Engineering
A Bad Photo and GPT-6
I've been interviewing for new roles, so I've been updating my wardrobe from my normal black t-shirt, jeans, and boots uniform I've been wearing for the past ten years. I've been told I look like I have a cartoon character's closet because it's all just black t-shirts.
I took a photo of a new sport jacket I got to send to my girlfriend. My room was a mess because I've been living in there the past two weeks since it's the only room in the house with AC and Huntington Beach is going through a heat wave.
I took this photo.

Looks pretty bad. Background is busy, there are two different types of lighting, plus it's through my phone.
Later that day, I'm working out in my room (shout out to Power Blocks) watching a YouTube video on GPT-6. There were a lot of bold claims about how it solved coding and we'll all be making video games now, but the person did show that GPT-6 is really good at image processing.
At this point in my photography career I pretty much only use Lightroom, a few profiles, and a few presets. I hate masking in Photoshop and spending hours in there. I used to love it and thought the time in post really made an image. Now I just want to light something and shoot at the right exposure to get what I want, then adjust the color.
So this evening I was working on my website and thought about using Codex with the Astra Light model to remove the background of that image, then resize it so it would be a standard crop.
I gave it a few simple prompts.

"Remove the background."

"Take the image and make it a dramatic black and white photo that I can use on my website's about page. You can crop if needed but don't lose too much quality. Add a background also, but it cannot look fake."

"Can you add some width to it so it's more of a standard crop."

"That looks great, can you give me a color version also."
Overall I think it did a really good job, especially with my hair. It was a good mask. It made a pretty shit photo pretty okay.