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.
A Mozilla analysis put the open to closed capability gap at about 4.4 months, and it went round the local AI world as good news. It is good news. It is also being read as something it does not say, because the open model doing the catching up is 2.78 trillion parameters and will never see the inside of your computer.
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.
Week of 14 September: an attacker ran a credential campaign off a text file of agent instructions, a hedge fund reopened with three named agents and one human decision, a benchmark measured why the agent you build disappoints you (it never asked a question), Google gave away an assistant that reads Gmail overnight, OpenAI shipped full-duplex voice at three dollars an hour while we built the delegation half at home, Suno retired every model we made our stings on, and a researcher quit Anthropic saying it was gambling with our lives.
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.
We took a stock Asus TUF gaming laptop with an RTX 3070 and 8 GB of video memory, and benched eight local models overnight across ten tasks. The result that surprised us: a 30 billion parameter model with two thirds of itself off the card beat a 14 billion parameter model that fit better, by nearly four times. File size does not predict speed. All 103 measured rows are free below, failures included.
Someone asked this morning which models to back up now that NVIDIA is buying Hugging Face. We told people to keep local copies a week ago, and four days ago we wrote about our own disk filling up and breaking our publishing. Both of those are true at once, so here is the version that reconciles them.
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.