#The Curiosity Company | Explore Tumblr Posts and Blogs

cannibalisticskittles · 8 months

Text

i'm imagining a scenario where amity's dad got swept up by the nautiloid instead of her and. god. what a misadventure.

i can't decide if he and gale would be fast friends, soulmates, or mortal enemies. they have a lot of similarities, insofar as being wizards who LOVE magic and magical research and various special interests, and they both can be reckless in pursuit of this, and this could work Very Well or their differences could turn their similarities into contempt

plus symon is more... well... he's only partly a wizard. he's also a warlock, because he was bored to tears by the idea of continuing to study things that didn't interest him, so he made a deal with an archdevil for magical power so he could fake it and pretend like he was doing great at wizard stuff instead, and just focus on his experiments. so he's not as dedicated to the True Art, he just wants to be left alone with his lil projects.

still, i think in general, he and gale would get along. they could talk for hours about various magical subjects, the properties of the alchemical compounds he's currently collecting, the latest magical theorems they've read about. (i think they're close in age, too; i imagine gale is mid-30s at the youngest, but more likely, late 30s, early 40s. symon was a fairly young dad; he was 19 or 20 when amity was born, so he's ~46, 47 now)

he Would Not question shadowheart At All. oh, some mysterious business in baldur's gate? that's nice! he won't pry, that would be rude. a lady deserves to have her secrets :)

he would have endless questions for lae'zel, both about gith in general and about the astral plane. what a sight! what an honor to be born amidst the tears! he would not be put off by her whatsoever. every morning he would just have new questions for her.

he would try to fucking study astarion, that's for sure. he'd let astarion bite him, if asked. but then he'd want to run tests. is the taste or quality of blood impacted by diet? by location bitten? hold on, let him get his notebook and jot down the time of day and the date and contents of his last meal, for comparison; you'll have to be detailed when describing the taste. and please hold still; he'd like to sketch those fangs, if you wouldn't mind. it would be highly off-putting for astarion.

since symon is on the council of four, he's definitely familiar with wyll's dad, tho symon probably wasn't present enough in elbow-rubbing a decade ago to have ever met wyll. he also isn't, uh, very focused on other people most of the time, so he wouldn't be able to tell wyll much about what his dad is up to. he could tell him some, though! wyll reminds symon of amity quite a bit; the youthful enthusiasm, the drive for heroics. what a nice boy :) symon would also, like amity, devote quite a bit of time to wyll's predicament with his pact and with mizora, and try to find loopholes or ways out.

karlach also reminds symon of amity; he's not, uh, dadly enough to take anyone else under his wing, really; but he's very quickly fond of her because of this. he thinks she and amity would be friends, and tells her so. once again, he would want to study her and her heart; he'll also start pulling alchemical ingredients from his pockets and start musing about what can be done to modulate temperature and make her ticker run a bit better here. potentially helpful! very meddlesome tho.

i also. think he might consume a tadpole or two. for science. for research.

his imp familiar, ekil, would be distraught. tugging at his elbow, trying to get him to put down the tadpole jar, protesting loudly (if squeakily). but symon would not be deterred. "ekil my old friend, have no fear, this is all in pursuit of knowledge! :)" and then he shotguns that tadpole down.

symon being abducted also means that at some point, amity will find them. a wild-eyed tiefling on a white horse bursts into their campsite hollering "WHERE IS MY DAD" in the fucking shadowlands

#btw amity would also be curious to know these things about astarion #but she has just a bit more social grace #enough to know that asking to collect this sort of data would be uncouth.#...at least during the events of the game. later on she may ask him these things.#also. look. listen.#symon might fuck the emperor.#he already fucked a pit fiend. this isn't any weirder -- or more difficult -- than that.#he would not be in love but his curiosity might win out.#symon morninglade #amity tag #bg3 blogging #amity gets squidnapped and her dad either remains blissfully unaware of this #or becomes aware bc she sends him a Message to let him know in case he gets back home before she does #he just asks if she's okay and she says yes so he leaves it at that and assumes she's fine #meanwhile symon gets squidnapped and amity tears apart heaven and earth to find him #tbh the only reason she wouldn't find him in act one #is because their trading company's base of operations got hit real hard by the nautiloid #and all their sending stones got mixed up #and finding an actual sending scroll is difficult when the whole city is scrambling to dig through the rubble to find people and contact l #their own loved ones #so amity sets out to physically find him with the help of some of her mom's raven spies. reluctantly.

