Yeah I still rely on fansubs for all my jdramas that don’t get official releases outside Japan (which is still most of them). The quality is instantly obvious compared to machine-generated crap.
Cultural context and an understanding of tone and narrative coherence is critical for subtitles if you want to do justice to the quality of the writing in the original language.
Yeah the irony of it all is that professionally subtitled shows were already inferior to fansubs, because the fansubbers were a self-selecting group of people who were fans of that show specifically. I guess Crunchyroll realized that their subs sucked anyway so why pay money for them?
Don’t forget that those fansubbers also want the watchers to get the feeling of the Japanese dubs, switching out sentences to match the voiced lines and still bring you the message, while non-fans just translate it.
I don’t see what the problem is with using AI for translations. if the translations are good enough and cheap enough, they should be used. If they’re not good enough, another business can offer better translations as a differentiator.
A good human translator is always the best solution.
But if the choice would be between crappy google translate or a LLM I would take the LLM translation.
There's no excuse for a big studio, they should hire translators. But for indie creators without a budget it can be the best way to get their creation to more people.
Even if it could it wouldn't matter, what makes art so valuable is the amount of work put into it. It's the highest effort form of communication you can imagine, it can express so much. An LLM chat bot could never do that.
If you need translation for just getting facts and information for say math equation and its annotation translated, there's little margin in variety, what you need is database - that's mostly fine.
Pieces need translation are usually not like that. They have cultural context, pun, wordplay in rhymes, structural parallel, underlying tone, a lot of things only work in the language originally written. Translation is always a (nearly impossible) challenge for the translator to reconstruct all of them in target language.
I did game translation for a while. Translation is a field where AI hit first and honestly I've seen people lowering standards. The criteria of "good enough", "passable" is not the same compared to pre-AI days, and will keep changing. I'm almost sure this trend is happening in every industry the same way, and "just translation" is a slippery slope.
I don’t know why it’s so difficult for these Ayn Rand readers to understand that almost no market is perfectly free, efficient, contestable or optimal. When it comes to licensing IP in particular, the overhead of trying to get permission for a show that’s already licensed by a competitor is an insurmountable barrier to entry for anyone who isn’t already a huge media conglomerate.
The only realistic alternative to shitty subtitles is piracy.
I don’t see what the problem is with using AI for translations. if the translations are good enough and cheap enough, they should be used.
Because machine translations for any large chunk of text are consistently awful: they don't get references right, they often miss the point of the original utterance, they ignore cultural context, so goes on. It's like wiping your arse with an old sock - sure, you could do it in a pinch, but you definitively don't want to do it regularly!
::: spoiler Verbose example, using Portuguese to English I'll give you an example, using PT→EN because I don't speak JP. Let's say Alice tells Bob "ma' tu é uma nota de três pila, né?" (literally: "bu[t] you're a three bucks bill, isn't it?") . A human translator will immediately notice a few things:
It's an informal and regional register. If Alice typically uses this register, it's part of her characterisation; else, it register shift is noteworthy. Either way, it's meaningful.
There's an idiom there; "nota de três pila" (three bucks bill). It conveys some[thing/one] is blatantly false.
There's a rhetorical question, worded like an accusation. The scene dictates how it should be interpreted.
So depending on the context, the translator might translate this as "ain't ya full of shit...", or perhaps "wow, you're as fake as Monopoly money, arentcha?". Now, check how chatbots do it:
GPT-4o mini: "But you're a three-buck note, right?"
Llama 4 Scout: "But you are a three-dollar bill, aren't you?"; or "You're a three-dollar bill, right?" (it offers both alternatives)
Both miss the mark. If you talk about three dollar bills in English, lots of people associate it with gay people, creating an association that simply does not exist in the original. The extremely informal and regional register is gone, as well as the accusatory tone.
With Claude shitting this pile of idiocy, that I had to screenshot because otherwise people wouldn't believe me:
\ [This is wrong on so many levels I don't... I don't even...]
This is what you get for AI translations between two IE languages in the same Sprachbund, that'll often do things in a similar way. It gets way worse for Japanese → English - because they're languages from different families, different cultures, that didn't historically interact that much. It's like the dumb shit above, multiplied by ten. :::
If they’re not good enough, another business can offer better translations as a differentiator.
That "business" is called watching pirated anime with fan subs, made by people who genuinely enjoy anime and want others to enjoy it too.
As much as people hate on dubs, this one reason I’m glad I tend to focus on them. The ADR team does a great job of translating and localization. Is it word for word the same as Japanese? No, if you want that, use AI. This keeps the spirit while making sure it’s more culturally understandable.
Time to cancel my Crunchyroll subscription. Oh wait I don't have one, I simply torrent my series.
Seriously now. The anime fansubbing scene is one that makes me genuinely happy. It shows me there are plenty amateurs out there that are as good or better than plenty professionals like me.
