Two separate finds landed today that work the same way. Draw a red circle on a reference image and the video happens there. Put a red dot where you want the eyes to look and they look there. Our own best result last week was the same idea in a different channel, and it explains why so much prompt wording does nothing.
Everyone plans their VRAM. Almost nobody plans their storage, and it fails in ways that do not look like storage. A note on the hidden running cost of local AI, including the 591 GB folder we did not know we had and the broken login it caused.
Twenty generators, eight prompts with invented specifics, one shot each, the prompt on every board. New this round: five Krea 2 builds, Qwen Image 2512, a real cloud column priced in Buzz, and a text ladder scored over four extra seeds. Every board at full size, every prompt in all three styles.
We ran MiniMax H3 for a full day on a single RTX 5090: talking shots driven by our own voice track, a walk and talk, text to speech, a trained voice, and an eight scene short film with its own soundtrack. Every number here came off our machine. So did the mistakes.
If your LTXVideo nodes stopped loading after updating ComfyUI, you did not break anything. A dependency called kornia gets pulled forward to 0.8.3 and the video node pack cannot import against it. The fix is one command, the issue has been open since June, and there is a real trade-off nobody mentions.
A published episode scored 86 failures against its own script. Here is every class of mistake we found making local AI video with LTX 2.5, SageAttention, Krea 2 and ComfyUI, what actually caught each one, and the checks we now run before anything ships.
Hugging Face says MiniMax H3 went up on 28 July. Five weeks later there are around 400 repositories with its name on them, new ones arriving at eight to twenty-four a day, and this week the community visibly stopped asking whether it runs and started asking how fast. Here is what that pace actually looks like, and why our own pipeline is not moving anywhere near that quickly.
Our radar pulled 240 items and the video half was almost entirely one model. Three separate projects landed in a day to squeeze MiniMax H3 onto consumer GPUs, and every one of them is impressive. The thing none of the posts mention is that the licence puts an application form in front of anyone in the US, EU, UK or South Korea.
How we made our AI presenter walk between sets in one continuous shot on a single RTX 5090: the walk-set plate design, guide strength 0.7 to 0.5, the 47% midpoint problem, generate-long-cut-short, chroma key at 0.28, and the honest numbers including the advice of ours that testing killed.
Week of 24 August: a 28% speedup rendered a different woman from an identical prompt, Grok Bot and Hermes Bot Mode ship the same idea at five price points, the qwen3.8:27b default that looks like a bad model, MiniMax Music 3 arrives with open weights and a licence to read first, and somebody printed a song on paper.