I am replacing the showerhead attachment. When I loosen the showerhead holder, I am seeing this metal ball ending. My question is this a separate part from the metal pipe that comes out of the wall? If so, can I safely remove at the part indicated by the red arrow?

Just saw this amazing AI-generated sci-fi short film on YouTube, and honestly, this is probably one of the best examples I've seen so far.

What really impressed me is that it feels much closer to a professionally produced short film than a collection of AI-generated clips. The character consistency, environments, lighting/color, camera movements, poses, and facial expressions all seem remarkably consistent throughout.

I'm really curious how something like this is actually made.

For people who have experience with AI filmmaking, what would the workflow for a production like this typically look like? Are they using a combination of image and video models, character/reference LoRAs, ControlNet, ComfyUI, etc.?

I'm particularly interested in how they maintain the same characters and visual style across many different shots and scenes and their expression are so precise and natural, with both short and long duration scens. Is most of the work done through carefully generated keyframes/reference images and then image-to-video, with a lot of manual selection and post-production?

And what models/setup would you expect to be behind something at this level? I'd be interested in both cloud-based workflows and local workflows (e.g. ComfyUI/Wan/etc.).

I've been experimenting with AI video myself, but there's a huge gap between getting one impressive 5-second clip and producing an entire film where everything feels like it belongs to the same movie.

Would love to hear from anyone who has actually worked on something at this level — even a rough breakdown of the workflow would be really interesting.

I am installing a vertical showerhead slide bar to a bathroom, my question is how can I be sure where I am drilling into doesn't contain any piping or electrical wires?

My guess would be to find the wall studs and extend it into the shower area? So for my example, I would use a stud finder on the brown wall above the shower area and find the studs, and then project it onto the shower area vertically?

Pictures attached are the before and after (using AI).

I'm trying to set up a clean Git/GitHub structure for an AI content environment that I'll use across multiple computers. The computers have different GPUs/specs, so I want the code/workflows to be shared, while allowing each machine to have its own environment.

My requirements

• I have multiple Windows PCs with different GPUs/VRAM.
• I want to use both ComfyUI and Wan2GP.
• I don't intend to heavily modify/fork either project.
• I may make occasional small changes to their source/configuration.
• I want to be able to:
  • Pull updates from the official upstream repositories periodically.
  • Keep my own changes/version history.
  • Roll back ComfyUI independently if an update breaks something.
  • Roll back Wan2GP independently.
  • Have the same setup available on all my computers.
• I also want to keep my workflows synchronized between computers.
• ComfyUI/Wan2GP require certain files to physically exist inside their directories, so I may have things like:
  • user/my-workflow.json
  • models/my-model.safetensors
• I don't want to put large model files in GitHub. I'd rather have Git ignore them while the files remain physically inside the ComfyUI/Wan2GP directories.
• The Python/Conda environments should not necessarily be identical between machines because the GPUs and hardware are different.
• Ideally, setup/update scripts should eventually automate setting up each machine.

The structure I'm considering

Dev/
└── ai-content-machine/              ← private GitHub repo
    │
    ├── .git/
    │
    ├── ComfyUI/                     ← private comfyui-custom repo
    │   └── .git/
    │
    ├── Wan2GP/                      ← private wan2gp-custom repo
    │   └── .git/
    │
    ├── script/
    │   ├── setup.ps1
    │   ├── setup-comfyui.ps1
    │   ├── setup-wan2gp.ps1
    │   ├── update.ps1
    │   └── launch.ps1
    │
    ├── config/
    │   └── machines/
    │       ├── desktop.yaml
    │       └── laptop.yaml
    │
    ├── workflow/
    │   ├── comfyui/
    │   └── wan2gp/
    │
    └── README.md

So there would be three independent repositories:

ai-content-machine
comfyui-custom
wan2gp-custom

And ComfyUI/ and Wan2GP/ are cloned into ai-content-machine/.

How I'm thinking about the private repos

For ComfyUI, for example:

official ComfyUI GitHub
        │
        │ upstream
        ▼
comfyui-custom (private GitHub repo)
        │
        │ origin
        ▼
local ComfyUI/

So I'd initially clone the public ComfyUI repo, change the remote:

origin   → my private comfyui-custom
upstream → official ComfyUI

and periodically pull/merge upstream changes.

Same idea for Wan2GP.

My main question

Is this actually a good architecture for my requirements, or am I overcomplicating it?

