{"id":55104,"date":"2026-07-24T14:26:45","date_gmt":"2026-07-24T06:26:45","guid":{"rendered":"https:\/\/www.1ai.net\/?p=55104"},"modified":"2026-07-24T14:26:45","modified_gmt":"2026-07-24T06:26:45","slug":"%e4%b8%80%e4%b8%aa%e4%ba%ba%e5%81%9aai%e7%9f%ad%e5%89%a7%e5%8f%98%e7%8e%b0%e9%a1%b9%e7%9b%ae%ef%bc%8cai%e7%94%9f%e6%88%90%e7%9f%ad%e5%89%a7%e5%87%ba%e6%b5%b7%e4%b8%9c%e5%8d%97%e4%ba%9a","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/55104.html","title":{"rendered":"ONE PERSON IS DOING THE AI SHORTSHOW PROJECT, AI GENERATION + SHORTSHOWING TO SOUTHEAST ASIA"},"content":{"rendered":"<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-55105\" title=\"99f2202 dj00to1qr00jzd000v9000cfp\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/07\/99f2202dj00tio1qr00jzd000v900cfp.jpg\" alt=\"99f2202 dj00to1qr00jzd000v9000cfp\" width=\"1125\" height=\"447\" \/><\/p>\n<p>In the last two months, I was trying to see what the locals were doing, what they talked about, what they cared about, and I stopped by a picture that looked a little magical in China and looked at it several times\u3002<\/p>\n<p>In TikTok, a Thai netizens uploaded a whole series of Chinese Qingmiya dramas with Thai voice, painted like a pirated disc 10 years ago, with hundreds of people in the comment area pushing for the next episode. On the other side of the country, you know, some people in Vietnam put a national sweets show on YouTube, with a much higher volume than the original one in the country. Further down, Indonesia's information stream is full of one-minute inverted, dog-blooded stand-by dramas, and one can't stop\u3002<\/p>\n<p>I don't give a shit about \"this shit,\" and there's only one more thing in my head:<\/p>\n<p>IT'S A HUGE CONTENT DEPRESSION, AND NOW ONE PERSON CAN FILL IN WITH AI IN THE AFTERNOON\u3002<\/p>\n<p>I don't have to say anything about the domestic course. But thousands of kilometres south of the horizon, and 600 million people in South-East Asia are hungryly consuming what we see as \u201cunnourished\u201d: an old play with a voice, a dog-blooded short play, a hard-on stand. They also wanted to see things, but there was no one in the area to do them, and the Chinese content industry was just good at this\u3002<\/p>\n<p>The hardest part of this business in the past was \"porn.\" AI has already kept the threshold down. I'm going to break the whole link: what are the Southeast Asians looking at, what kind of \"AI\" tools, how to use ChatGPT to turn the play in their head into an \"AI-enabled\" directive, how a person finishes a set, what pits are localized, where the money comes from\u3002<\/p>\n<p>After that, you can do it\u3002<\/p>\n<p><strong>First, let's figure out what it looks like<\/strong><\/p>\n<p>Many people listen to the word \u201coff to sea\u201d and the first reaction is to do it better than at home. This is the first recognition to come from\u3002<\/p>\n<p>South-East Asia should not be considered a high ground, but rather a low ground. Here's the drama habit, probably ten years ago in China\u3002<\/p>\n<p>One of the publishers who had been out of the country for 20 years said one thing, and I think it's the best judgment of the whole market: Fashion films can be taken in every country, but only in China. South-East Asian viewers want to see modern urban dramas, they can do it in their own country, and they don't have to have you; the Qing Ming Palace, the Ming Dynasty, the Veterans, the Dynasty. Globally, only China has developed this set of industrial capabilities and culture. So it's been the first time in the country to get into South-East Asia, and it's been the first time in the country, and it's not bad for all generations\u3002<\/p>\n<p>The data is also going in this direction. In the first half of this year, the content of an international platform increased by the same amount of 1301 TP3T, the top of which was an ancient play, which was the highest Google search for heat in the world's two years. The interest of overseas users in antiques is extending from mere personality attachments to curiosity about historical background, traditional culture and the aesthetics of