So I got to know that Amazon app store doesn't let you delete apps that you have published !! And now I can't even update my old web apps there !!! I have 7 apps in Amazon app store and I want to remove them because of poor developer experience of Amazon, but get this?!! They will not let you even remove apps. To get it published you have to dance to their drums!! To get it removed you have to dance to their drums!!! I HATE AMAZON !!!
14 Programming App Languages for Mobile App Development - A mobile application whether it is android or iOS, helps with the brand awareness in the market.
Duolingo is one of the best language learning apps. It provides users with modest exercises and takes a gamified approach. These exercises focus on summarizing words, sentences, and phrases from the language users are learning into their aboriginal language, & vice versa.
Do you wish to build an app like Duolingo, and do you want to hire a top mobile app development company in Los Angeles, then read our latest blog about How To Develop A Language Learning App Like Duolingo!
It can safely be said that the object has been the driving force in the programming industry for a very long time and will continue to be so for the foreseeable future.
YouTube will use AI to monitor your viewing habits to determine if you're under 18 and require an ID for you to view certain content, yes even if you're in the US.
Now. I have some questions. Like if I watch kids shows from the 90s is that a mark for me being over 30 because it's an old TV show? Or under 30 because it's a kid's show?
Okay question aside.
There's legitimate reasons to have "odd" viewing habits. I for one have chronic fatigue. I've fallen asleep with Youtube's autoplay on and will wake up 3 hours later with the weirdest shit playing (sometimes it's kids stuff).
I'm learning 2 languages. I will watch kids shows in those languages to help me practice and pick up the language. (Muzzy in Gondoland is SO CUTE). But it's really helpful to watch shows that are like "Red orange yellow green! One two three!" To learn the basics like your colors, numbers, days of the week.
Also. Bronies... adults like kids shows. Sometimes they got a really good message like MLP. Sometimes for Nostalgia, I mean Pokémon is still going, and I watch that for nostalgia.
"No, but AI would be smarter than that!"
Is it? Is it really? Bitch if kids want to watch YouTube, that's what YouTube Kids is for. We don't need your 1984 bullshit monitoring our fucking watching habits.
"I don't understand why you're so upset. Just give them your ID." Because I'm a developer. I watch programmers increasingly use AI to fucking program their goddamn software every fucking day. And the thing is AI is really fucking bad at it. It leads to issues like the Tea app leak.
The Tea app was an app where women warned other women in the dating scene about red flag men in the area and they could ask other women if the guy that were dating was a red flag. That app had women upload their Drivers Licenses, and they recently had a leak due to shitty security protocols caused by AI coding causing a legacy database to literally be open to the public, with so security features to keep any old random person from accessing it. All of the data in that database what just... dumped to the open internet. Location data. IDs. All of it.
So, no... I'm not gonna me uploading my ID anywhere BECAUSE I'VE SEEN HOW POORLY SOME OF YOUR ASSHOLES CODE. And by "how poorly you code" I mean you don't fucking code. You just give it to AI and the fucking dumbass system wouldn't know a security feature if it punched it in the face.
Suck my dick. You're not "keeping the kids safe" you're exposing everyone to identity threats.
If you don't mind me asking, how did you find the experience of migrating to Twine from ChoiceScript last year? I'm so curious about the whole process - both the initial migration and how you've found Twine to work with ever since. Do you feel it has affected your writing and bug testing? I always thought Twine was better for shorter less complex stuff, but you've proved me super wrong with SoH only getting bigger and better since the move!
Thanks for your question, I'm always happy to talk about things like this! I talk a bit about the actual logistics of the code in the #coding and #resources tags, but the gist of it is, ChoiceScript was like learning how to build a table. A table is more complicated than one might think--there are all kinds of joints and bolts and interlocking parts that one has to be aware of--but eventually one learns to master the table pretty completely. There's not much surprise to building a table, after a while, if you get the essentials down pat. But migrating to Twine was like taking that knowledge and using it to build a house: not just the scaffolding and structure of that house, but also the exterior paint and landscaping and roofing and cosmetic details, and the interior pipes and systems and wiring, and everything else in between. You not only have to know the principles of building a house, you also have to know, say, how electricity works, and how to install a water heater, and what kind of paint won't mess with the HOA regulations because it's an eyesore... That felt complicated enough: it took 3 years from the conception of the idea of moving to Twine to actually getting the Twine alpha build version up (though I think I could have managed it in less time if I'd only been working on the port and not updating the ChoiceScript version with new content simultaneously). But I can't even fathom how people who work with more complex engines (like Unity or Unreal) do it. They deserve all the flowers!