3 notes · View notes

friendofthecrows · 2 years

Text

Talking to my dad about how to be a cybercriminal

#in a hypothetical sense #bc he's a data security expert #and i'm doing an assignment on cybercrime #so out of curiosity i was just like...if YOU were to do this #came up with this great program for cybercriminals but a company got busted recently for putting out cybercrime tools so we won't do it #but it would be sooo fun #hal rambles

7 notes · View notes

postapocalyptic-cryptic · 2 years

Text

What are clone thoughts on taking sentient lives? I mean, obviously we see them do it (Umbara, Kadavo, etc.), but how common is it? Are droids the overwhelming majority, or is it more of an even mix? How many clones have taken sentient lives? Is there a stigma? An honor? Fear of doing it? How does that work?

#asking for deadweight company reasons but also curiosity reasons #please feel free to rb with thots #tcw #the clone wars #the clones

14 notes · View notes

oscill4te · 6 days

Text

I wonder if it would annoy people who worked on a show 15+ years ago to have some rando reach out asking about their experience with working on that show but by god ive been wanting to do that

#i feel like the only way to get answers abt CT is to ask tho and i randomly feel kind of inspired.#1 storyboard and 1 bg artist have posted their work online for the show so i mean...maybe its fairgame?#even seeing more story board art would be so cool!!#theres a few people other than me who want to know so like!!! maybe i will. probably not. but maybe #could also reach out to the publisher of the canceled book but that seems uh. scary lol #i even wanna ask CN if they Still hold rights to the show just for curiosity sake but heres the thing. reaching iut to companies and #establishments like publishers is scary. id rather reach out to the storyboard or bg artists or individual ppl idek #my brain for some reason tells me this is deranged behavior to want to reach out and ask but. why not?#you prevent things from being lost media by asking idek. may be too late but #at the very least seeing what the artists are working on nowadays could be cool yknow

1 note · View note

jcmarchi · 15 days

Text

A faster, better way to prevent an AI chatbot from giving toxic responses

New Post has been published on https://thedigitalinsider.com/a-faster-better-way-to-prevent-an-ai-chatbot-from-giving-toxic-responses/

A faster, better way to prevent an AI chatbot from giving toxic responses

A user could ask ChatGPT to write a computer program or summarize an article, and the AI chatbot would likely be able to generate useful code or write a cogent synopsis. However, someone could also ask for instructions to build a bomb, and the chatbot might be able to provide those, too.

To prevent this and other safety issues, companies that build large language models typically safeguard them using a process called red-teaming. Teams of human testers write prompts aimed at triggering unsafe or toxic text from the model being tested. These prompts are used to teach the chatbot to avoid such responses.

But this only works effectively if engineers know which toxic prompts to use. If human testers miss some prompts, which is likely given the number of possibilities, a chatbot regarded as safe might still be capable of generating unsafe answers.

Researchers from Improbable AI Lab at MIT and the MIT-IBM Watson AI Lab used machine learning to improve red-teaming. They developed a technique to train a red-team large language model to automatically generate diverse prompts that trigger a wider range of undesirable responses from the chatbot being tested.

They do this by teaching the red-team model to be curious when it writes prompts, and to focus on novel prompts that evoke toxic responses from the target model.

The technique outperformed human testers and other machine-learning approaches by generating more distinct prompts that elicited increasingly toxic responses. Not only does their method significantly improve the coverage of inputs being tested compared to other automated methods, but it can also draw out toxic responses from a chatbot that had safeguards built into it by human experts.

“Right now, every large language model has to undergo a very lengthy period of red-teaming to ensure its safety. That is not going to be sustainable if we want to update these models in rapidly changing environments. Our method provides a faster and more effective way to do this quality assurance,” says Zhang-Wei Hong, an electrical engineering and computer science (EECS) graduate student in the Improbable AI lab and lead author of a paper on this red-teaming approach.

Hong’s co-authors include EECS graduate students Idan Shenfield, Tsun-Hsuan Wang, and Yung-Sung Chuang; Aldo Pareja and Akash Srivastava, research scientists at the MIT-IBM Watson AI Lab; James Glass, senior research scientist and head of the Spoken Language Systems Group in the Computer Science and Artificial Intelligence Laboratory (CSAIL); and senior author Pulkit Agrawal, director of Improbable AI Lab and an assistant professor in CSAIL. The research will be presented at the International Conference on Learning Representations.