In particular, I'm wondering:

1. Would you keep ai-content-machine, comfyui-custom, and wan2gp-custom as three separate repos?
2. Would you instead just clone the public ComfyUI/Wan2GP repos directly inside ai-content-machine and manage everything from there?
3. Would Git submodules be preferable?
4. Is there a better way to handle the workflows that need to live physically inside ComfyUI/user/ or similar directories?
5. How would you handle large models that need to physically live under ComfyUI/models/ but shouldn't be committed to Git?
6. Does it make sense to have the same source repos across all computers but create different Python/PyTorch/CUDA environments based on each machine's GPU?
7. Is there anything fundamentally wrong with having independent .git directories nested like this?

I'm mainly looking for a practical setup that won't become a maintenance headache, rather than the most sophisticated Git architecture.

Would appreciate advice from people who maintain AI/ComfyUI setups across multiple machines.Here’s a Reddit-ready version that lays out the requirements and the decision I’m trying to make, rather than over-explaining the implementation:Title: Best Git repo structure for managing ComfyUI/Wan2GP across multiple PCs?I'm trying to set up a clean Git/GitHub structure for an AI content environment that I'll use across multiple computers. The computers have different GPUs/specs, so I want the code/workflows to be shared, while allowing each machine to have its own environment.My requirementsI have multiple Windows PCs with different GPUs/VRAM.

I want to use both ComfyUI and Wan2GP.

I don't intend to heavily modify/fork either project.

I may make occasional small changes to their source/configuration.

I want to be able to:

Pull updates from the official upstream repositories periodically.

Keep my own changes/version history.

Roll back ComfyUI independently if an update breaks something.

Roll back Wan2GP independently.

Have the same setup available on all my computers.

I also want to keep my workflows synchronized between computers.

ComfyUI/Wan2GP require certain files to physically exist inside their directories, so I may have things like:

user/my-workflow.json

models/my-model.safetensors

I don't want to put large model files in GitHub. I'd rather have Git ignore them while the files remain physically inside the ComfyUI/Wan2GP directories.

The Python/Conda environments should not necessarily be identical between machines because the GPUs and hardware are different.

Ideally, setup/update scripts should eventually automate setting up each machine.The structure I'm consideringDev/
└── ai-content-machine/ ← private GitHub repo
│
├── .git/
│
├── ComfyUI/ ← private comfyui-custom repo
│ └── .git/
│
├── Wan2GP/ ← private wan2gp-custom repo
│ └── .git/
│
├── script/
│ ├── setup.ps1
│ ├── setup-comfyui.ps1
│ ├── setup-wan2gp.ps1
│ ├── update.ps1
│ └── launch.ps1
│
├── config/
│ └── machines/
│ ├── desktop.yaml
│ └── laptop.yaml
│
├── workflow/
│ ├── comfyui/
│ └── wan2gp/
│
└── README.md
So there would be three independent repositories:ai-content-machine
comfyui-custom
wan2gp-custom
And ComfyUI/ and Wan2GP/ are cloned into ai-content-machine/.How I'm thinking about the private reposFor ComfyUI, for example:official ComfyUI GitHub
│
│ upstream
▼
comfyui-custom (private GitHub repo)
│
│ origin
▼
local ComfyUI/
So I'd initially clone the public ComfyUI repo, change the remote:origin → my private comfyui-custom
upstream → official ComfyUI
and periodically pull/merge upstream changes.Same idea for Wan2GP.My main questionIs this actually a good architecture for my requirements, or am I overcomplicating it?In particular, I'm wondering:Would you keep ai-content-machine, comfyui-custom, and wan2gp-custom as three separate repos?

Would you instead just clone the public ComfyUI/Wan2GP repos directly inside ai-content-machine and manage everything from there?

Would Git submodules be preferable?

Is there a better way to handle the workflows that need to live physically inside ComfyUI/user/ or similar directories?

How would you handle large models that need to physically live under ComfyUI/models/ but shouldn't be committed to Git?

Does it make sense to have the same source repos across all computers but create different Python/PyTorch/CUDA environments based on each machine's GPU?

Is there anything fundamentally wrong with having independent .git directories nested like this?I'm mainly looking for a practical setup that won't become a maintenance headache, rather than the most sophisticated Git architecture.Would appreciate advice from people who maintain AI/ComfyUI setups across multiple machines.

There are more and more weeds and crabgrass growing in my rock areas. The rock areas have landscape fabric underneath, with about 4–5 inches of rock on top. However, over the years, I think a small amount of soil and organic matter has accumulated underneath and between the rocks. It seems that the weeds and crabgrass only need a tiny amount of soil to take root and grow.

Is there anything I can do to completely clear the weeds and crabgrass and prevent them from growing back, ideally for the long term?

I notice this white tube that's right outside my window within the window well area. Currently it is completely covered with dirt and rocks inside. Can I fill it up with dirt and rocks over it?

I notice this white tube that's right outside my window within the window well area. Currently it is completely covered with dirt and rocks inside. Can I fill it up with dirt and rocks over it?