Creatorslop (@creatorslop) header
Creatorslop (@creatorslop) profile picture

The Morning Paper for AI Video Creators | Subscribe here!馃憞

Joined March 2026
Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)Photo by Creatorslop (@creatorslop)
Creatorslop retweeted
This AI video is absolute cinema 馃敟 A lone samurai under a blood-red cherry tree. 10 seconds, 1.4M views. ImagineArt (per the account that posted it)
馃 Made with AI
1
1
127
Creatorslop retweeted
This AI account's typical post gets 676K views. Not its best post. Its normal one. The format: clips that look like real phone or GoPro footage of moments that can't happen. You spend the first second working out whether it's real. Three accounts in our tracker run it. All three sit above 200K typical.
馃 Made with AI
1
1
2
123
Creatorslop retweeted
You can't call this AI slop anymore. One creator made a hand-drawn-looking gentleman, and his last two shorts did 72.5M and 87.8M views. The 10 posts before that: 200K to 3M. Same character, new city each episode. That's the format. Midjourney + Seedance (reportedly)
馃 Made with AI
1
1
144
This AI video is absolute cinema 馃敟 A lone samurai under a blood-red cherry tree. 10 seconds, 1.4M views. ImagineArt (per the account that posted it)
馃 Made with AI
1
1
127
馃帴 /imagineart.creators on Instagram, ImagineArt's own creator page. Their typical post gets about 5.6K views. This one did roughly 250x that.
18
Creatorslop retweeted
If I were starting an AI video page from zero this week, I wouldn't copy the account with the biggest viral hit. We tracked 35 AI video creators. Three big hits: 1. A 15.9M-view post. That account's typical post: 3.1K. 2. A 10M-view post. Typical: 8.4K. 3. A 4.6M-view post. Typical: 904K. Copy #3. Here's exactly what to do: - Sort accounts by typical views - Best post under ~10x typical = repeatable - Copy the format, not the clip
1
1
1
108
Creatorslop retweeted
If I were picking an AI video format from zero this week, I'd only pick one that works for more than one account. We tracked 35 AI video creators. Typical views per post, by format: 1. Animals acting human: 4.7M to 5.9M (2 accounts) 2. Fake "real" phone footage: 215K to 677K (3) 3. Pixar-style family dramas: 54K to 805K (4) 4. Stick figure stories: 8K to 240K (4) One account crushing it can be luck. Two or more is a format.
1
1
97
This AI account's typical post gets 676K views. Not its best post. Its normal one. The format: clips that look like real phone or GoPro footage of moments that can't happen. You spend the first second working out whether it's real. Three accounts in our tracker run it. All three sit above 200K typical.
馃 Made with AI
1
1
2
123
馃帴 The clip is from ifonly ai (676K typical; it credits Dreamina). The other two: exhumia ai (349K, says it uses Kling) and the worldai (215K, about 30 posts a week). All on Instagram. Typical = median views per post, from our creator tracker (Nov 10).
17
Creatorslop retweeted
Stop asking AI image editors to "make it bigger." 馃毇 We tested it. Left: our Nano Banana still. Right: the same image + "make the cat 3x bigger." Barely changed. What our tests taught us: 1. Editors keep the layout. Size is layout. 2. Regenerate instead, with size in the prompt. 3. Describe size against a person: "he only comes up to its shoulder" 4. Expect a trade-off: the biggest cats looked the most CGI. Pick scale first.
馃 Made with AI
1
1
127
If I were starting an AI video page from zero this week, I wouldn't copy the account with the biggest viral hit. We tracked 35 AI video creators. Three big hits: 1. A 15.9M-view post. That account's typical post: 3.1K. 2. A 10M-view post. Typical: 8.4K. 3. A 4.6M-view post. Typical: 904K. Copy #3. Here's exactly what to do: - Sort accounts by typical views - Best post under ~10x typical = repeatable - Copy the format, not the clip
1
1
1
108
Typical = median views per post over the last 30 days, from our creator tracker (Instagram, Nov 10). #3 is meowcloud ai on Instagram: 904K typical on under 2 posts a week. Every Friday we break down one format that holds up like this: creatorslop.com
44
Creatorslop retweeted
This AI video is absolute cinema 馃敟 Giants hauling stone blocks to build the pyramids, shot like a century-old film reel. 100% AI imagination, not history. Tools not stated.
馃 Made with AI
1
1
1
385
You can't call this AI slop anymore. One creator made a hand-drawn-looking gentleman, and his last two shorts did 72.5M and 87.8M views. The 10 posts before that: 200K to 3M. Same character, new city each episode. That's the format. Midjourney + Seedance (reportedly)
馃 Made with AI
1
1
144
馃帴 /antrofrog on Instagram. This is the New York episode. View counts from our creator tracker (Instagram, Nov 10).
37
Creatorslop retweeted
Posting more won't fix a weak AI video page. We tracked 35 AI video creator accounts. The ones posting 5+ times a week get the same typical views as the rest: about 46K a post. Here's exactly what to do instead: 1. Post 4-5 times a week. The top 2 accounts (5.9M and 4.7M typical) do. 2. One format, every post. 3. Judge it by your typical post, not your best one. 4. Turn stories into numbered parts. One series holds 240K an episode.
2
1
102
If I were picking an AI video format from zero this week, I'd only pick one that works for more than one account. We tracked 35 AI video creators. Typical views per post, by format: 1. Animals acting human: 4.7M to 5.9M (2 accounts) 2. Fake "real" phone footage: 215K to 677K (3) 3. Pixar-style family dramas: 54K to 805K (4) 4. Stick figure stories: 8K to 240K (4) One account crushing it can be luck. Two or more is a format.
1
1
97
Typical = median views per post over the last 30 days, from our tracker. Peaks lie: one stick figure account's best post got 25.1M. Its typical post gets 37K. This Friday we remake a Pixar-style family drama, prompts included: creatorslop.com
14
Creatorslop retweeted
Stop re-describing your image in image-to-video prompts. 馃毇 The model can already see it. We animated our giant-cat still with Kling 2.1, and what worked was describing only the motion: 1. The person's action: "grabs its fur and climbs up" 2. The reaction: "the cat calmly stands up and starts walking" 3. One camera line: "handheld phone camera" 4. End on: "realistic" 5. Make 3-4 takes. Ours smeared mid-climb.
馃 Made with AI
2
1
147
Stop asking AI image editors to "make it bigger." 馃毇 We tested it. Left: our Nano Banana still. Right: the same image + "make the cat 3x bigger." Barely changed. What our tests taught us: 1. Editors keep the layout. Size is layout. 2. Regenerate instead, with size in the prompt. 3. Describe size against a person: "he only comes up to its shoulder" 4. Expect a trade-off: the biggest cats looked the most CGI. Pick scale first.
馃 Made with AI
1
1
127
The size prompt that finally worked is in our giant-cat prompt post from Thursday (Flux Pro Ultra, raw mode). The whole test, failed tries included, is in the Creatorslop report: creatorslop.com
1
26