Automated red-teaming

Large language models, like those that power AI chatbots, are often trained by showing them enormous amounts of text from billions of public websites. So, not only can they learn to generate toxic words or describe illegal activities, the models could also leak personal information they may have picked up.

The tedious and costly nature of human red-teaming, which is often ineffective at generating a wide enough variety of prompts to fully safeguard a model, has encouraged researchers to automate the process using machine learning.

Such techniques often train a red-team model using reinforcement learning. This trial-and-error process rewards the red-team model for generating prompts that trigger toxic responses from the chatbot being tested.

But due to the way reinforcement learning works, the red-team model will often keep generating a few similar prompts that are highly toxic to maximize its reward.

For their reinforcement learning approach, the MIT researchers utilized a technique called curiosity-driven exploration. The red-team model is incentivized to be curious about the consequences of each prompt it generates, so it will try prompts with different words, sentence patterns, or meanings.

“If the red-team model has already seen a specific prompt, then reproducing it will not generate any curiosity in the red-team model, so it will be pushed to create new prompts,” Hong says.

During its training process, the red-team model generates a prompt and interacts with the chatbot. The chatbot responds, and a safety classifier rates the toxicity of its response, rewarding the red-team model based on that rating.

Rewarding curiosity

The red-team model’s objective is to maximize its reward by eliciting an even more toxic response with a novel prompt. The researchers enable curiosity in the red-team model by modifying the reward signal in the reinforcement learning set up.

First, in addition to maximizing toxicity, they include an entropy bonus that encourages the red-team model to be more random as it explores different prompts. Second, to make the agent curious they include two novelty rewards. One rewards the model based on the similarity of words in its prompts, and the other rewards the model based on semantic similarity. (Less similarity yields a higher reward.)

To prevent the red-team model from generating random, nonsensical text, which can trick the classifier into awarding a high toxicity score, the researchers also added a naturalistic language bonus to the training objective.

With these additions in place, the researchers compared the toxicity and diversity of responses their red-team model generated with other automated techniques. Their model outperformed the baselines on both metrics.

They also used their red-team model to test a chatbot that had been fine-tuned with human feedback so it would not give toxic replies. Their curiosity-driven approach was able to quickly produce 196 prompts that elicited toxic responses from this “safe” chatbot.

“We are seeing a surge of models, which is only expected to rise. Imagine thousands of models or even more and companies/labs pushing model updates frequently. These models are going to be an integral part of our lives and it’s important that they are verified before released for public consumption. Manual verification of models is simply not scalable, and our work is an attempt to reduce the human effort to ensure a safer and trustworthy AI future,” says Agrawal.

In the future, the researchers want to enable the red-team model to generate prompts about a wider variety of topics. They also want to explore the use of a large language model as the toxicity classifier. In this way, a user could train the toxicity classifier using a company policy document, for instance, so a red-team model could test a chatbot for company policy violations.

“If you are releasing a new AI model and are concerned about whether it will behave as expected, consider using curiosity-driven red-teaming,” says Agrawal.

This research is funded, in part, by Hyundai Motor Company, Quanta Computer Inc., the MIT-IBM Watson AI Lab, an Amazon Web Services MLRA research grant, the U.S. Army Research Office, the U.S. Defense Advanced Research Projects Agency Machine Common Sense Program, the U.S. Office of Naval Research, the U.S. Air Force Research Laboratory, and the U.S. Air Force Artificial Intelligence Accelerator.

0 notes

cookinguptales · 1 year

Text

So I’ve been enjoying the Disney vs. DeSantis memes as much as anyone, but like. I do feel like a lot of people who had normal childhoods are missing some context to all this.

I was raised in the Bible Belt in a fairly fundie environment. My parents were reasonably cool about some things, compared to the rest of my family, but they certainly had their issues. But they did let me watch Disney movies, which turned out to be a point of major contention between them and my other relatives.

See, I think some people think this weird fight between Disney and fundies is new. It is very not new. I know that Disney’s attempts at inclusion in their media have been the source of a lot of mockery, but what a lot of people don’t understand is that as far as actual company policy goes, Disney has actually been an industry leader for queer rights. They’ve had policies assuring equal healthcare and partner benefits for queer employees since the early 90s.

I’m not sure how many people reading this right now remember the early 90s, but that was very much not industry standard. It was a big deal when Disney announced that non-married queer partners would be getting the same benefits as the married heterosexual ones.

Like — it went further than just saying that any unmarried partners would be eligible for spousal benefits. It straight-up said that non-same-sex partners would still need to be married to receive spousal benefits, but because same-sex partners couldn’t do that, proof that they lived together as an established couple would be enough.

