Someone opened a tab on Monday to try a ComfyUI node pack, came back on Tuesday, and the repository was a 404. Minutes later the author’s entire GitHub account was gone. The model weights it wrapped are still sitting on Hugging Face, untouched. Most people back up the wrong half of their install, and we were doing it too.
A 31-second photoreal film of a consistent AI presenter, made on one RTX 5090 with no cuts: reference pictures given jobs, a locked voice track, three cheap seed previews then one upscaled finish, each clip extending the last. The five ways it broke, the picture that fixed each one, the plate rules, the LoRA retrain, and the bill.
Qwen released a new image model today at 7.1 billion parameters, roughly a third the size of everything else in the line, with ComfyUI support the same afternoon. Every earlier Qwen image release was Apache 2.0. This one is research and evaluation only, and we read the licence file rather than the thread about it.
Somebody asked today how you count the real cost of AI video once the failed generations are included. We published a 16 minute episode this morning, we keep every take we render, so here are our actual numbers: what we made, what we used, what the retries cost, and the line item nobody prices.
Turn a photograph into an anime character, put that character in six worlds, and make a short film with no dialogue. All of it local, in ComfyUI, on one gaming card. Five things went wrong on the way and all five had the same fix.
YuE2 landed last week, and what happened next is the best argument for open weights we have seen in a while. Quantised to 8 GB by day two. The missing encoder trained from scratch by a stranger on day five. A LoRA trainer the same day. One thing almost nobody is mentioning, though, and it matters if you plan to earn anything.
One group photograph, three people, and we wanted two of them gone without the third one changing. Five free image editors running on our own card took the job in turn. The first thing that happened was the room came back empty. What fixed it was grammar, and the rest of what we learned is below, with the five workflow files.
Every identity benchmark we can find scores the face. We put one full-length figure through five local image editors with an explicit instruction to keep her proportions, scaled the results so the heads match, and measured. Every engine kept her face. Four of five changed her body. Here is the method, the tool, and what the control test says about who this happens to.
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.