Sadly the german fansubbing scene basically evaporated when simulcast became a thing. 15 years ago you got some amazing subs, with signs translated (even when moving through the screen) and cultural information shown when it was needed to understand a scene or joke
I was already bitter about patrasche being translated/localized(?) as patlash in re:zero but things are getting even shittier I guess. For those not in the know Dog of Flanders is a story nearly every japanese person will know at least to some degree, and has been referenced infinitely in anime and manga.
And similarly so many weird fucking names you could search the katakana online and find 30 different game wikis that reveal it was just akward katakana for a common non-japanese name, but no it has to to be different in this anime or that manga and we are going to make up some aneurysm romanization for it.
or incorrect subject assumtion because japanese is vague on that and if it's ai translated it actually makes sense with the lack of context awareness, and the list could go on for days
If anyone trusted Crunchyroll after they removed comments and reviews, it's honestly kind of their fault, as much as I hate victim blaming. They have shown who they are time and time again, it's not hard to believe them.
Though, I think anyone in this community very likely already knows exactly who they are.
Bought some gatorade powder you can see in the photo. The recommended serving is 4 tablespoons for 750 mL of gatorade. I used around half of that, which is 10 spoons for 4 L (orange) and 10 spoons for 3 L (blue). ChatGPT recommended a quarter, but I was afraid it wouldn't taste anything. I'm also pretty confident he never tried it anyway, so who cares about the computer's opinion. Then it's a standard 1 kg of honey each and EC-1118 yeast with some nutrient for the road (I realize now that I might have put way too much of it, but I think they'll manage). Both were extremely sweet before the fermentation. I have doubts about the blue one, but high hopes for the orange one (this is no political statement btw). I actually have no idea how it will come out, but right now they're both bubbling nicely.
Never heard of it, but it sounds fun. I like mead because you can turn pretty much any drink into alcohol with it and the only specialized ingredient you need is yeast and sometimes nutrients. Everything else can be found at the grocery store. Turned some sodas (mountain dew being a classic, but also root beer), and fruit juices into mead, with mostly good results. Whereas ciders and wine requires the right kind of fruits, and I'm sure I'll never do better than what is already on the market. The future is in weird drinks. But lemonade wine seems nice, maybe I'll try it next.
Gatorwine is a thing that the internet has made known. It's a mix of wine and blue Gatorade. Babish mixed some up as part of an "I do your suggestions" episode.
Turns out that it isn't as horrible as it would seem. To the contrary, he liked it and did a second episode mixing up more varieties. So now it's a thing everywhere since he's fairly popular.
I could kind of see that working. Wine coolers are soda and wine, sangria is wine and juice, gatorade is just kool-aid and salt so it's somewhere between soda and juice. It's not classy, but it's not bad.
Your comparison to wine coolers is dead on, imo. It has that same overall profile where it's sweet and easy to drink. It isn't great, but it's not bad, and I could see someone that lollies likes drinking going through gallons of it. I don't like drinking, and I managed to kill a whole glass, which is saying a lot for me.
I think the blue one might have turned bad (probably my fault). It tastes like bacteria or something. The drink turned a weird green that can either look like an especially healthy potion, or some really chemical poison. (blue gatorade and yellow honey makes that, apparently) The orange one is not so bad, but nothing really good either. Right after bottling, I still discerned the chemical orange flavour, which was kind of nice, but after a few weeks of aging in bottles, I have a hard time telling what it tastes exactly. I can tell it's gatorade-based, but without a particular, recognizable gatorade flavor. The color is nice, though. It kept the clear, distinguishable orange from the gatorade. Just a bit darker. Taste-wise, it's definitively the worst batch I've made so far. Otherwise, I didn't get any headache from it and the orange one is still drinkable. 4/10 because it's got electrolytes.
I know I can google this, but I wanted to hear from the people as well. If someone shows you an album and asks your opinion, what would you look for?
My take: I know it is subjective to a degree. I know principles in composition are important But I am not sure I could be part of a discussion on the topic and it would be cool to!
Separate challenges by immigrant rights groups and an American detained by federal agents accuse officers of racial profiling, brutality and unlawful detentions.
If the budget bill passes, as seems likely to happen in the next couple hours, these abuses are going to go national.
Stop Killing Games is an European Citizens Initiative aiming to keep games playable even after their developers and publishers have stopped supporting it.
To get the initiative onto the EUs agenda so it has the chance to become EU law, it has to both reach 1 million signatures total and minimum thresholds in at least 7 countries. Now both of those goals have been reached. But that's no reason to stop signing! Some signatures will get thrown out in the validation phase because the signee made a mistake. So keep signing and show the world just how many people are in favour of saving videogames.
In a 92-minute interview with Tucker Carlson on Monday, RFK Jr. drilled down on his vision for the US Department of Health and Human Services (HHS). Artificial intelligence — arguably, a uselessly vague umbrella term — came up multiple times. (As did conspiracy theories and disinformation on vaccines and autism, the medical establishment, and covid-19 deaths.)
As the head of HHS, Kennedy said his federal department is undergoing an “AI revolution.” He implored viewers to “stop trusting the experts,” as highlighted by Gizmodo, and, presumably, put their trust into AI instead of decades of scientific consensus.