the East\u3002<\/p>\n<p>But the real increase is in the short play. In the first quarter of this year, there was an increase of 2381 TP3T in the global short play application, and it took an average of 37.73 minutes a day after 30 days for the user to load App. The market for the overseas micro-shorts broke out on a total scale of $4 billion in 2025 and is expected to burst to $5 billion in 2026. With 100 million people already living in the month and 1.8 billion short video users worldwide, there is still considerable space in the middle\u3002<\/p>\n<p>Southeast Asia is a strategic high ground on this board. The short play application downloads its number one in the world and Indonesia ranks second in the single country. The 600 million people, with an average age of 30.9 years, are much younger than in Europe and America and are born to eat fragments and strong emotions\u3002<\/p>\n<p>Call it a lowland, not a gold mine. The other side is brutal: Southeast Asian users are far less willing to pay than Europe and America\u3002<\/p>\n<p>Also a short play, App, South-East Asia has a global lead in downloads, but single downloads generate much lower revenues than mature markets in Europe and America. This determines your liquidity path: in North America, you can make money from a collection of pay-off locks by users, and in South-East Asia, it is more realistic to make money from ads, from big play and branding. This difference is repeated when it comes to liquidity, remember\u3002<\/p>\n<p>There's another layer of difference in taste, and it's not clear:<\/p>\n<ul>\n<li>Indonesia needs a very fast rhythm. The local video industry is underdeveloped, people grew up on YouTube with free short videos, used to emotional, conflict-intensive things, and you put him on a slow-paced mattress, and he cut it out\u3002<\/li>\n<li>Thailand eats old costumes, and it eats very much, but there is an absolute red line: can't touch the royal family, can't black out the Thai image\u3002<\/li>\n<li>Indonesia and Malaysia have a large Muslim population, with no direct access to the aesthetics, and a strict review of religious-related dramas, with close access\u3002<\/li>\n<\/ul>\n<p>South-East Asian viewers want to be straightforward: fast, cool, close to the locals. It doesn't matter if you take out the bottom of the box. It's the rhythm and the touch. Thinking about it, the power behind it will not be wasted\u3002<\/p>\n<p><strong>Two, the threshold fell in two years<\/strong><\/p>\n<p>The traditional reality shorts go out to sea and are stuck in three places: high cost, long cycle, and culturally difficult. You have to rent a place, find an actor, film, clip, play, subtitle, a play that hits hundreds of thousands, not necessarily a water and earth suit\u3002<\/p>\n<p><a href=\"https:\/\/www.1ai.net\/en\/tag\/ai%e7%9f%ad%e5%89%a7\" title=\"[View articles tagged with [AI skit]]\" target=\"_blank\" >AI skits<\/a>It's just the three walls\u3002<\/p>\n<p>This piece of cost, AI generates scripts, AI generates images, auto-syncs, and single-set costs can press up to one tenth of a reality play. Multilingual, a set of scripts is automatically translated with voice synthesis and can be available in English, Indonesian and Thai at the same time. This piece was distributed, and the same content was placed in TikTok, YouTube, and local media to measure the repercussions\u3002<\/p>\n<p>Someone's running through. There's a company that's doing a micro-short show, and it's only been in North America since May. One of them, an AI short play that combines Western supernatural elements like vampires and werewolves, broke 15 million on TikTok. They rely on the national production of AIGC, the big audio-visual model, which adds to the country ' s mature short play industrialization experience. In the words of their leaders, this road offers an opportunity to recreate the success of the game\u3002<\/p>\n<p>I'm going to pour cold water here, and one of the most important judgments:<\/p>\n<p>AI KNOCKED DOWN THE ENTRANCE THRESHOLD, AND THE REAL ACCESS TO THE SEA WAS ALMOST IMPOSSIBLE\u3002<\/p>\n<p>ONCE THE THRESHOLD COLLAPSES, EVERYONE