So far, I haven't found that Twine has limited me in any way when it comes to writing a huge, long project. I do think it has been historically used for shorter projects, but I think that largely applies to its app interface, with all of the little boxes and visual branches. For example, it takes about 429 "boxes" to get to meeting Hal in Chapter 1:
...so that wouldn't have been feasible if I'd kept going with that method. I shudder to think how many boxes that would have been! But the workflow I use is text-based, using VSCode, and very similar to ChoiceScript's CSIDE, and I haven't found any limitations with doing that that so far!
If you're a complete beginner to interactive fiction, I still recommend starting with ChoiceScript first. It's a narrow enough scope--with good documentation and a great community--that it's easier to learn the foundational principles of interactive fiction and game development through that lens. Twine is so flexible and powerful that it's almost a blank-canvas--you often have to figure out things and build from scratch--which allows for more customization, but is also a much steeper learning curve. Like building a house without a blueprint! Just go nuts and see what works!
I would say the most difficult aspects about the migration to Twine was the lack of a really concentrated community (at least in terms of how concentrated and easy-to-find ChoiceScript's is), a bit scattered documentation (there are a variety of versions, language names, and story formats for Twine, so it's hard to find a definitive "START HERE" point with basic 101 documentation), and so far, I'm finding the bugtesting much less robust than CSIDE's. Not only am I using a lot of my own custom language/programming syntax, but I don't think there's an automated way to test for bugs with the way I'm doing it?? In ChoiceScript you could literally press a button and have the program run through your game and test random choices for you 1,000 times, 10,000 times, whatever. It would automatically throw a bug with the exact line number if it found any, which made bugtesting a breeze. With Twine, I've been bugtesting everything manually, relying on player reports as well as my own playthroughs... 🥹 So that is certainly a consideration!
But otherwise, I've managed to get my Twine system very similar to my ChoiceScript workflow, so it hasn't affected my writing at all!
Thanks for your question! If I didn't quite answer anything well enough, feel free to ask any follow-ups!
Ellipsus is a freemium writing and collaboration tool for writers, offering both free and paid versions. As a browser-based text editor, it has a distinct focus on anti-AI offerings and web-based publishing and collaboration. Users can use themes to change the colours and fonts displayed by the interface, and can easily export or copy text in multiple formats for online publishing. Fan writers will find the AO3 export tool especially useful. However, limitations to collaboration and tools for organising longer works make it a tough sell for power users. A lack of automated backups, reported server connection and sync issues, and unclear terms for allowed/prohibited content make for a murky user experience.
As a writer and poet, I think that the most important part of writing is the actual writing. Any platform that works for you? That’s the right one, and my first and foremost advice is to always back up your work, no matter what software you use.
For writers looking at trying out something new, I’ve taken a look at what’s going on with Ellipsus.
One of the web app’s biggest selling points? It’s easy to get started with and use for basic writing, and it’s dead simple to switch between devices with.
Pros and Cons
Positives
The company’s anti-AI ethos is attractive to authors concerned about scraping of private and publicly exposed documents for large language model (LLM) training.
Headquartered in Berlin, Germany, Ellipsus GmbH is a European company with European users, and has laid out its privacy policy to showcase its compliance with European GDPR and German data protection law, including an explicit promise to “never use your writing to train AI models.” Its hosting with Amazon Web Services is also provided through EU regions (see “Service providers we work with“).
According to the terms of service, a user’s “content” within the service refers not only to the user’s files, documents, and drafts, but also to comments and chat messages. Users explicitly own that content. Ellipsus explicitly promises not to use any of it for “training or developing machine learning or AI models of any kind” and this is echoed in the privacy policy.
“Snippets” sharing tool is a hit with writers who like to share quotes from their manuscripts and works-in-progress with an external audience, showcasing customisable, social-media-friendly designs (with additional customisation in the Plus version).
Theme customisation: users enjoy a similar level of customisation with the interface, mostly involving adjustments to to colours and fonts. Dark and light themes—among others—are available in the free version, and custom themes are available in the Plus plan.
Convenient export functionality allows users to download any individual document as a PDF, DocX, or Markdown file; or to copy it as HTML, Markdown, or Rich Text. Fan writers may appreciate the export functionality that allows linking an account directly to Archive of Our Own.