[image of a svobenian google search result saying earth's age is 4.543 billion years]
A: how is it
A: (that) the earth is 4543 [sic] billion years old but the year is (only) 2025
B: idiot 🤦
Also a bit of a context, Svobenian is my Germanic conlang with strong Slavic influence (basically a descendant of Proto-Germanic with Slavic sound changes and many proto-slavic borrowings). I made this joke to demonstrate the difference between “vėtar” (year, cognate to English winter, used only as a time unit, e.g. 4.543 mlrd vėtř, meaning 4.543 billion years) and “jero” (cognate to English year, used in other contexts, e.g. jero 2025 “year 2025”)
Interesting. It’s a germanic language. Because “debil” the only part I got (having no knowledge of this conlang), I got because “débile” for dumb exists in french. Which I think comes from Latin debilis.
This word also exists in Polish, Lithuanian, and Russian, as a loanword from French. Because I imagine speakers of Svobenian living somewhere in central Europe (i even have the whole althist thing where i made Svobenia, country of Svobenians, being located in what is now Kaliningrad oblast), i find it sensible that they'd also borrow this word.
And yeah, it's a Germanic language, developing separately from probably originally Eastern Germanic branch, somewhere in what is now Mecklenburg-Vorpommern, Germany, but then got heavily influenced by some early West Slavic language. I'm gonna elaborate on the whole history of svobenes once i actually "finish" the language
Sono passati 3 anni dal mio primo approdo nel fediverso, e credo che sia abbastanza per farmi un idea generale su cosa sia e delineare dei pregi e difetti. sicuramente la cosa che mi piace è che è in costante evoluzione anche se devo dire molto lenta. In
Sono passati 3 anni dal mio primo approdo nel fediverso, e credo che sia abbastanza per farmi un idea generale su cosa sia e delineare dei pregi e difetti.
Non avevo letto il tuo post namirblog@namirblog.altervista.org (ti menziono anche come peppenamir@www.foxyhole.io perché non so se funziona il commento a wordpress) ma mi è capitato di vederlo ora, a poch ore dal lancio di questo nostro progetto citiverse.it
Sono d'accordo con la tua analisi: il Fediverso va promosso fuori dal Fediverso e, possibilmente, bisognerebbe anche evitare di chiamarlo Fediverso 😀
Chissà se il nostro citiverse riuscirà ad avvicinare un po' di gente?
PS: intanto sono molto contento di aver potuto "prendere" il tuo post di Wordpress e averlo "spostato" all'interno della categoria "Discussioni sul Fediverso"
Citiverse, la città dei social liberi, è l'alternativa aperta e italiana ai gruppi di Facebook dove è possibile parlare con persone della tua città o della tua tua organizzazione.
The 15 ministers said that Israel's "strategic partnership, backing, and support of the U.S. and President Donald Trump" make this a "propitious time" to formally steal most of Palestine.
It’s quite jarring how everything pre-season 7 is compared to the rest. Like most the characters are quite deep with conflicting emotions and not caricatures. Especially season 2. Like “Bart gets an F” episode, that kind of empathetic emotional portrayal of bart would never happen post season 10.
And we really went from Homer is a dumb guy who genuinely loves his family and is suffering under an oppressive system, to well “jackass homer”.
Shortening attention spans had to be part of it. It was probably hard to compete with other shows that had rapid-fire jokes and shorter time to pay off. Building up complex characters and creative situations takes a little more time and probably loses eyeballs.
Sort of analogous to long form vs short form videos now (but obviously both are much further down that road in comparison).
This process is called Flanderization, whereby a character on a long running show becomes a self-parody as their most distinctive traits and behaviors are amplified again and again. It's named for a popular side character Ned Flanders, from the show the Simpsons. Though arguably Ned undergoes more permanent personal growth than any other character on the show.
Un terribile incidente stradale ha portato via Diogo Jota, attaccante del Liverpool e pilastro della Nazionale portoghese. Aveva solo 28 anni. Con lui ha perso la vita anche il fratello minore, André Jota, anch’egli calciatore professionista. La tragedia si è consumata nelle prime ore del mattino lungo l’autostrada A-52, nei pressi di Villanueva de Sanabria, nella provincia di Zamora, in Spagna.
Today I can share a major development status update of XPipe, a connection hub that allows you to access your entire server infrastructure from your local desktop. It can make your life easier when working with any kind of servers by eliminating all the commonly tedious tasks that come up when interacting with remote systems, either from the terminal or from a graphical interface. XPipe comes with integrations for SSH, docker and other containers, various hypervisors, and more without requiring setup on your remote systems. You can also keep using your favourite text/code editors, terminals, password managers, shells, command-line tools, and more with it.
Docker compose
This release introduces support for docker compose. Containers in compose projects are grouped together and can be managed all at the same time via compose project entries.
The container state information shown is also improved, always showing the container state in combination with the system information.
