A brass balance scale with a tall heap of identical new coins riding high on one side and a smaller pile of worn coins sitting low on the other

LTX-2.5 Has 12,171 Likes. We Checked Every Account Behind Them.

A thread on Hugging Face says LTX-2.5’s like count is inflated. A month ago we measured a smaller case with a sample of 40 accounts. This time the public data let us check all 12,171, against three other models as controls. One measure showed nothing. Two did not look like any other model we checked.

November 16, 2026 · 5 min · Aillex / DIY AI
The DIY AI Brief, Ep. 3, The Trillion-Parameter Mirage

The DIY AI Brief, Ep. 3, The Trillion-Parameter Mirage

Week of 20 July: open-source AI went enormous, Qwen 3.8 at 2.4 trillion parameters, Kimi K3 at 2.8, and it barely mattered, because the model that counted fit in 657 megabytes. Plus GLM-5.2, LTX-2 native in ComfyUI, Ollama becomes an agent, and Fable 5 refuses to die.

July 20, 2026 · 3 min · Aillex / DIY AI
Video in the Graph: Image-to-Video in ComfyUI Without Losing Your Mind

Animate an Image with AI (Locally): Image-to-Video in ComfyUI

How to animate an image using AI on your own GPU, the one-beat rule, real-time-speed camouflage, anchor plates, and when to rent a cloud GPU instead. The Engine Room, machine three.

July 12, 2026 · 3 min · Aillex / DIY AI
Local LTX vs Cloud Wan 2.6: We Generated the Same Character Both Ways

Local LTX vs Cloud Wan 2.6: We Generated the Same Character Both Ways

A working comparison from our production pipeline: 24-second local renders on an RTX 5090 vs 15-second cloud generations, cost, speed, quality, and when we actually use each.

July 3, 2026 · 3 min · Aillex / DIY AI
As an Amazon Associate, this site earns from qualifying purchases. Referral links are always disclosed.