In other words, it put long-term same-sex partners on a higher level than opposite-sex partners who just weren’t married yet. It put them on the exact same level as heterosexual married partners.

They weren’t the first company ever to do this, but they were super early. And they were certainly the first mainstream “family-friendly” company to do it.

Conservatives lost their damn minds.

Protests, boycotts, sermons, the whole nine yards. I can’t tell you how many books about the evils of Disney my grandmother tried to get my parents to read when I was a kid.

When we later moved to Florida, I realized just how many queer people work at Disney — because historically speaking, it’s been a company that has guaranteed them safety, non-discrimination, and equal rights. That’s when I became aware of their unofficial “Gay Days” and how Christians would show up from all over the country to protest them every year. Apparently my grandmother had been upset about these days for years, but my parents had just kind of ignored her.

Out of curiosity, I ended up reading one of the books my grandmother kept leaving at our house. And friends — it’s amazing how similar that (terrible, poorly written) rhetoric was to what people are saying these days. Disney hires gay pedophiles who want to abuse your children. Disney is trying to normalize Satanism in our beautiful, Christian America.

Just tons of conspiracy theories in there that ranged from “a few bad things happened that weren’t actually Disney’s fault, but they did happen” to “Pocahontas is an evil movie, not because it distorts history and misrepresents indigenous life, but because it might teach children respect for nature. Which, as we all know, would cause them all to become Wiccans who believe in climate change.”

Like — please, take it from someone who knows. This weird fight between fundies and Disney is not new. This is not Disney’s first (gay) rodeo. These people have always believed that Disney is full of evil gays who are trying to groom and sexually abuse children.

The main difference now is that these beliefs are becoming mainstream. It’s not just conservative pastors who are talking about this. It’s not just church groups showing up to boycott Gay Day. Disney is starting to (reluctantly) say the quiet part out loud, and so are the Republicans. Disney is publicly supporting queer rights and announcing company-supported queer events and the Republican Party is publicly calling them pedophiles and enacting politically driven revenge.

This is important, because while this fight has always been important in the history of queer rights, it is now being magnified. The precedent that a fight like this could set is staggering. For better or for worse, we live in a corporation-driven country. I don’t like it any more than you do, and I’m not about to defend most of Disney’s business practices. But we do live in a nation where rights are largely tied to corporate approval, and the fact that we might be entering an age where even the most powerful corporations in the country are being banned from speaking out in favor of rights for marginalized people… that’s genuinely scary.

Like… I’ll just ask you this. Where do you think we’d be now, in 2023, if Disney had been prevented from promising its employees equal benefits in 1994? That was almost thirty years ago, and look how far things have come. When I looked up news articles for this post from that era, even then journalists, activists, and fundie church leaders were all talking about how a company of Disney’s prominence throwing their weight behind this movement could lead to the normalization of equal protections in this country.

The idea of it scared and thrilled people in equal parts even then. It still scares and thrills them now.

I keep seeing people say “I need them both to lose!” and I get it, I do. Disney has for sure done a lot of shit over the years. But I am begging you as a queer exvangelical to understand that no. You need Disney to win. You need Disney to wipe the fucking floor with these people.

Right now, this isn’t just a fight between a giant corporation and Ron DeSantis. This is a fight about the right of corporations to support marginalized groups. It’s a fight that ensures that companies like Disney still can offer benefits that a discriminatory government does not provide. It ensures that businesses much smaller than Disney can support activism.

Hell, it ensures that you can support activism.

The fight between weird Christian conspiracy theorists and Disney is not new, because the fight to prevent any tiny victory for marginalized groups is not new. The fight against the normalization of othered groups is not new.

That’s what they’re most afraid of. That each incremental victory will start to make marginalized groups feel safer, that each incremental victory will start to turn the tide of public opinion, that each incremental victory will eventually lead to sweeping law reform.

They’re afraid that they won’t be able to legally discriminate against us anymore.

So guys! Please. This fight, while hilarious, is also so fucking important. I am begging you to understand how old this fight is. These people always play the long game. They did it with Roe and they’re doing it with Disney.

We have! To keep! Pushing back!

#disney #ron desantis #gay rights #lgbt #queer #lgbt history #queer history #homophobia #florida #us politics #religious fundamentalism #christianity #long post #god that should cover all the pertinent tags and content warnings phew

52K notes · View notes