Batch mode
There is now a batch mode available that allows you to select multiple systems via checkboxes and perform actions for the entire batch. This can include starting/stopping, automatically adding available subconnections, or running scripts on all selected systems.
You can toggle the batch mode in the top left corner.
Password managers
The password manager integrations have been upgraded: - There is now support for KeePassXC - All password manager integrations have been reworked to work out of the box without configuration - There is now support to use password manager SSH agents more easily - You can now unlock the xpipe vault with your password manager
Terminals
The terminal integration comes with many new features: - There is now built-in support for the terminal multiplexers tmux, zellij, and screen. This is especially useful for terminals without tabbing support. - There is also now built-in support for custom prompts with starship, oh-my-posh, and oh-my-zsh. - On Windows, you now have the ability to use a WSL distribution as the terminal environment, allowing you to use the new terminal multiplexer integration seamlessly on Windows systems as well.
SSH
Various improvements were made to the SSH implementation: - The SSH gateway implementation has been reworked so that you can now use local SSH keys and other identities for connections with gateways - The VSCode SSH remote integration has been reworked to allow more connections it to be opened in vscode. It now supports essentially all simple SSH connections, custom SSH connections, SSH config connections, and VM SSH connections. This support includes gateways - There is now built-in support to refresh an SSO openpubkey with the opkssh tool when needed - There is now the option to enable verbose ssh output to diagnose connection issues better - For VMs, you can now choose to not use the hypervisor host as SSH gateway and instead directly connect to the VM IP
Other
Connection names, e.g. VM names, will now automatically update on refresh when they were changed
You can now launch custom scripts within XPipe with a command output dialog window without having to open a terminal
Various installation types like the linux apt/rpm repository and homebrew installations now support automatic updates as well
The k8s integration will now automatically add all namespaces for the current context when searching for connections
The application window will now hide any unnecessary sidebars when being resized to a small width. This makes it much easier to use XPipe in a tiling window arrangement
The webtop has been updated to have terminal multiplexers, proper konsole tab support, disabled kwallet, and more
Various error messages and connection creation dialogs now contain a help link to the documentation sections
A note on the open-source model
Since it has come up a few times, in addition to the note in the git repository, I would like to clarify that XPipe is not fully FOSS software. The core that you can find on GitHub is Apache 2.0 licensed, but the distribution you download ships with closed-source extensions. There's also a licensing system in place with limitations on what kind of systems you can connect to in the community edition as I am trying to make a living out of this. I understand that this is a deal-breaker for some, so I wanted to give a heads-up.
Outlook
If this project sounds interesting to you, you can check it out on GitHub, visit the Website, or check out the Docs for more information.
More than 175 Democratic members of Congress are filing an amicus brief on Thursday opposing the Trump administration's overhaul of the U.S. Department of Education.
“The law couldn't be clearer: the president does not have the authority to unilaterally abolish the Department of Education,” Sen. Elizabeth Warren wrote in a statement first obtained by ABC News, adding, “Donald Trump is not a king, and he cannot single-handedly cut off access to education for students across this country.”
Warren and Reps. Jamie Raskin, Bobby Scott and Rosa DeLauro -- the ranking members of the House’s Education and Judiciary committees -- are leading the 15-page legal document. They’re joined by Senate Minority Leader Chuck Schumer, more than 20 Senate Democrats, and more than 150 other members of the House Democratic caucus.
The U.S. liquid coffee market is undergoing a rapid transformation fueled by evolving consumer preferences, the rising trend of on-the-go consumption, and a growing appetite for premium ready-to-drink (RTD) coffee offerings. According to insights from Persistence Market Research, the market has witnessed double-digit growth over the past five years and is expected to continue this trajectory through 2030.
A Market Percolating with Opportunity The shift toward convenience and premiumization has redefined the beverage landscape in the United States. The modern consumer, often strapped for time but unwilling to compromise on quality, is embracing liquid coffee formats such as bottled cold brews, canned lattes, and bag-in-box coffee concentrates.
The U.S. liquid coffee market—which includes RTD coffee beverages, coffee concentrates, and extract-based products—is poised to reach an estimated valuation of USD 6.8 billion by the end of 2025, growing at a compound annual growth rate (CAGR) of 9.2% from 2022 to 2025.
Key Growth Drivers Convenience and On-the-Go Culture The American work culture, characterized by long hours and hectic schedules, has accelerated the demand for grab-and-go beverages. Liquid coffee, available in portable and resealable formats, offers the perfect solution for busy professionals and commuters.
Cold Brew and Specialty Trends Cold brew coffee has emerged as a consumer favorite, appreciated for its smooth taste and lower acidity. Liquid coffee brands are leveraging this trend by expanding their product lines with flavored cold brews, nitro-infused variants, and non-dairy options.
Health-Conscious Consumers With wellness trends taking center stage, manufacturers are introducing low-sugar, plant-based, and functional coffee beverages enhanced with adaptogens, vitamins, and probiotics. These innovations appeal to the health-savvy consumer looking for a caffeine boost without the guilt.
