AI Art Creator & Adobe Firefly Ambassador | Surreal visuals & glitch aesthetics
Mar 30, 2026
•
9 min read
One of Them Has a Serious Problem
Mar 28, 2026
3 min read
Impressions swung 2.3x. Bookmarks swung 6x. The payout didn't care.
Mar 27, 2026
What happens when you give AI a container it cannot see through.
Mar 26, 2026
10 min read
Flux 1.1 Pro: 8.50. Ultra: 7.83. Ultra Raw: 7.46. The pattern holds across images and video.
Mar 25, 2026
12 min read
Four models. Sixteen images. Not one scored below 8.78. Here is the formula.
Mar 24, 2026
11 min read
How Nano Banana Pro went from 8.26 to 9.23 without changing a single word of the prompt.
Mar 23, 2026
13 min read
How container choice and ecosystem design outperformed switching between 12 AI models.
Mar 22, 2026
How I use Adobe Firefly Boards to R&D my prompts before they reach your timeline. #AdobeFireflyAmbassadors #Ad #HowToAdobeFirefly
Mar 20, 2026
8 min read
One prompt. Seven models. Five completely different interpretations of the same concept.
Mar 19, 2026
Every prompt element broken down. One emergent behavior I didn't expect.
Mar 18, 2026
Why AI art gets more convincing the further you push from reality.
Mar 17, 2026
How describing what you can touch changed what Firefly could see.
Mar 16, 2026
24 images. 6 surfaces. One model, three interpretations.
Mar 13, 2026
Part 2: Three-Era Temporal Portals | Adobe Firefly Image 5 | 30 images
Mar 10, 2026
When research has a deadline and a judge, every choice means something different.
14 min read
Why a Scored Failure Is Worth More Than an Undocumented Win
Mar 9, 2026
The community asked. Creators answered. And the responses were more honest than I expected.
Mar 6, 2026
60 images, 15 variations, one repeatable formula. This is what the data actually says.
Mar 5, 2026
68 images. 17 materials. 4 articles of discovery. Here's the entire system compressed into one decision framework you can use on your next prompt.
Mar 4, 2026
17 min read
I built a suspension bridge from melting candle wax. The burning wicks were the only light source in the scene. It scored the highest of 68 images. Here's why everything I'd learned stacked into one frame.
Mar 3, 2026
16 min read
I put my three best materials on a Gothic rose window. Two performed as expected. One dropped almost a full point and forced me to rewrite the rules.
Mar 2, 2026
I took the three underperformers from my first test, made targeted prompt fixes, and every one improved. But one change produced a result I didn't expect.
Feb 27, 2026
I tested 5 impossible materials on the same architecture. 68 images later, I found the rule that separates "interesting" from "impossible."
Feb 26, 2026
The single variable that matters more than your prompt, your material, or your subject.
Feb 25, 2026
Why the best creators never stop learning.