Elsi Hj

@yapper_so Code Higgs sounds about right.

Use code "HIGGS" for 10% off your first credit pack. yapper.so/promos/credits

H*ggsfield tricks you into an overpriced plan with "unlimited" credits, then makes it near impossible to cancel when you realize you were fooled. They make money when you give up or forget about your subscription. Yapper gives you the same models for 40% less with NO SUBSCRIPTION. Make the switch & save!

@Diesol Problem is not will they watch (unless there is product placement in it) but will they pay to watch. Do we have proof that anyone paid a penny to watch an AI movie in a movie theather. That's something you guys should work on.

Proof people will indeed watch long-form narrative AI films. Go tell your story.

@VigoCreativeAI We are very much excited about the Chinese models in US as well, especially on how the Gov lately is taking over who gets to play with the new models and who is not.

Qwen3.8 announced: 2.4T params, open weights. As a Qwen Dev Ambassador, three things I actually care about: 1. Can it take over the meta-prompting stage of my pipeline? That slot is Opus today — if 3.8 holds up, my entire visual workflow gets a tier cheaper. 2. JSON output stability. My whole "fixed structure × variable elements" method rides on this. 3. Long-form copy quality — including Chinese, which Western evals almost never test. Max-Preview is live on Token Plan. Testing first. Full report soon. #Qwen

@Gossip_Goblin I laughed so hard at the gulp moment ;p

Another sneak peek.

@HendoAi Is this real audience going to pay just like they do for real movies?

I’m sick of watching the best AI films in the world on a screen the size of my hand. So I’m doing something about it. Today I’m launching WVLNGTH A new kind of film festival. A giant cinema screen. A packed room. A real audience. Brighton. 26 September. 🌊

@koltregaskes At least his name is Legit

Gemini 3.6 Flash is due for release near the end of this month. Remember, no Gemini 3.5 Pro for a while. h/t OGC More news like this at discord.gg/devmode

@higgsfield LMAO dunno why I find this so funny

Free Mode is live in Higgsfield Supercomputer. The LLM inside is now free to use for brainstorming, prompt writing, scripting, research, and project planning. Credits kick in only when you generate. A Supercomputer-level LLM, yours to use.

@venturetwins @DecartAI You can tell its not the person, which makes it unusable. It has to be at the very least at the same levels of realism as the video diffusion models. Kling has a great motion transfer model but it does not work in real time.

We're rapidly approaching the point where you can livestream as any character you want 🤯 Tested out the new Lucy 2.5 model from @DecartAI - latency and character consistency are impressive. It can also edit things in your surroundings or restyle the scene. How it works 👇

@AmirMushich @trymoda You could have just used a simple mask selection area for this one on Photoshop or even a color selection masking. I tried layerize with a more difficult task (presentation slides lol) and it gave me 2 layers in total ;p

This 12-sec video gives you more value than 10 hours of doomscrolling People are building careers using AI as their asset machine - in UI, web design, pitch decks for clients etc Please, get serious - dig into it.. My free layerize template: amirmushich.link/layerize Generate icons 🔽 x.com/AmirMushich/st…

@maxescu Did you try prompting supercomputer for lipsyncing the verses? Nowadays if it doesnt sing along it does not feel like a music clip (it reminds me of 90s hed candi summer house music clips.

CHERRY COLA I woke up this morning, and I prompted some music to get the groove on for the day. And then I generated this track. 6 hours later, I had a complete music video. Here's the full process, from start to finish 🧵🍒

@techhalla But I do not get it, did you do all of these real estate vids with image ref or not? Because i've noticed once you go from t2v to i2v it doesn't look nearly as good. What's the point of t2v when it comes to real estate? Selling a villa which does not exist?