WILL BE ABLE TO GET IN, AND THE INTERNET WILL SOON BE FLOODED WITH AI-GENERATED GARBAGE. ONE PERSON SPECIALIZING IN THE AI ANIMATION CHANNEL SAID IT OUT LOUD: THE INTERNET DOESN'T LACK AI ANIMATION, THE MISSING STORY IS INTERESTING\u3002<\/p>\n<p>Tools are available to everyone, and naturally they are not an advantage. What really opens the gap is your strategy and your ability to tell stories. All the tools, all the processes that follow, will be responsible only for getting you from the \"not done\" to the \"do done\" starting line. How far can you go after that\u3002<\/p>\n<p><strong>III. Tool selection: Don't be fooled by the list<\/strong><\/p>\n<p>NOW OPEN ANY OF THE TESTS, AND THE RANKING OF THE AI VIDEO TOOL CAN LIST A DOZEN SECTIONS AND SEE WHAT YOU SEE. I'LL CUT IT TO THE POINT WHERE YOU CAN MAKE DECISIONS\u3002<\/p>\n<p>Let's split up into two groups, one standard: want to go over the wall\u3002<\/p>\n<p>Four things that can be used directly in the country without going over the wall:<\/p>\n<ul>\n<li>CLEVER. IT'S THE ONLY ORIGINAL 4K, 60 FRAME, AND TUSHENG IS THE BEST VIDEO PLAYER, AND HE'S GOT A BIG MOVE, LIKE DANCING, FIGHTING, AND HE'S GOT A PRETTY GOOD MOUTH\u3002<\/li>\n<li>The dream \/ Seedance. It's the best thing to do in a series\u3002<\/li>\n<li>Mini Max. Cheapest, one video around $2.8\u3002<\/li>\n<li>Mansion. Move fast, catch the eastern aesthetic best-placed, do something Asian-style\u3002<\/li>\n<\/ul>\n<p>Four of the high-thresholds:<\/p>\n<ul>\n<li>Veo 3 (Google). The mouth and audio are the best in the field, producing 48 kHz of a professional-level synchronized audio, and the words are automatically heard when they are written directly into the hint. The disadvantage is that a single piece is only 8 seconds, and long content has to be kept\u3002<\/li>\n<li>Sora 2 (OpenAI). Physical simulations and spectroscopy editing are the strongest, but they are severe and scrutinized, and the community generally reflects the loss of quality after their release\u3002<\/li>\n<li>Runway, Pika. The former is the most powerful, the best filmer, and the latter is the strong\u3002<\/li>\n<\/ul>\n<p>The choice of tools depends on one thing: can it do the work you have? I'll give you a summary of my own:<\/p>\n<p>Why am I going to be first in a serial short play? The hardest thing to do is for character consistency, which can be used as a reference for nine maps, three videos, three audio segments, which can be used to keep the same character in camera, in a place where it's a part, and in an automatic lens. Cream 3.0 second, it produces a spectroscopy of up to six lenses at a time, with a free amount of 66 points per day for rapid validation of the rhythm\u3002<\/p>\n<p>Then there were two big pits, specifically for those who were prepared to throw money\u3002<\/p>\n<p><strong>First pit: the price is not equal to the real cost\u3002<\/strong><\/p>\n<p>All the tools will fail, they'll draw cards, a satisfactory shot you might have to generate several times. Counting the failure rate, the real cost of the material available per minute is as follows: the dream is about $6 to $8, the most cost-effective; the Vime 10 to $15; Sora goes straight to $22 to $30 because it's the toughest, and a perfect shot could be 10 times. So don't look at Sora. He's famous\u3002<\/p>\n<p><strong>The second pit: Don't chase new. It'll work\u3002<\/strong><\/p>\n<p>The number one on this track rotates by week. At the beginning of the year, the spirit managed to control the fire, and in March, a new version of the dream took away the wind, and then a new model came out and stepped on the first two. Every time a new king changes and learns, you're always on the line. The choice of a domestic, wall-free, adequate tool to eat it out is much less\u3002<\/p>\n<p>By the way, there's a risk to know: the conch is cheap, but it's only six seconds alone, and it's not enough to do the drama; and it's been followed by Disney and the Globe film industry and has initiated copyright litigation. Cheap is cheap, and it's done before commercial\u3002<\/p>\n<p><strong>IV. TRANSLATING THE PLAY IN YOUR HEAD INTO WHAT AI