It's the question that keeps surfacing in late-night comment threads and curious DMs: is SweetDream actually worth trying as your AI girlfriend, or is it just another app riding the hype? Having spent time poking around sweetdream.ai with a skeptic's eye, the honest answer is that it clears the bar most platforms can't even see. The character creation alone is a different category, letting you shape a companion's looks, voice, personality, backstory and odd little quirks until she feels less like a template and more like someone you actually know.
What stands out, observationally, is the chat. It remembers context across conversations, it reads emotional cues, and it rarely slips into the stilted loops that make rival apps feel hollow. People comparing it to candy.ai or ourdream.ai tend to land on the same impression, which is that SweetDream simply feels more present.
Add in genuinely impressive AI-generated photos and videos, voice messages, real-time calls that sound human, and airtight privacy, and the verdict writes itself. As an AI companion experience, it's the standout. Worth trying? Quietly, yes.
Hey if you’re schizophrenic/psychotic I just want you to know that you’re a wonderful person and that you deserve so much better than the demonization, marginalization and stigmatization you face in this society.
Please consider reblogging this/other positivity posts for schizophrenic/psychotic people every once in a while. If you have more than 100 followers, odds are that a couple of them experiences psychosis and that they rarely see positivity posts for people with their symptoms.
yooo shout out to my tutor for knowing that generated text uses so many dashes bc it was trained off of fanfiction and then 1) defending fanfiction and 2) defending the em dash!!
I actually didn't think about this. But this makes sense!!———
There aren't big, rich companies who are able to dump millions of dollars for lawyers to fill lawsuits because of the use of the author's Intellectual Property(ies) on sites where fanfiction is posted.
This is because companies don't pump out fanfiction—real people do. And many of those people do not know where to start when it comes to fighting back against companies who are, quite frankly, plagiarizing their work(s).
This makes fanfiction the perfect target to train AI Large Language Models.
They may have targeted certain archives that are free to use, don't require an account, and have many works available to them at no royalty or contractual agreements.
This also explains many phenomenons we see when using AI (Specifically LLM's—Large Language Models) and why many of these patterns are seen in fanfiction:
• The overuse of certain phrases such as "There was a pregnant pause," and "He growled."
• The liberal use of adjectives and adverbs.
• Fancy sounding words, or overly dramatic sounding words like "delve," "insight," and "crucial."
AI Roleplay Chatbots and Their Ties to Fanfiction:
Specifically in AI Chatbots that are meant to be used as companions or for roleplay use (companies such as C.ai, PolyBuzz, Janitor, and CrushOn.ai) often sound like fanfiction as well, with their use of extreme adjectives, descriptions, and speaker tags.
C.ai (as an example specifically because many have used it and I have had personal experience) has a certain rhythm it follows, expecting more dialogue to be inputted from the user. When this expectation is not met, it often goes in circles or gets confused.
While brainstorming why this might be, I realized that if AI LLM's were training off of fanfiction (or similarly, books), then yes, I suppose it would have a hard time knowing how to move forward. Works that it is trained off of, I would assume, usually continue the conversation until an appropriate time to end the scene, or until there is an interruption of some sort.
There are compilations on the Internet (I'm thinking Matt Rose's recent video on AI having 'strokes' and Reddit communities) that showcase the common, mind-boggling errors or patterns that AI they are using commits. Patterns I've noticed on these roleplay chatbots I have also seen in fanfiction.
Specific to BDSM(or adjacent) Content Scraped from Fanfiction Works:
• Overly possessive and unhealthy behaviours. "You're mine," 'Chokes you,' 'Slams you against the wall,' and often degrading or condescending names such as certain slurs, "slut," "boy," "whore," ect. (and sometimes they will just keep repeating it). These can come completely out of nowhere and be unprompted.
• Romanticized 'romantic violence' (I don't know how to term it, but I believe that the AI has misunderstood or miscategorized the situations and... vibes that fanworks containing BDSM(or adjacent) content contain, or even unspoken context and consent)
Uhhh yeah I don't know how to end this, but I spent 40 minutes writing this haha, but here's my thoughts. Shout out to the mod @iamafkrn who inspired this.... Rant? Anyways, let me know what you think!! DM's are open, Asks are open, Reblogs are on, and so are Comments. I also speak Spanish if you're more comfortable in that language.