SuReal Estate Seedance 2.0 prompts are below 👇 Raw vertical iPhone footage style, real mobile phone recording shot on iPhone 16 Pro, natural bright daylight on a private clifftop, strong sun, endless horizon views, handheld with natural shake and micro movements, authentic phone camera look, no cinematic emulation, no film grain, no photorealism, looks exactly like a real influencer video posted on Instagram, natural vibrant colors. [IMAGE REFERENCES] No reference images provided. Generate a fully consistent female influencer: mid 20s, attractive, energetic Dragon Ball fan vibe, tanned skin, long dark hair in a high ponytail, wearing a stylish casual-luxury outfit with subtle yellow and green accents, very expressive and excited when showing details. [TIMELINE SECOND BY SECOND] 0-3s: [Handheld phone shot, natural shake] The influencer walks toward a modern luxury villa built on a private clifftop with 360º views. The villa has a distinctive rounded yellow and white design. Parked outside is a small, quirky luxury car in bright yellow. She points at it smiling: "Look at this. They actually made the car real." 3-6s: [Quick cut to main living area] She enters the bright open-plan living space. The entire back wall is glass with insane views of the sea and horizon. She gestures at the minimalist but warm interior and says: "Everything is designed so you feel like you're floating above the world." 6-9s: [Quick cut to garden terrace] She walks out onto a large circular terrace with perfectly cut grass and a small garden area. There’s a big round stone table in the center. She runs her hand over it and says: "This is where you sit and watch the clouds go by for hours." 9-12s: [Quick cut inside special room] She enters a round meditation-style room with floor-to-ceiling curved windows. The light fills the entire space. She turns slowly and says with a smile: "This room… you can literally see the whole world from here. No exaggeration." 12-15s: [Quick cut to edge of terrace] She stands at the very edge of the terrace, wind in her hair, with nothing but open sky and sea behind her. She looks at the camera and says: "Private clifftop villa. Feels like your own little planet. Would you live here?" [STYLE & QUALITY BOOSTERS] Real iPhone 16 Pro footage look, natural bright daylight, authentic mobile camera movement and slight shake, natural vibrant colors, coherent physics, stable character, real phone video quality, no film look, no artifacts, looks like genuine Instagram Reel footage shot on location.

@venturetwins Does this imply that the lack of quality in gen AI so far was actually in the tool and not the creative direction?

If the public version of Seedance 2.5 is anywhere near this good, we're going to see a massive shift to AI generated video for almost every use case.

@maxescu It was very very soon even 3 weeks ago. I am sure there is some internal problems or they want some pre-hype to cover up or smth. Otherwise I cannot explain why Chinese model have 1+ months of prereleases while google releases stuff within the same day.

@eeelsi It's coming very, very soon. And it's fantastic!

This is it. The AI promised land.

Always amazed on how much people put confidence in demo reels which usually fall under the 5% outlier of all generations using the same prompt. I really wish SD 2.5 keeps at least half of its promises but the truth is we will only know when they release it. Anything before the release its only fluff. Prove me wrong.

@martini_film excellent

No more prompt roulette. Use Martini's Camera Motion for precise camera control over Seedance 2.0. Set keyframes for each composition and select the speed to create a controlled transition inside of a shot. Get precision beyond what text prompting can offer. #StepIntoSet #Martini #MadeWithMartini #CameraControl #AIFilmmaking

@maxescu Thats what we said since the release of Veo 2.0 and here we are. I do not think any substantial change will come up until version number 15 or so.

If Seedance 2.0 is this good, imagine how crazy Seedance 3.0 will be. Prompt below:

@BytePlusGlobal Is this the same guy who has pinned a vid how he is creating an agent from prompting, to scripting... same imagination is the only moat guy?

Every breakthrough starts with a creator pushing the boundaries of what's possible. It's inspiring to see REALITY brought to life with Seedance 2.0 4K. Excited for what creators imagine next.

@MayorKingAI Can AI be art? (since many tool artists I know, especially on Linkedin, are hating on anything AI)

Share your AI Art Have you made something amazing with AI? Images, videos, anything goes! Tomorrow (Sunday), I’ll feature the top creations in a special post Drop your best work below 👇

@IXITimmyIXI What's the purpose of this video. I am looking at it as a marketer and thinking, what does it mean? Is it just art for arts sake? Is it showing the capabilities of a gen AI model?

Day 831 #RunwayDailyChallenge Prompt Thankful 🙏 Have fun y'all