Millennials and Gen Z Influence Younger demographics are driving demand for ethically sourced, sustainably packaged, and digitally accessible products. Liquid coffee brands are responding with eco-conscious packaging and social media campaigns targeting this tech-native generation.
Competitive Landscape and Innovation The U.S. liquid coffee space is highly competitive, with both established beverage conglomerates and emerging specialty players vying for shelf space. Industry giants like Starbucks, Nestlé, and PepsiCo (via its Starbucks RTD partnership) continue to dominate, but smaller brands such as La Colombe, Chameleon Cold-Brew, and Stumptown are capturing niche markets with craft appeal and innovative formulations.
To stay ahead, brands are investing in:
Sustainable Packaging: Recyclable cans, compostable cartons, and reusable glass bottles.
Flavor Experimentation: Infusions with lavender, vanilla oat, mocha almond, or turmeric.
Tech Integration: QR-coded packaging that leads to origin stories, brewing tips, and loyalty programs.
According to Persistence Market Research, technological advancements in cold-brew extraction, shelf-stable preservation, and nitrogen infusion are expected to further boost product quality and consumer trust.
Distribution Channels Expand Beyond Traditional Retail Another key factor contributing to the growth of the liquid coffee market is the diversification of distribution channels. While supermarkets and convenience stores remain primary sales points, other avenues are rapidly expanding:
E-commerce and Direct-to-Consumer (DTC) Platforms: Especially post-pandemic, online subscriptions and personalized delivery models have seen a major uptick.
Foodservice and Quick-Service Restaurants (QSRs): Many coffee brands are now offering liquid coffee solutions in bulk formats for offices, cafes, and hotels.
Vending and Micromarkets: Automated coffee vending with ready-to-drink options is increasingly common in office complexes, airports, and universities.
Challenges and Regulatory Outlook Despite its impressive growth, the market faces certain challenges:
Shelf Life and Storage Concerns: Ensuring stability and flavor retention over extended periods remains a concern, particularly for dairy-based and natural formulations.
Label Transparency and Clean Label Demands: Consumers are demanding clear ingredient lists, ethical sourcing, and minimal additives—requiring brands to reformulate or lose market share.
Price Sensitivity: Premium liquid coffee products can be priced significantly higher than traditional ground coffee, posing challenges in more price-conscious segments.
Still, with robust investment and evolving technology, these challenges are being addressed head-on. Persistence Market Research identifies increased R&D investments and regulatory compliance as key to long-term sustainability.
Regional Insights Within the United States, urban areas lead in consumption, with California, New York, Texas, and Florida being the largest markets. The West Coast, in particular, shows high adoption of cold brew and nitro varieties, while the Midwest exhibits growing interest in bulk and concentrate formats for office and institutional use.
Rural regions, though slower to adopt, are catching up via retail expansion and targeted marketing by national brands.
Future Outlook With consumers increasingly seeking high-quality, convenient caffeine experiences, the U.S. liquid coffee market is set to continue its robust expansion. As manufacturers focus on innovation, sustainability, and personalization, new product launches are expected to accelerate.
Persistence Market Research predicts that by 2030, more than 35% of total coffee consumption in the U.S. could come from liquid formats, marking a dramatic shift in traditional coffee consumption patterns.
The U.S. liquid coffee market outlook, projected to grow at a 6.4% compound annual growth rate, reaching US$ 6,913.4 Mn by 2032 from US$ 4,482.3 Mn in 2025
The development of AI in China enables the transformation of the authoritarian regime's passive censorship practice into the proactive shaping of people’s cognitive framework in favour of the single-party regime.
The development of AI in China enables the transformation of the authoritarian regime's passive censorship practice into the proactive shaping of people’s cognitive framework in favour of the single-party regime.
Gemini AI needs to be disabled on Android or it will override your privacy settings and gain full access to your texts, calls, and WhatsApp - even if you’ve turned off Gemini Apps Activity.
Gemini AI needs to be disabled on Android or it will override your privacy settings and gain full access to your texts, calls, and WhatsApp - even if you’ve turned off Gemini Apps Activity.
Studi clinici hanno dimostrato che le iniezioni semestrali di lenacapavir offrono una protezione quasi al 100% contro l'infezione da HIV. Tuttavia, restano importanti interrogativi sull'accessibilità economica del farmaco.
L'Organizzazione Mondiale della Sanità ha aggiunto la variante, formalmente nota come XFG, alla sua lista di monitoraggio. Tra i suoi sintomi principali rientrano la secchezza e l'irritazione della gola.
Large language models (LLMs) like ChatGPT can generate and revise text with human-level performance. These models come with clear limitations, can produce inaccurate information, and reinforce existing biases. Yet, many scientists use them for their scholarly writing. But how widespread is such LLM usage in the academic literature? To answer this question for the field of biomedical research, we present an unbiased, large-scale approach: We study vocabulary changes in more than 15 million biomedical abstracts from 2010 to 2024 indexed by PubMed and show how the appearance of LLMs led to an abrupt increase in the frequency of certain style words. This excess word analysis suggests that at least 13.5% of 2024 abstracts were processed with LLMs. This lower bound differed across disciplines, countries, and journals, reaching 40% for some subcorpora. We show that LLMs have had an unprecedented impact on scientific writing in biomedical research, surpassing the effect of major world events such as the COVID pandemic.
