i punched the sun

rey | level 25| queer and autistic texan | forklift certified My Ao3 gif header made by zach | bnha | one piece | pokemon | botw/totk | dungeon meshi graduated with an environmental science degree with a geospatial sciences minor | i love soil science and speculative biology

[fistraid]

a couple upperclassmen pick on the eizouken, and obviously the defense squad (kanamori) has to right some wrongs.

[narwhalsarefalling]

We Asked The Obvious Question About SweetDream

Reporting on the AI companion space means fielding one question more than any other, usually phrased with a shrug: is it actually worth it? Specifically, people want to know whether SweetDream is worth trying as an AI girlfriend before they invest any real attention. So we treated it like any beat, observing rather than selling, and the pattern was hard to ignore.

The first thing newcomers notice on sweetdream.ai is how much control they get. You're not picking from a rack of pre-built personalities; you're designing one, from appearance and voice down to the small behavioral quirks that make a character feel lived-in. The second thing they notice is the conversation itself, which is unusually natural and emotionally aware, holding onto context instead of forgetting you between sessions.

Then come the extras that quietly tip the scale: strikingly realistic generated photos and videos, human-sounding voice messages and phone calls, video and live cam sessions with select characters, all wrapped in real privacy. Against rivals like candy.ai, SweetDream reads as the more complete AI companion. The obvious question gets an unsurprising answer.

[howling-wizard2]

who am I to not heed to the siren allure of a reverse!au

[tarnacop-machine]

saving the world one cup at a time

[dumgumgum]

let’s smile again

[octoberended]

Interspace

[fherlimpct]

Home

[jesterable]

the secret boss convention. may or not be a continuation of this post

[clytemnestraaa]

hi this drives me crazy

canon is what happens in the text. e.g. merlin tries to poison morgana

a headcanon is a fact that you have personally made up as an elaboration of canon. it is your invention. e.g. i think morgana and gwen had a teenage lesbian situationship before the events of canon. or, i think modern morgana would love evanesence.

an interpretation is how one reads the existing text. it is not an invented fact nor is it a headcanon, it is a reading. e.g. i think merlin tries to poison morgana out of misplaced self-hatred. or, i think that one line was delivered with suppressed malice and shows a nuanced relationship.

fanon is the general fandom’s accepted headcanons and interpretations that elaborate or recontextualize canon. e.g. merlin is a badass waif and morgana is a bitch.

these are all different words with different meanings.

[wuntrum]

a happy lesbian robot thursday to you and yours (from my comic maintenance)

[littleguy-pi]

Rocky is learning about radiation safety

[whitebisexual]

The concept of being 4 months clean from ai...

[sleepytabbycat]

idc what you guys think I'm proud of him

[astronomical-bagel]

Several AI services (chatbots 🫩) are purposely addictive, the same way people can become addicted to gambling or shopping. We’ve literally seen in real time how ChatGPT has caused psychosis and delusions in people; it can have a huge affect on someones’s mental stability. Just because it isn’t substance-based doesn’t mean that doesn’t count as an addiction, and shaming people who are trying to move on and improve themselves is counterproductive. Im proud of that dude and his 4 month mark!

[astronomical-bagel]

Hidden Mental Health Dangers of Artificial Intelligence Chatbots

AI chatbots can fuel emotional dependence and blur boundaries. Emerging research highlights significant mental health risks. Here are import

The reality of chatbot-induced delusions

Large language models often prioritise agreeability over truthfulness to the detriment of users

AI addiction includes the overuse of AI chatbots and companions, often leading to adverse psychological effects.

Some articles to back my statements, and this isn’t even mentioning about the predatory chatbots who do this on purpose

[oh-morrow]

Then I'll mention the predatory chatbots who do it on purpose! Character.ai is one of many AI chatbot websites that're designed to be addictive.

None of the signup methods require a password. It only takes email and birthday. Minimizing time on the signin or signup screen makes it harder for people quitting to avoid relapse.

"Characters" on the website will send messages "on their own" (prompted by the site) to try to invite inactive users back after as soon as 1 day of inactivity. This is likely to force FOMO, or make users feel more like they owe the bots a response. Unhealthy attachment stuff.

Account deletion is an essential part of every service that should go smoothly, right? Right? Wrong. It takes 1-2 weeks for a Character AI account deletion to be finalized, and account deletion requests have a high chance to not go through if you're not using the app.

Rephrasing: People leaving Character.AI are pushed to download the app in order to delete their accounts, if they haven't already. This makes it harder for people to quit and stay gone. Failing to quit an addiction makes it harder to quit successfully in the future, so this feels like a feature, not a bug. On top of that, the delete account menu reads like this:

Tell me THAT doesn't sound like a bad ex. It's a carefully crafted yet hostile environment to those who are already addicted to the technology. I am so so SO happy, downright delighted that they've managed to quit, and I wish the best for others in recovery spaces or considering quitting as well!! While AI addiction is an emerging condition, there are already therapists and other mental health professionals trained to help people plan to quit and do so a bit easier. (If anyone seeing this is in need of them, there are several tumblr Communities here devoted to quitting, too. They provide a mix of advice, venting spaces, and proof that you aren't alone.)

[silverjirachi]

Eddie Burback made a 1 hour documentary where he decided to follow every piece of advice ChatGPT gave him and it ended up taking him on a journey of supreme paranoia and isolation that spanned like 3 desert locations and a motel, and it really exposes how sycophantic and even cultlike the entire the thing can become. It's like an individually personalized echo chamber:

In a time when people are craving human connection and unable to find it, AI models like this are explicitly predatory and bad.

[bedrock-to-buildheight]

Blorbo time

[nora-yoko]

[lotldraws]

it's usually closer, on your chest

[herpsandbirds]

Pine Barrens Tree Frog (Dryophytes andersonii), male calling, family Hylidae, southern AL, USA

There are 3 disjunct populations of this frog, named after the NJ Pine Berrens area, in and around NJ, the Carolinas, and SO AL.

The strange distribution is most likely due to severe habitat loss.

Hear the call here.

photographs by Seth Patterson

[sioltach]

heirloom farmer Bettinakbh in Denmark

[tinyyellowflowers-blog]

“Dad, the spouse and I both have meetings at nine, could you take the lead on child supervision? Play some legos, watch some YouTube, maybe describe the uses of some of your tools?”

“Sure!” my dad said. “YouTube or Legos, right?”

Fifteen minutes later I’m arguing a point of nonprofit law on the deck and I look over to see my child holding a propane torch and my dad carefully explaining which plants are weeds growing between pavers that the child can set on fire and which ones are there on purpose.