CAN UNDERSTAND<\/strong><\/p>\n<p>This section is the most appropriate time to look at it. Because the tools are so strong, what you feed it is wrong, it's just a piece of shit\u3002<\/p>\n<p>Let's start with a general perception that saves you months of detour:<\/p>\n<p>IT'S LIKE I'M A VERY EXECUTIVE PHOTOGRAPHER WHO DOESN'T HAVE ANY BRAINS. YOU'RE THE WRITER AND DIRECTOR\u3002<\/p>\n<p>YOU SAID TO HER, \"THE HOSTESS WALKED INTO THE ROOM AND FOUND A LETTER ON THE TABLE, AND SHE WAS SHOCKED, REMEMBERING THE PAST,\" AND PEOPLE COULD READ, AND AI WOULD BE BLIND. IT DID NOT KNOW WHAT THE ROOM LOOKED LIKE, HOW THE HOSTESS WENT, WHERE THE LETTER WAS, WHETHER THE SHOCK WAS STARING OR RETREATING AND WHETHER THE MEMORIES WERE GOING TO CUT THE PICTURE. IT'S GOOD AT SEEING THINGS. IT DOESN'T MAKE UP THE MOVIE IN YOUR HEAD\u3002<\/p>\n<p>SO THE ONLY CRITERION FOR A GOOD SCRIPT, A GOOD SPECULATION IS ONE: LET AI GUESS LESS\u3002<\/p>\n<p>How to land, in three steps, all can be done with ChatGPT\u3002<\/p>\n<p><strong>Step one: use ChatGPT to make a script skeleton<\/strong><\/p>\n<p>Give it three things first: a one-word premise (e.g. \u201cA woman receives a letter from a father who she thinks is dead\u201d), a recurring role, one or three scenes. Everything in the back comes from these three\u3002<\/p>\n<p>Then let it be written by the beat of the short play. One episode, 75 seconds or so<\/p>\n<ul>\n<li>0 to 10 seconds: hook. One image added a line that brought the audience into the situation\u3002<\/li>\n<li>10 to 55 seconds: upgrade. One thing changes, the bets keep rising\u3002<\/li>\n<li>55 to 75 seconds: Suspended. There's an iron law here that cuts off the \"discovery\" and never waits for the \"resolved\" and lets the audience get itch to the next episode\u3002<\/li>\n<\/ul>\n<p>AND THERE'S ANOTHER DETAIL: THE SCENE SWITCH IS AS CONTAINED AS POSSIBLE EVERY 20 SECONDS. EVERY TIME WE SWITCH, IT'S AN OPPORTUNITY FOR THE AI IMAGE TO DRIFT, THE MORE IT CUTS, THE MORE STEADY IT GETS\u3002<\/p>\n<p><strong>Step 2: Use a set of fixed Prompt to get ChatGPT out of the lens<\/strong><\/p>\n<p>This is the best spectroscopy I've ever seen, Prompt skeleton, which can be copied directly. Set up the ChatGPT make-up mirrors, set the table fields, and make it one step at a time:<\/p>\n<p>THE RULE OF \u201cONE STEP AT A TIME, ONE STEP AT A TIME\u201d IS CRUCIAL, NOT TO MENTION ITS SLOWNESS. ONCE IT'S DONE, IT'S LEAKING, IT'S DRIFTING, IT'S A STEP, YOU CAN FIX IT. THIS PROCESS HAS BEEN USED TO CRUSH THE SPECTROSCOPE, WHICH TOOK ONE MORNING AND FOUR HOURS TO COMPLETE, TO 10 MINUTES\u3002<\/p>\n<p><strong>STEP 3: WRITE THE SPECTROSCOPY AS AN AI\/ VISIBLE VIDEO TIP<\/strong><\/p>\n<p>To feed the video model, remember a formula:<\/p>\n<p>scene + person + action + lens + emotion\u3002<\/p>\n<p>Take an eight-second lens. Feel the density:<\/p>\n<blockquote>\n<ul>\n<li>On rainy nights, older residents are downstairs, with water and street lights on the ground. Twenty-five-year-old female, black windsuit, short hair, nervous. She stood at the entrance to the building and looked back to the corner, with a blurry background. Medium view, the camera slowly moves forward. She held the phone tight and hesitated to enter the building in two seconds. Cold-temperature, weak light, slight handheld. No cartoons, no sci-fi\u3002<\/li>\n<\/ul>\n<\/blockquote>\n<p>There are four of the most easy places to roll over and die early in the script:<\/p>\n<ol>\n<li>PEOPLE ARE CONSISTENT. THE SHORT PLAY IS THE MOST AFRAID OF THE MAIN ACTOR HAVING LONG HAIR AND SHORT HAIR. THE SOLUTION IS TO REPEAT THE KEY SET IN EACH SHOT: \"25 YEARS OLD, SHORT HAIR, BLACK WINDSUIT, SILVER WATCH IN THE LEFT HAND.