Browser-based text editor offers ease in switching between devices, making it a popular alternative to writing in Google Docs (also: unlike in Ellipsus, GDocs’s scroll bar makes me lose my place every time my thumb brushes the side of the screen). As a writing app, Ellipsus boasts a low barrier-to-entry compared to more privacy-focused tools that focus on encryption. Personally, I also enjoy writing with Markdown shortcuts enabled in Ellipsus, which allow for quick formatting when typing on mobile.
“Writing Insights” (Plus plan) tool uses “simple, programmed algorithms that run locally on your device”—the ability to highlight frequently used words and passive voice can be a useful tool while editing. These simple metrics are similar to word count or the “find and replace” tool in complexity; there’s zero indication that they use any sort of “AI” or LLM-based processing, and Ellipsus confirms that the feature does not use AI.
Queer-founded, with support from Gaingels Syndicate, a US-based venture capital firm that aims to support LGBTQ-friendly tech investment, as well as diversity and inclusion in hiring.
Introduction of a paid plan (“Ellipsus Plus“) indicates the potential for a sustainable business plan, and the “yours to keep” model provides a lifetime licence after 24 months of payments (currently $6 USD per month, with a final price of $144 USD). No credit card is required for the free 7-day trial.
Drawbacks
note: numbers in parenthesis are footnotes; see the end of the post for the notes!
No desktop version or offline version. Currently, the only way to access Ellipsus is through a browser on your computer or mobile device.
Lack of automated backups—or any simple way to regularly export all of the files in a user’s account—makes Ellipsus concerning as a primary place to write and store documents. With any platform, loss of data is always a possibility: always back up your work! (1)
Feature limitations:
Editing: A lack of track changes functionality (also known as selective merging, suggestions, or inline track changes) means that editors or beta readers need to leave comments on the work instead of being able to provide individual suggestions at a granular level. I also find it clunky to manage editing in the platform; writers need to create a separate “draft” to allow comments. Drafts can be merged, but doing so isn’t as intuitive or specific as track changes.
Writing: This becomes especially confusing when drafts are also used as a workaround for a lack of chapter/tab functionality, and a common criticism online is that the platform lags on documents with longer word counts. There is no support for tables, and nested lists often break upon import/export, even in Markdown. (2)
Collaboration: Commenters need their own Ellipsus accounts, so depending on how common Ellipsus already is in your own writing circles, it might not be the easiest platform to onboard beta readers to.
Version history: only 30 days are included with the free plan.
No encryption for documents, which might otherwise alleviate concerns about allowed content.
Security concerns: no multi-factor authentication is currently available for logging in, despite Ellipsus now accepting payments. (3)
Bug reports online include issues with syncing and loss of work (always click the sync button to ensure your work syncs), server connection issues, random movement through the document, keys sticking, and unexpected formatting changes to paragraph spacing and fonts. A ticketing system is available, but the best way to reach support is through the Discord server, which is very unfortunately very common with modern apps.
Unclear terms for allowed content: as a young platform, Ellipsus is largely untested when it comes to policing of user content within the platform. The terms of service assure that content written “privately within Ellipsus” isn’t policed, but some content is against the rules, namely plagiarism and planning/committing crimes (4), and that prohibited content is subject to reporting from readers who receive access to the document, as well as legal requests (keeping in mind Ellipsus’s European jurisdiction). Ellipsus also recognises the community’s concerns that the payment processor Stripe might “pressure” Ellipsus over “adult or NSFW content” and has responded that in the case of a “serious conflict with our values, user privacy, or the kinds of (lawful) writing our users create, we would evaluate alternatives. If necessary, we would migrate payment providers.” (5)(6) Please remember that I am not a lawyer, and that this summary is not legal advice.
Abrupt changes are always possible with any SaaS (software as a service), especially to the free version. Growth of the platform is imperative for a company with Ellipsus’s funding structure, based on both its venture capital backing (7) and its lifetime licence structure. It will require additional users to continue bringing in cash flow, and as with any company, it could shift focus from its current niche of fan writers and toward a larger potential userbase.
The opt-in tool Emboss, available as part of the Plus plan, appears to have been planned by the company from early on (8). The purpose of this tool is to provide evidence of human authorship by combining a number of metrics with the ability to provide a public link to past versions of the document (“selected version history snapshots“). While the tool doesn’t provide any kind of score or assessment as to whether the document was or was not human authored, I personally don’t think it was well-executed, and I was concerned when Ellipsus seemed to be suggesting that sharing Emboss links might provide a writer with a public shield against accusations of AI use. While this tool might be able to provide evidence of the evolution of a work to an editor or publisher through version history, it doesn’t seem to provide any more value than other tools with version control. (9)
Footnotes
The “Right to portability” section under the privacy policy specifies that users can request a “machine-readable export” which includes your writing.
