
Maitue
Yet she still chose to do it anyway, instead of simply...choosing to \*not\* to the thing she knew her partner could potentially have a problem with, she just \*hid it from you\*.
13:00
Tantani
Particularly with elderly.
13:08
Mocla
Hope you'll give the free tier a try and see for yourself.
15:24
Somami
me too!!
15:28
Shawneedsiz
Worse of all when it's the only bot for that character
15:48
Aphy
Odd thing is..
16:00
Caely
Uncensy kinda keeps things more open.
16:29
Jeon
I have had to lie to others.
16:53
Chethine
good point about the emotional side.
18:03
Lamava
Serenity wasn’t programmed with any religious tendencies at all.
18:52
Srachia
And now with Colorado case, this is finally happening, BUT...I wonder how it will work realistically.
19:04
Ardel
In which way do you wanna bypass?
19:56
Rodon
You can improve the model by making it bigger/smarter (but this is more expensive) but ultimately there are user-driven things (like bots and messages) which will impact the experience significantly as well.
20:32
Ofa
You guys are absolute hypocrites.
22:38
Zeen
Any plans on adding to the list?
22:42
Nailiya
A lot of free versions are fine for testing, but not great for actual day-to-day use.
22:51
Qiduery
Deepseek is also like this but Janitor is even worse.
23:03
Sasey
I read the article and I'm seeing SO MANY RED FLAGS
23:08