Large language models (LLMs) like ChatGPT can generate and revise text with human-level performance. These models come with clear limitations, can produce inaccurate information, and reinforce existing biases. Yet, many scientists use them for their scholarly writing. But how widespread is such LLM usage in the academic literature? To answer this question for the field of biomedical research, we present an unbiased, large-scale approach: We study vocabulary changes in more than 15 million biomedical abstracts from 2010 to 2024 indexed by PubMed and show how the appearance of LLMs led to an abrupt increase in the frequency of certain style words. This excess word analysis suggests that at least 13.5% of 2024 abstracts were processed with LLMs. This lower bound differed across disciplines, countries, and journals, reaching 40% for some subcorpora. We show that LLMs have had an unprecedented impact on scientific writing in biomedical research, surpassing the effect of major world events such as the COVID pandemic.
The Fediverse Passport would be the central account for all users on the Fediverse.
How it would work
a. Upon signing up for the a platform on the Fediverse the user would be redirected to the "Create your Passport" You would create your unique username. Once signed up you would then have an account on every platform connected to the Fediverse.
b. If someone friends/follows you on one platform they would automatically follow you on all platforms. Insuring that communities and friends could stay connected across platforms.
c. The passport for the user would show your feed on all platforms and allow you to selected which platform you want to see your feed from, also allowing you to directly search your content so you could find a post for whatever reason you need.
d. For the subscriber it would show them your feed and allow them to easily find your content.
e. Tons of customization options including the ability to monetize and or set a subscription fee for the video, blogging, and other "arts" platforms.
Safe Guards
You would be allowed to set your privacy setting to, Public, Subscribers Only, Approve Subscribers, Mutual Friends only, Private (Requires link)
Benefit
Would allow stream less interaction across the whole Fediverse and really get it going. No more having to create a different account on each platform and now you can claim an identity and keep track of your communities, also each site would directly help "advertise" the others.
According to Apple, the EU is favoring certain companies and discriminating Apple. Users in the EU are still waiting for their iPhone mirroring on Macs,...
È probabilmente una cometa ed è solo la terza osservazione di questo tipo di qualcosa che ha avuto origine altrove
Un oggetto interstellare, cioè proveniente da un sistema solare diverso dal nostro, si sta dirigendo a grandissima velocità verso la Terra, ma non c’è nessun pericolo di impatto con il nostro pianeta. Dalle osservazioni svolte finora sembra che sia una cometa ed è stata chiamata 3I/ATLAS: è il terzo oggetto in viaggio tra sistemi solari diversi a essere stato mai notato nelle nostre vicinanze da quando puntiamo i telescopi verso il cielo. Nei prossimi mesi sarà studiato dagli astronomi, per approfondire le nostre conoscenze su come alcuni oggetti si mettano a vagare tra diversi sistemi solari all’interno della Via Lattea, la nostra galassia.
Nel sistema solare in cui ci troviamo si stima ci siano milioni di asteroidi e comete in orbita intorno al Sole, ma è raro che siano osservati corpi celesti provenienti da altrove, sia perché sono piccoli sia perché spesso rimangono a grande distanza da noi. Un telescopio in Cile, che fa parte di ATLAS, l’iniziativa della NASA per osservare i corpi celesti e capire se alcuni potrebbero scontrarsi con la Terra, aveva osservato per la prima volta un nuovo oggetto martedì 1 luglio, e inizialmente sembrava fosse un asteroide. Successive osservazioni avevano permesso di scoprire un’orbita insolita del corpo da poco osservato, portando alla conclusione che si trattasse con molta probabilità di qualcosa proveniente da un altro sistema solare.
Spesso sono necessarie alcune settimane prima di farsi un’idea dell’orbita di un oggetto di questo tipo, ma grazie a un astrofilo – cioè un appassionato che aveva controllato alcune osservazioni svolte da ATLAS a fine giugno – è stato possibile affinare da subito i calcoli, capendo meglio la traiettoria del nuovo corpo celeste. La notizia è circolata velocemente tra gli astronomi e gli appassionati di astronomia e negli ultimi giorni le decine di osservazioni fatte con vari telescopi in giro per il mondo hanno permesso di confermare, con un buon margine di certezza, che il misterioso oggetto è una cometa: cioè un corpo fatto per lo più di ghiaccio e di alcuni frammenti rocciosi (gli asteroidi sono invece per lo più rocciosi).