One of the co-founders addressed the lack of some of these features in a reply to a document posted by both co-founders in June 2026.
The company updated the privacy policy by June 10, 2026 to indicate that full credit card numbers were not being stored by Ellipsus, but were handled by the payment processor. The co-founder responded to a user question on June 13, 2026 saying that multi-factor authentication/two-factor authentication would move to a planned feature, and this has been added as a “big priority” to the June 2026 document (item 35).
Item 28 in the June 2026 document elaborates that “NSFW content is not illegal” and “is permitted in Ellipsus” while users should look to the terms of service “and refer to your local laws for clarification.” The terms prohibit “Us[ing] our services to do anything illegal: Write about fictional or historical crime as much as you want—please don’t use our services to plan or commit any IRL illegal activity.” In the June 2026 document, illegal content is defined as subject to the user’s local law; Ellipsus claims that they “do not actively look at or police your documents” but elaborates that users sharing their documents do so at their own risk. “We don’t actively police public work, unless it’s reported to us/we receive a valid legal or DMCA request. (If so, we will handle that according to our policies and legal obligations).”
Item 26 in the June 2026 document.
Users are also concerned because of the reality that classification systems often include all queer content within the label “adult” and that “Several regions make it illegal to discuss, promote, or teach LGBTQ+ topics.”
Including Better Angle, Circle & Co Ventures, Common Magic, SpeedInvest, and Gaingels.
An argument for the Emboss human attribution tool is laid out in some early blog posts: The Watchmakers (2023) and The Distributors.
As with the writing metrics, there is no reason to think that the metrics provided through Emboss use any sort of “AI” or LLM-based processing, and Ellipsus confirms that “No AI is used in the compiling of these metrics.”
Language Apps Suck, Now What?: A Guide to Actually Becoming "Fluent"
The much requested sequel to my DL post that was promised almost a year ago.
I'm going to address all of the techniques that have helped me in my language learning journeys. Since 95% of these came from the fact that in a past language learning mistake, they are titled as my mistakes (and how I would/did things differently going forward). For those that read to the bottom there is a "best universal resources" list.
Disclaimers:
"Fluency" is hard to define and everyone has their own goals. So for the purpose of this post, "fluency" will be defined as "your personal mastery target of the language".
If you just want to pick up a bit of a language to not sound like a total foreigner on vacation or just exchange a few words in a friend's native language, feel free to ignore what doesn't apply, but maybe something here could help make it a little easier.
This is based on my own personal experience and (some) research.
Mistake 1: Asymmetrical Studying
Assuming you don't just want to do a single activity in a language, or are learning a language like ASL, a language requires 4 parts to be studied: Speaking, Listening, Writing, Reading. While these have overlap, you can't learn speaking from reading, or even learn speaking from just listening. One of my first Chinese teachers told me how he would listen to the textbook dialogues while he was biking to classes and it helped him. I took this information, thought "Yeah that's an idea, but sounds boring" and now regret not taking his advice nearly every day.
I think a lot of us find methods we enjoy to study (mine was reading) and assume that if we just do that method more ™ it will eventually help us in other areas (sometimes it does, but that's only sometimes). Find a method that works for you for each area of study, even better find more than one method since we use these skills in a variety of manners! I can understand a TV program pretty well since I have a lot of context clues and body language to fill in any gaps of understanding, but taking a phone call is much harder—the audio is rougher, there's no body language to read, and since most Chinese programs have hard coded subtitles, no subtitles to fall back on either. If I were to compare the number of hours I spent reading in Chinese to (actively) training my listening? Probably a ratio of 100 to 1. When I started to learn Korean, the first thing I did was find a variety of listening resources for my level.
Fix: Find a variety of study methods that challenge all aspects of the language in different ways.
A variety of methods will help you develop a more well-rounded level of mastery, and probably help you keep from getting bored. Which is important because...
Mistake 2: Inconsistent Studying
If there is one positive to a language app, it is the pressure it puts on keeping a streak. Making studying a part of your everyday routine is the best thing you can do. I benefited a lot from taking a college language course since I had a dedicated time to study and practice Chinese 5 days out of the week (and homework usually filled the other two). Memorization is a huge part of language learning, and stopping and starting is terrible for memorization. When I was in elementary school, we had Spanish maybe a couple times a month. Looking back, it seems like it was the first class to be cut if we needed to catch up on a more important course. Needless to say, I can't even speak Spanish at an elementary level.