\" FOR AI, IT NEEDS THIS STABLE ANCHORAGE TO REPEAT IT SEVERAL TIMES. BUT DON'T OVER-SPECIFY IT. IF YOU WRITE THE BONE STRUCTURE, THE PACE, IT'S A FACE CHANGE. PICK TWO TO THREE SIGNATURE PROPERTIES, ONE SIGN COSTUME PLUS ONE HAIR AND ONE ACCESSORIES\u3002<\/li>\n<li>Space direction. The first shot goes from the left to the right, the second shot goes from the right to the left, and the audience feels that the space is confused. Just write, \"Persons keep moving to the right from the left side of the picture\" -- that's the axial consciousness in the traditional lens\u3002<\/li>\n<li>Emotions don't just write adjectives. The word \"sorture\" and \"irror\" are too abstract. Replace it with the following: \u201cShe looked down at the paper, her lips shivered slightly and her eyes were red, but she did not cry\u201d. Emotions need to be presented through action\u3002<\/li>\n<li>Do not pile the camera language. Don't say, \"Motion sense, aerial shoot, close-up, slow shot, push and roll.\" One shot specifies only one major lens language, and the more concrete the more stable\u3002<\/li>\n<\/ol>\n<p>And one more technique I learned from cartoon clips: more specific action verbs. It's a hundred times better to write \"Jerry travels quickly and avoids seagrass\" than \"Jerry works hard to escape\". \u201cScrambling\u201d has a clear pattern of actions that can drive the picture directly, while \u201cworking\u201d is an empty word\u3002<\/p>\n<p><strong>Five, one person in one afternoon, how do you make a episode<\/strong><\/p>\n<p>The parts are in front of them, and this section links them to the currents. The goal is for you to know how to move your hand from zero to one\u3002<\/p>\n<p>Step 1, write the script. It's free of charge. Throw your story into it and give a spectroscopy in seconds. It's the source of everything behind it\u3002<\/p>\n<p>Step 2, do the lens table. Don't jump directly from the script. The middle watch is a bridge. Let the bean bag spell out the scenes, the movement of the person, the motion of the camera, the lines. The closer the lens, the higher the success rate of video generation. First, three to five anchor-point lenses with emotional weight (characterized feature of the place, confiding landscape, key reaction lenses) are planned, then the middle view is filled. Single shot controls in 3 to 6 seconds\u3002<\/p>\n<p>Step 3, lock the role and scene. It's the easiest part of the car to pull out and do it alone. A visual reference file was created for each of the main players, and all subsequent footage was generated on that file. For every scene that appears more than once, it also locks up one thing (a red brick wall, a brass lamp) and repeats it word for word in every hint of this scene. With three or four references, you have a database, each of which can be reused without additional costs\u3002<\/p>\n<p>Step 4, generate a lens. The locked camera forms are fed together with role, scene reference, and a reference image for each line\u3002<\/p>\n<p>Step 5, Tusheng video. Here are provincial money strategies: simple lenses (one person, fixed slots) using cheap models such as Vidu walk; complex lenses (many interactions, special effects) using a more stable model; and repeatedly failed bug lenses, cutting to dream Seedance 2.0, with its own \"director thinking\" and a high pass rate\u3002<\/p>\n<p>STEP 6, ACCENT AND ACCENT. IT'S THE VAST MAJORITY OF AI SHORT-TIME SHOWS OUT THERE. THE PREFERRED MODEL IS A SYNCHRONIZED EXPORT-TYPE MODEL THAT PRODUCES THE SAME VIDEO. IT'S NOT RIGHT. THERE'S THREE HIDDEN TRICKS: THE LENS CUTS EVEN TIGHTER (2 SECONDS OF THE CAMERA, SIX SECONDS OF THE PUDDLE); THE LENS IS DETACHED FROM THE MOUTH WHEN THE LINE IS LONG; THERE'S A SOUND AT THE BOTTOM OF EACH SPEECH, AND THERE'S A SENSE OF QUALITY BELOW THE SOUND, AND THERE'S NOT THAT IN THE EAR. AND THERE'S A HARD RULE: DON'T DO A MONOLOGUE SHOT FOR MORE THAN FIVE SECONDS, AND THE LONGER IT TAKES, THE MORE THE MOUTH DOESN'T SYNCHRONIZE AND THE MORE THE PROPORTION DRIFT\u3002<\/p>\n<p>Step 7, cut, color, subtitle. One principle: the same play, a set of mosaics and a light, without discussion. One light from the left, two light from the right, and the audience will think it's two places\u3002<\/p>\n<p>Step 8, export. The vertical platform is very condensed, so it has to be expanded, sampled, pressed 9:16, 1080 x 1920, or even higher, to hold the squeezing of the platform\u3002<\/p>\n<p>STEP 9, COVER. USE THE DREAM PHONE VERSION, SEARCH FOR SHORT PLAY FONTS, DO THE SAME TEXT, GO BACK TO THE BACKGROUND, AND WORK WITH THE BACKGROUND IMAGE