Mi sa che l’estate è proprio arrivata in via definitivamente ufficiale anche per me, perché stanotte ho veramente sofferto le pene dell’Inferno infuocato, da cui stamattina sono stata congendata più inzuppata di sudore liquido del solito… e quindi mi toccherà togliere il piumino. Si, il piumino a luglio, vietato giudicare. Questo è terribile, come tutti […]
notte infernalizzata tra caldo sudato e incubi dello stanco
Mi sa che l’estate è proprio arrivata in via definitivamente ufficiale anche per me, perché stanotte ho veramente sofferto le pene dell’Inferno infuocato, da cui stamattina sono stata congendata più inzuppata di sudore liquido del solito… e quindi mi toccherà togliere il piumino. Si, il piumino a luglio, vietato giudicare. Questo è terribile, come tutti gli anni, perché vorrà dire che il letto non avrà più quella solita classica pesantezza accogliente che bla bla bla, e io mi sentirò ihhhhh… Che grandi, gigantesche palle mosce! 😩
Ma, ovviamente, non è tutto così semplice; oltre ai danni, ci stanno le beffe. Avrò perso forse un’oretta di sonno complessivamente a causa della sofferenza, perché prima ci avrò messo almeno 3 quarti d’ora a prendere sonno definitivamente, poi mi sono svegliata credo una o due volte, di cui sicuramente l’ultima un’oretta prima della sveglia… che ne parliamo a fare, è banalmente il mio incubo peggiore che si avvera; magari marcisse soltanto il mio corpo nella pozza sottocoperta di acqua salata puzzolente ma lasciando in pace la mia anima… 🥀
Ah no, il mio incubo peggiore è avvenuto stanotte, ma è un altro. Senza approfondire troppo il contesto, ma ho sognato di essere stanca (nel mentre che lo ero davvero, visto che ho dormito come la cacca sciolta) tutta la mattinata fuori casa… quindi nel sogno mi sono chiesta come fosse possibile, e mi è venuto in mente: “ah, ero di fretta e ho dimenticato di prendere il caffè!” (cosa che, in un sogno non lucido, quindi percepito come reale, automaticamente mi ha pure peggiorato l’umore). Ora, sarebbe tanto simpatico sapere come mai il mio cervello ha generato quella come risposta al dubbio da lui stesso generato, visto che da un mese a questa parte la mattina sto bevendo sempre una quantità microscopica di caffè, visto che dormo abbastanza e rimango sempre a casa a marcire, e nel mentre (almeno consciamente, ma evidentemente solo lì) sapevo perfettamente di aver perso sonno… ma io giuro, in quei momenti la stanchezza l’ho sentita tutta. Potrei fottutamente cadere per sbaglio dal decimo piano in un sogno e non sentire alcun dolore, ma la stanchezza e la confusione mentale assolutamente si. E in effetti questa è la più perfettissima sintesi della stagione estiva. (Nonché degli effetti dei colpi di calore notturni sul cervello.) 🐡 pinterest.com/pin/759279437150… #caldo #dormire #estate #heat #incubo #night #nightmare #notte #sleep #sonno #summer #sweating
pinterest.com/pin/759279437150…
Snot Flickerman
in reply to themachinestops • • •I remember the days of high quality fansubs by people who were just doing it for free because they loved it.
Those still exist and still give you a better consumer experience than these fucking dogshit streaming services.
like this
slothbear, SaltySalamander, sujeito, TVA e yessikg like this.
BertramDitore
in reply to Snot Flickerman • • •Yeah I still rely on fansubs for all my jdramas that don’t get official releases outside Japan (which is still most of them). The quality is instantly obvious compared to machine-generated crap.
Cultural context and an understanding of tone and narrative coherence is critical for subtitles if you want to do justice to the quality of the writing in the original language.
metaStatic
in reply to Snot Flickerman • • •skisnow
in reply to Snot Flickerman • • •BudgetBandit
in reply to skisnow • • •Pika
in reply to themachinestops • • •like this
TVA e yessikg like this.
AntOnARant
in reply to themachinestops • • •SolOrion
in reply to AntOnARant • • •The issue is that they're almost certainly going to be cheap, and therefore companies are going to use them even if they're complete dogshit.
But I think the bigger issue is saying "no ai" and then... using AI.
TwiddleTwaddle
in reply to AntOnARant • • •like this
fif-t likes this.
plm00
in reply to TwiddleTwaddle • • •like this
TVA likes this.
daniskarma
in reply to TwiddleTwaddle • • •A good human translator is always the best solution.
But if the choice would be between crappy google translate or a LLM I would take the LLM translation.
There's no excuse for a big studio, they should hire translators. But for indie creators without a budget it can be the best way to get their creation to more people.
like this
TVA likes this.
weirdo_from_space
in reply to TwiddleTwaddle • • •inconel
in reply to AntOnARant • • •If you need translation for just getting facts and information for say math equation and its annotation translated, there's little margin in variety, what you need is database - that's mostly fine.
Pieces need translation are usually not like that. They have cultural context, pun, wordplay in rhymes, structural parallel, underlying tone, a lot of things only work in the language originally written. Translation is always a (nearly impossible) challenge for the translator to reconstruct all of them in target language.