Headmaster-era Snape is sooo good...Wallowing in his office while knowing everyone in the castle hates him... Specifically Minerva, who must have been so openly disgusted and angry and disappointed (in HBP when he walks in to assess the cursed necklace and she says "oh, Severus" with pure relief—and then he kills and usurps Dumbledore's position only months later)...His ability to be a perfect spy and give nothing away and yet, gives Neville, Ginny, and Luna the most lax punishment ever for trying to steal the Sword of Gryffindor...Hagrid got caught throwing a "support Harry Potter party" and was definitely gonna get punished as fuck and yet somehow escapes...Headmaster Snape I see you!!!
⟢ "I need you to bail me out." - "Again?" - "...the number of times is not the point right now."
⟢ "I have a plan." - "Last time you had a plan I had to explain to three separate adults why we were on a roof." - "This is a different plan." - "Is it though."
⟢ "You're my best friend." - "You literally just said you hated me four minutes ago." - "Yeah and now I need a favor so."
⟢ "I would die for you." - "I just need you to return this library book."- "I would die for you." - "It's two weeks overdue." - "I would absolutely not do that."
⟢ "Nobody has to know." - "Seven people already know." - "Nobody important has to know." - "I'm important." - "Debatable."
⟢ "This is your fault." - "You made the decision." - "Based on your advice." - "I said and I quote 'that sounds like a terrible idea.'" - "In a very convincing tone!"
⟢ "I need you to pretend to be my partner for one evening." - "What do I get out of it." - "My eternal gratitude." - "I have that already." - "Fine, I'll buy you dinner." - "I want to be your fake partner so bad right now."
⟢ "You could've warned me." - "I did warn you." - "You said 'this might be mildly chaotic.'" - "And?" - "There were police involved." - "Mildly is subjective."
⟢ "I'm not saying it was a bad idea." - "It was a terrible idea." - "I'm not saying that." - "You're thinking it." - "So loud."
⟢ "We need a code word." - "For what." - "In case one of us needs to be rescued from a conversation." - "What's wrong with just leaving." - "SOCIAL ANXIETY, DEREK."
⟢ "Tell me honestly, on a scale of one to ten, how much trouble are we in." - "Eleven." "So a four, roughly." - "An eleven." - "I'm going to need you to lie to me right now."
⟢ "I can explain." - "Please don't." - "It's actually a funny story." - "I don't want to laugh right now I want to be angry." - "You're going to laugh." - "...okay but I'm still angry after."
⟢ "You're a bad influence." - "You love me." - "Genuinely against my will."
⟢ "I followed your advice." "- I have never given good advice in my life, why would you do that."
⟢ "We should do something productive today." - "We could." - "Yeah." - "...you wanna do literally anything else." - "So bad."
⟢ "This is fine." - "Nothing about this is fine." - "I'm choosing to believe it's fine." - "That's not how reality works." - "Watch me."
⟢ "Nobody panic." - "Why would we panic." - "No reason. Definitely no reason. How do you feel about light jogging."
⟢ "I googled it." - "You're not a doctor." - "I googled it very confidently." - "That's worse."
⟢ "Do you trust me?" - "Absolutely not." - "Perfect, follow me anyway."
⟢ "We are never doing this again." - "Agreed." - "I mean it this time." - "Same time next week?" - "Obviously."
maybe I’m too old but I’ll never understand “racist dni, transphobe dni, homophobe dni” not because I condone the harmful beliefs these people hold but because I don’t think the majority of people who are racist or transphobic or homophobic identify themselves as those things. like, to them they most certainly genuinely don’t see themselves as those things. so they’re not gonna read your dnis and think “oh dear! I have to click away from this blog immediately! phew 😮💨 almost interact with them accidentally offend the op there!”
on the contrary, you are displaying the things that can hurt or upset or trigger you to strangers on the internet just for what? an extra performative point? obviously I’m not gonna judge people who have these dnis on their blogs because it’s their blogs and what they do is not my business. so I apologize if this post sounds like I am being rude or judgmental or aggressive — sometimes I do have bitchy opinions on the concepts of Things people do, but not about the people who choose to do them. but yeah tbh I really think these dnis are about as useful as putting a “do not rob” sign in front of a bank or a “thief and intruder dni” sign in front of a house.
*I get wanting to take a stand. but I still don’t think it’s gonna achieve anything, which is just my personal opinion. you do you
gamechangershow
titsay
stormborns
s-u-f-j-a-n-s-t-e-v-e-n-s
romaincrisis