CREATED BY AI, WITHOUT DESIGNING THE WORK\u3002<\/p>\n<p>The cost anchor gives you the concept of a one-minute set of short dramas, done in full, for about two to four hours, and the cost of tools can be up to a hundred dollars, some of the platforms in seconds, down to a dime a second\u3002<\/p>\n<p>If you don't want to spell so many tools yourself, there's already a do-stop platform on the market, where you can enter a script that goes all the way to a piece: it's got a byte clip, it's got hundreds of characters that don't fail, it's got an automatic collection that supports the import of 50,000 long scripts, it's got three hundred or so sound in it (in Cantonese dialect and dozens of foreign languages), and it's got a direct connection to the Acchi Express to the one square. The tools are always changing. Pick one to use. Don't get stuck in the selection\u3002<\/p>\n<p>AFTER ALL, DO HIGH-QUALITY CONTENT OR DO AI TRASH AS BAD AS YOU CAN KEEP IT TOGETHER. IT'S THE SAME ROLE, THE SAME SCENE, THE SAME LIGHT, THE TIGHTNESS OF THE CLIPS, AND IT'S THE FIRST SET OF AI CLIPS THAT LOOK LIKE A \"PEOPLE ARE GUIDING\" PLAY, AND THEY DON'T TURN INTO A MATERIAL PILE\u3002<\/p>\n<p><strong>VI. Localization is the real moat<\/strong><\/p>\n<p>You've come here to make something that you can see. But it's two things that can be made and accepted by the locals\u3002<\/p>\n<p>LET'S START WITH A JUDGMENT THAT MANY PEOPLE WILL IGNORE: IN THIS TRACK, THE PART THAT AI CAN DO FOR YOU IS VERY LOW. IMAGES, AUDIO, TRANSLATIONS, EVERYONE CAN DO IT QUICKLY. WHAT REALLY OPENS THE GAP IS THE LOCALIZATION OF THINGS THAT NEED PEOPLE TO BOARD\u3002<\/p>\n<p>LANGUAGE IS ONLY THE FIRST AND THE EASIEST. NOW AI TRANSLATES WITH SMART VOICE, A PLAY SYNCHRONISED WITH MORE THAN A DOZEN LANGUAGE SUBTITLES, SEVERAL LANGUAGE SOUNDS, AND HAS BEEN CONFIGURED. THERE IS A PLATFORM FOR AN ANCIENT PLAY WITH 14 LANGUAGE SUBTITLES AND 5 LANGUAGE SOUNDS. THIS FLOOR WILL BE DELIVERED TO AI\u3002<\/p>\n<p>THE SECOND LEVEL IS DIFFICULT: TRANSLATION ITSELF IS RECREATED, AND VERBATIM TRANSLATION IS FAR FROM SUFFICIENT. NAMES, MANNERS, CEREMONIALS, AND LAUGHTER MUST BE REORGANIZED IN LOCAL TERMS. TITLES, POSTERS, TRAILERS, ONLINE RHYTHMS ARE ALSO TO BE DESIGNED FOR REGIONALIZATION. A CUTE LINE IN CHINESE THAT TRANSLATES DIRECTLY INTO THAI MAY DIE, AND YOU HAVE TO LET SOMEONE WHO UNDERSTANDS LOCAL CULTURE THINK ABOUT IT AGAIN. AI CAN WRAP IT UP WORD FOR WORD, BUT IT HAS TO BE DECIDED, \u201cWHAT IS TO BE SAID ABOUT THIS?\u201d\u3002<\/p>\n<p>The third level is to review the red line and step on it is nothing. As mentioned earlier, this is a hard-on, because it is a hard-on: Thailand cannot touch the royal family, it cannot hack the image of Thailand; Indonesia, Malaysia and the United States are forbidden to delay, religious drama is thoroughly reviewed and close-in footage ends; the Middle East market is virtually impassable; Europe and America should be careful about ethnic issues\u3002<\/p>\n<p>The last floor is the one that is most easily ignored by the content, but that directly determines whether you can receive the money: pay for localization\u3002<\/p>\n<p>The payment ecology in South-East Asia is highly fragmented. Indonesians use DANA, GoPay, Malaysians use Touch'n Go, Thais rely on TrueMoney and PromptPay, and Vietnam has Momo, ZaloPay. Local electronic wallets already account for more than 40 per cent of electricity payments. If you're doing the best you can, and you end up with just a credit card or an international payment channel, you'll be pushing a large number of local users out of the door. A good content meets an unlocalized payment, which is nothing\u3002<\/p>\n<p>IT'S BEEN A LONG TIME SINCE WE'VE BEEN OUT OF THE SEA FOR MORE THAN THE CHEAPEST. YOU CAN'T BE ACCEPTED BY ANOTHER MARKET FOR A LONG TIME TO DECIDE HOW LONG BUSINESS CAN GO. AND THE \"ACCEPTED\" THING DEPENDS ON THE SLOW WORK THAT AI CAN'T REPLACE\u3002<\/p>\n<p><strong>Where did the money come from<\/strong><\/p>\n<p>It's finally time to get to the drop point: the realization\u3002<\/p>\n<p>First for a combination that has been validated: using YouTube to raise IP, using