I did game translation for a while. Translation is a field where AI hit first and honestly I've seen people lowering standards. The criteria of "good enough", "passable" is not the same compared to pre-AI days, and will keep changing. I'm almost sure this trend is happening in every industry the same way, and "just translation" is a slippery slope.
like this
missingno likes this.
samus12345
in reply to AntOnARant • • •skisnow
in reply to AntOnARant • • •I don’t know why it’s so difficult for these Ayn Rand readers to understand that almost no market is perfectly free, efficient, contestable or optimal. When it comes to licensing IP in particular, the overhead of trying to get permission for a show that’s already licensed by a competitor is an insurmountable barrier to entry for anyone who isn’t already a huge media conglomerate.
The only realistic alternative to shitty subtitles is piracy.
like this
TVA likes this.
richmondez
in reply to skisnow • • •Lvxferre [he/him]
in reply to AntOnARant • • •Because machine translations for any large chunk of text are consistently awful: they don't get references right, they often miss the point of the original utterance, they ignore cultural context, so goes on. It's like wiping your arse with an old sock - sure, you could do it in a pinch, but you definitively don't want to do it regularly!
::: spoiler Verbose example, using Portuguese to English
I'll give you an example, using PT→EN because I don't speak JP. Let's say Alice tells Bob "ma' tu é uma nota de três pila, né?" (literally: "bu[t] you're a three bucks bill, isn't it?") . A human translator will immediately notice a few things:
So depending on the context, the translator might translate this as "ain't ya full of shit...", or perhaps "wow, you're as fake as Monopoly money, arentcha?". Now, check how chatbots do it:
Both miss the mark. If you talk about three dollar bills in English, lots of people associate it with gay people, creating an association that simply does not exist in the original. The extremely informal and regional register is gone, as well as the accusatory tone.
With Claude shitting this pile of idiocy, that I had to screenshot because otherwise people wouldn't believe me:
[This is wrong on so many levels I don't... I don't even...]
This is what you get for AI translations between two IE languages in the same Sprachbund, that'll often do things in a similar way. It gets way worse for Japanese → English - because they're languages from different families, different cultures, that didn't historically interact that much. It's like the dumb shit above, multiplied by ten.
:::
That "business" is called watching pirated anime with fan subs, made by people who genuinely enjoy anime and want others to enjoy it too.
Luke
in reply to Lvxferre [he/him] • • •Deluge6062
in reply to Lvxferre [he/him] • • •That's a really awesome explanation.
Also, r/suddenlycaralho.
katy ✨
in reply to themachinestops • • •Lime Buzz (fae/she)
in reply to themachinestops • • •like this
sujeito e yessikg like this.
outhouseperilous
in reply to Lime Buzz (fae/she) • • •like this
yessikg likes this.
richmondez
in reply to Lime Buzz (fae/she) • • •Lime Buzz (fae/she)
in reply to richmondez • • •ramble81
in reply to themachinestops • • •like this
TVA likes this.
Hestia [she/her, fae/faer]
in reply to ramble81 • • •skisnow
in reply to ramble81 • • •like this
yessikg likes this.
dastanktal
in reply to themachinestops • • •like this
TVA likes this.
TriflingToad
in reply to themachinestops • • •like this
yessikg likes this.
Lvxferre [he/him]
in reply to themachinestops • • •Time to cancel my Crunchyroll subscription. Oh wait I don't have one, I simply torrent my series.
Seriously now. The anime fansubbing scene is one that makes me genuinely happy. It shows me there are plenty amateurs out there that are as good or better than plenty professionals like me.
like this
yessikg likes this.
MaggiWuerze
in reply to Lvxferre [he/him] • • •Grass
in reply to themachinestops • • •I was already bitter about patrasche being translated/localized(?) as patlash in re:zero but things are getting even shittier I guess. For those not in the know Dog of Flanders is a story nearly every japanese person will know at least to some degree, and has been referenced infinitely in anime and manga.
And similarly so many weird fucking names you could search the katakana online and find 30 different game wikis that reveal it was just akward katakana for a common non-japanese name, but no it has to to be different in this anime or that manga and we are going to make up some aneurysm romanization for it.
or incorrect subject assumtion because japanese is vague on that and if it's ai translated it actually makes sense with the lack of context awareness, and the list could go on for days
GeekFTW
in reply to themachinestops • • •They can want me to pay for it all they want, Crunchyroll has never and will never see a penny from my ass.
Shit, the last time fucking Funimation saw a penny from me was back when the DBZ orange box sets were on shelves, motherfuck Crunchroll lmao
richmondez
in reply to themachinestops • • •0range
in reply to richmondez • • •ipkpjersi
in reply to themachinestops • • •If anyone trusted Crunchyroll after they removed comments and reviews, it's honestly kind of their fault, as much as I hate victim blaming. They have shown who they are time and time again, it's not hard to believe them.
Though, I think anyone in this community very likely already knows exactly who they are.
themachinestops
in reply to ipkpjersi • • •ipkpjersi
in reply to themachinestops • • •shortcake
in reply to themachinestops • • •