However, I'm sure many people reading this don't have the time to do ultra-immersion 4-hour study sessions every day either. Find what days during the week you have time to focus on learning new vocab and grammar, and use the rest of the week to review. This can be done on your commute to school/work, while you do the dishes, or as a part of your morning/evening routine. Making this as realistic as possible will help you actually succeed in making this a habit. (Check this out for how to set realistic study goals)
Fix: Study regularly (ideally daily) by setting realistic goals. Avoid "binge" studying since remembering requires consistent repetition to be most effective.
Mistake 3: Resource Choice
This is really composed of two mistakes, but I have a good example that will cover them both.
First, finding resources that are at or slightly above your level is the most important thing. Easy resources will not challenge you enough and difficult resources will overwhelm you. The ideal is n+1, with n as what you know plus 1 new thing.
Second, getting distracted by fancy, new technology. Newer isn't always better, and there are often advantages that are lost when we've made technological developments. I often found myself wanting to try out new browser extensions or organizational methods and honestly I would've benefitted from just using that time to study. (Also, you're probably reading this because of my DL post so I don't think it has to be said that AI resources suck.)
A good example of this was my time using Clozemaster. I had actually recommended it when I first started using it since I thought the foundation was really solid. However, after long term use, I found that it just wasn't a good fit. The sentences were often too simple or too long and strange for memorization at higher levels or were too difficult at lower levels. I think that taking my textbook's example sentences from dialogues into something like Anki would've been a far better use of my time (and money) as they were already designed to be at that n+1 level.
Fix: "Vet" your resources—make sure they will actually help you. If something is working for you, then keep using it! You don't always have to upgrade to the newest tool/method.
Mistake 3.5: Classrooms and Textbooks
A .5 since it's not my mistake, but an addendum of caution. I think there is a significant part of the language learning community that views textbooks and classroom learning as the worst possible resource. They are "boring", "outdated", and "ineffective" (ironically one of the most interesting modern language learning methods, ALG, is only done in a classroom setting). Classrooms and textbooks bring back memories of being surrounded by mostly uninterested classmates, minimal priority, and a focus on grades rather than personal achievement (imagine the difference between a class of middle schoolers who were forced to choose a foreign language vs. adult learners who self-selected!) People have used these exact methods, or even "cruder" methods, to successfully learn a language. It all comes down to what works best for you. I specifically recommend textbooks for learning grammar and the plentiful number of dialogues and written passages that can function great as graded readers and listening resources. (Also the distinction made between "a youtube lesson on a grammatical principle" which is totally cool, and "a passage in a grammar textbook" is more one of tone and audio/written than efficacy).
Classrooms can be really great for speaking practice since they can be a lot less intimidating speaking to someone who is also learning while receiving corrections. Speech can be awkward to train on your own (not impossible if you're good at just talking aloud to yourself!), and classrooms can work nicely for this. Homework and class schedules also have built in accountability!
Fix: Explore resources available to you and try to think holistically about your approach. CI+Traditional Methods is my go to "Learning Cocktail"
Mistake 4: Yes, Immersion, But...
I realized this relatively quickly while learning Chinese, but immersion at a level much higher than your current level will do very little for you. What is sometimes left out of those "Just watch anime to learn Japanese" discussions is that you first need to have a chance at understanding what is being said. Choosing materials that are much higher than your level will not teach you the language. It doesn't matter how many times someone at HSK 1 hears “他是甘露之惠，我并无此水可还”, they will not get very far. Actual deduction and learning comes from having enough familiar components to be able to make deductions—something different than guessing. An HSK 1 learner, never having heard the word 老虎 will be able to understand "tiger" if someone says “这是我的老虎” while standing next to a tiger. This is not to say you can never try something more difficult—things should be challenging—but if you can't make heads or tails of what's being said, then it's time to find something a bit easier. If mistake 2 is about the type of method, this is about the level. If you wouldn't give a kindergartener The Great Gatsby to learn how to read, why would you watch Full Metal Alchemist to start learning a language?
Side note: Interesting video here on the Comprehensible Input hypothesis and how it relates to neurodivergence.
Fix: Immerse yourself in appropriate content for your level. It's called comprehensible input for a reason.