TikTok to convert. The algorithms and the liquidity logic of the two platforms are different and fit well. It is proposed to do 12 to 15 series of standardized content per month to spread the test\u3002<\/p>\n<p>TikTok is responsible for traffic conversion\u3002<\/p>\n<p>Its algorithms have several clear preferences: the first frame is expected to give a strong conflict or suspense within 3 seconds, and the system automatically gives weight to recommendations; vertical labels such as #ShortDrama can enter the exclusive flow pool; single-set hours are pressured in 90 seconds, with a completion rate exceeding 70% triggers a second recommendation. There is also a particularly friendly mechanism for going to sea: the content of your account number in South-East Asia can be synchronized with the markets in the Middle East and Latin America\u3002<\/p>\n<p>THERE ARE SEVERAL WAYS TO MAKE MONEY: THE CREATOR'S FUND, WHICH PLAYS ABOUT $100 TO $200 PER MILLION (WHICH REQUIRES MORE THAN A MILLION FANS); BRAND IMPLANTS, WHICH DIVIDE INTO 601 TP3T; NATURAL COMMODITY-LINKING MOTORS, COMMISSION 151 TP3T TO 251 TP3T; AND SHORT-CASTERS, WHICH ATTRACTS PEOPLE TO LIVE STUDIOS, WHERE ALL THE GIFTS ARE YOURS. THERE'S AN INTERESTING FIGURE: LIVE TRANSMISSION OF SUSPICIOUS SHORTS, 371 TP3T HIGHER THAN THAT OF SWEET PATS, AND THE CHOICE OF THE SUBJECT IS ON THE OTHER HAND\u3002<\/p>\n<p>YouTube is better suited to raise assets\u3002<\/p>\n<p>Shorts has three times the exposure weight of the short screen, and the series has a playlist that directs the audience to the main channel. The liquidity system is more complete: ads are divided into approximately US$ 0.8 to US$ 1.2 per 1,000 broadcast; channel members, paid subscriptions unlocked and creators receive 70%; super-appreciated reward, platform draws 30%; and high-quality plays with a volume of over 5 million are shown and may enter YouTube's procurement library. A good fee design: free pre-three, post-member-only five, end-of-the-life single-payment, step-by-step push-out\u3002<\/p>\n<p>HERE WE HAVE TO TAKE BACK THE JUDGMENT THAT I BURIED IN SECTION I: SOUTH-EAST ASIA IS LESS WILLING TO PAY, AND ADVERTISING (IAA MODEL) IS USUALLY MORE REALISTIC THAN PAYING SUBSCRIPTIONS. NORTH AMERICA CAN PAY FOR ROLLS, AND SOUTH-EAST ASIA EARNS ITS MONEY FROM LARGE-SCALE BROADCASTING, ADVERTISING AND BRAND COOPERATION. DON'T COUNT ON LOCAL USERS TO COLLECT MONEY. IT'S HARD TO GET THROUGH\u3002<\/p>\n<p>THIS IS HOW THE MATRIX WORKS: A MASTER ACCOUNT, A SUB-ACCOUNT, AND A SMALL NUMBER FOR THE A\/B TEST. THE TIME OF UPLOADING IS BETWEEN 8 AND 10 P.M., AND IT'S JUST IN TIME FOR THE PEAK. IT IS SUFFICIENT TO FOCUS ON THREE INDICATORS: THOUSANDS OF BROADCAST INCOME, AVERAGE INCOME OF PAID USERS, AND VALUE OF USER LIFE CYCLE\u3002<\/p>\n<p><strong>Eight, don't make a sign for the bonus<\/strong><\/p>\n<p>This last section is the one that would most like us to read it carefully, because all the efforts that have been made before it could have been planted on it\u3002<\/p>\n<p>The platform is actively cleaning up AI garbage. YouTube has a \u201ccommon or duplicate content\u201d policy that specifically targets content-based farm-based production. Don't think every video is different. So long as the whole is the same set, the same hook template, the same script structure, the same rhythm, the same cover only a few words, it will put you in the repeat, the light is limited, and the seal is the same\u3002<\/p>\n<p>IN CONTRAST TO THE TYPICAL PRACTICES OF BEING SENTENCED TO DEATH, YOU AVOID: CHANGING A TEMPLATE FOR A FEW WORDS IN BULK (\u201cCOLD KNOWLEDGE OF A GIVEN COUNTRY\u201d AND A SCRIPT FOR THE SAME NAME EACH DAY); RE-READING WIKIPEDIA WITH THE IMAGE OF THE MATERIAL LIBRARY AND THE AL CHORUS, WITHOUT ANY OF YOUR OWN VIEWS; REPEATING THINGS IN THE SAME FORMAT WITH NOTHING NEW; AND A PURE SLIDE WITH THE AI BYWORD FACT\u3002<\/p>\n<p>So what, AI? You can think of the auditing criteria very simply: a real person who looks at them and thinks, \"It's worth it\" or just says, \"It's trash again\"? As long as the creative direction is people-driven, each one has a unique value (your comments, your perspectives, your narratives), as well as a sense of visibility on the channel, and it's usually fine with AI. Remember value over volume, value greater than quantity\u3002<\/p>\n<p>THERE'S A GOOD THING ABOUT CARTOONS, ANIMATION CREATORS. MANY PEOPLE DON'T KNOW YET: CARTOONS, OBVIOUS CGIS, ANIMATIONS, NO NEED TO LABEL \"AI GENERATION.\" WHAT THE PLATFORM IS ASKING FOR IS A PICTURE OF WHAT MAY BE DECEPTIVELY SYNTHESIZED, FOR EXAMPLE, BY TELLING THE REAL PERSON THAT HE DID NOT SAY ANYTHING, OR TAMPERING WITH THE REAL EVENTS. YOU CAN SEE AT FIRST SIGHT THAT IT'S ANIMATED, IT'S A FICTIONAL SHORT PLAY, AND IT'S NOT USUALLY WITHIN THE SCOPE OF DISCLOSURE. IN TURN, THE WRITTEN CONTENT, WHICH COVERS SENSITIVE AREAS SUCH AS PUBLIC INFORMATION, HEALTH, FINANCE AND ELECTIONS, IS THE MOST CRITICAL AND SHOULD NOT BE TOUCHED UPON\u3002<\/p>\n<p>If it's true that it's written without a note, the price is higher: the first warning, the second suspension 90 days, the third permanent kickout. The disclosure was cheap, the recovery was expensive, and the bill was drawn\u3002<\/p>\n<p>Copyright is also a thunder. As mentioned earlier, the conch has been prosecuted by Disney and Globe, which reminds you of the source of copyright for training data when selecting tools. On music, no commercial copyrighted music from the platform, no royalties. As for copying the script structure, the cost of cross-border advocacy is extremely high and the success rate is extremely low, and you're likely to suffer from a lack of understanding\u3002<\/p>\n<p>This section, which is also an entire article, I would like to put it in a few words:<\/p>\n<p>THIS WAVE OF AI-RELATED DIVIDENDS, THE THRESHOLD OF WHICH IS SO LOW THAT IT IS ACCESSIBLE TO ALL, IS WHY AI ALONE DOES NOT CONSTITUTE ANY ADVANTAGE\u3002<\/p>\n<p>TECHNOLOGY CHANGES EVERY MONTH, AND GOOD STORIES CAN RUN FOR YEARS. THE INTERNET NEVER LACKS AN AI-GENERATED PICTURE, AND WHAT IS MISSING IS STORIES THAT PEOPLE WANT TO KEEP LOOKING AT. WHEN EVERYONE CAN PRODUCE A NICE SHOT, AND A PERSON CAN WRITE A 10-MINUTE SCRIPT, IT'S MORE SCARCE AND VALUABLE\u3002<\/p>\n<p>AI IS JUST THAT PEN, THE MAN WHO SHAKES THE PEN IS YOU\u3002<\/p>\n<p>THE REAL WINDOW PERIOD, I GUESS, SHOULD BE BETWEEN ONE AND A HALF AND TWO YEARS. AND WHEN THE LOCAL TEAM IN SOUTH-EAST ASIA REACTS, AND EVERYONE COMES IN, THE FLOOR RISES. NOW, THE OPPORTUNITY BELONGS TO THOSE WHO WANT TO DO IT TODAY, AND WHO WANT TO SAVE AI TIME, TO GRIND STORIES AND READ THE LAND\u3002<\/p>\n<p>The threshold has collapsed, and the scarce things have changed their place\u3002<\/p>","protected":false},"excerpt":{"rendered":"<p>In the last two months, I was trying to see what the locals were doing, what they talked about, what they cared about, and I stopped by a picture that looked a little magical in China and looked at it several times. In TikTok, a Thai netizens uploaded a whole series of Chinese Qingmiya dramas with Thai voice, painted like a pirated disc 10 years ago, with hundreds of people in the comment area pushing for the next episode. On the other side of the country, you know, some people in Vietnam put a national sweets show on YouTube, with a much higher volume than the original one in the country. Further down, Indonesia's information stream is full of one-minute inverted, dog-blooded stand-by dramas, and one can't stop. I don't give a shit about \"this stuff sucks,\" and there's only one more thing in my head: It's a huge content plume<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[144],"tags":[2268,4080],"collection":[],"class_list":["post-55104","post","type-post","status-publish","format-standard","hentry","category-baike","tag-ai"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/55104","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=55104"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/55104\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=55104"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=55104"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=55104"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=55104"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}