Mistake 5: On Translation
I work as a translator, so do you really think I'm going to say translation is all bad? Of course not. It's a separate skill that can be added on to the basic skills, but is really only required if you are A. someone who is an intermediary between two languages (say you have to translate for a spouse or family member) or B. It is your job/hobby. In the context of sitting down and learning, it can be harmful. I think my brain often goes to translation too often because that's how I used to learn. Trying to unlearn that is difficult because, well, what do people even mean when they say "don't translate"? They mean when someone says "thank you", you should not go to your primary language and translate "you're welcome" from that. You should train yourself to go to your target language first when you hear the word for "thank you". A very literally translated "thank you" in Chinese "谢谢你" can come off as cold and sarcastic. I don't tell my friends that, I say "谢啦～". Direct translation can take away the difference in culture, grammar, and politeness in a language. If there is a reason you sound awkward while writing and speaking, it's probably because you're imposing your primary language on your target language.
Fix: Try as hard as you can to not work from your primary language into the target language, but to work from the structures, set phrases, and grammar within the target language that you know first.
Mistake 6: The Secret Language Learners Don't Want You To Know...
...is that there is no one easy method. You are not going to learn French while you sleep, or master Korean by doing this one easy trick. Learning a language requires work and dedication, the people that succeed are those that push through the boredom of repetition and failure. The "I learned X in 1 year/month/week/day!" crowd is hiding large asterisks, be it their actual level, the assistance and free time available to them, "well actually I had already studied this for 4 years", or just straight-up lying. Our own journeys in our native tongue were not easy, they required years and years of constant immersion and instruction. While we are now older and wiser people that can make quick connections, we are also burdened with things like "jobs", "house work", "school work", and the digital black hole that is "social media" that take up our time and energy. Everything above is to help make this journey a little bit easier, quicker, and painless, but it will never be magic.
I find that language learning has a lot in common with the fitness community. People will talk about the workout that changed their life and how no other one will do the same—and it really can be the truth that it changed their life and that they feel it is the ultimate way. The real workout that will change your life is the one you're most consistent with, that you enjoy the most. Language learning is just trying to find the brain exercise that you can be the most consistent with.
Fix: Save your energy looking for shortcuts, and do the work, fail, and come back for more. If someone tells you that you can become fluent in a ridiculously short amount of time, they are selling you a fantasy (and likely a product). You get out what you put in.
For those that made it to the end, here are some of my "universal resources":
Refold Method: I don't agree with their actual method 100%, but they've collected a lot of great resources for learning languages. I've found their Chinese and Korean discords to also be really helpful and provided even more resources than what's given in their starter guides.
Language Reactor: Very useful, and have recently added podcasts as a material! The free version is honestly all you need.
Anki: If I do not mention it, the people with 4+ year streaks with a 5K word deck will not let me forget it. It can be used on desktop or on your phone as an app. If you need a replacement for a language learning app, this is one of them. Justin Sung has a lot of great info on how to best utilize Anki (as does Refold). It's not my favorite, but it could be yours!
LingQ: "But I thought you said language apps are bad!" In isolation, yes. Sorry for the clickbait. This one is pretty good, and more interested in immersing you in the language than selling a subscription to allow you to freeze your streak so the number goes up.
Grammar Textbooks: For self-taught learning, these are going to be the best resource since it's focused on the hardest part of the language, and only that. If you're tired of seeing group work activities, look for a textbook that is just on grammar (Modern Mandarin Chinese Grammar is my rec for Chinese, and A Guide to Japanese Grammar by Tae Kim is the most common/enthusiastic rec I've heard for Japanese).
Shadowing: Simply repeat what you hear. Matt vs Japan talks about his setup here for optimized shadowing (which you can probably build for a lot cheaper now), but it can also just be you watching a video and pausing to repeat after each sentence or near simultaneously if you're able.
Youtube: Be it "Short Story for Beginners", "How to use X", "250 Essential Phrases", or a GRWM in your target language, Youtube is the best. Sometimes you have to dig to find what works for you, but I imagine there is something for everyone at every level. (Pro tip: People upload textbook audio dialogues often, you don't even have to buy the textbook to be able to learn from it!)
A Friend: Be it a fellow learner, or someone who has already mastered the language, it is easier when you have someone, not only to speak to, but to remind you why you're doing this. I write far more in Chinese because I have friends I can text in Chinese.
Pen and Paper: Study after study, writing on paper continues to be the best method for memorization. Typing or using a pen and tablet still can't compare to traditional methods.
The Replies (Probably): Lots of people were happy to give alternatives for specific languages in the replies of my DL post. The community here is pretty active, so if this post blows up at least 20% of what the last one did, you might be able to find some great stuff in the replies and reblogs.
