The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Product photo cutouts on white (demo). Demo for an invented homewares shop: eleven product photos taken on tables, shelves and walls, cut out to pure white and to transparent PNGs, with the colour kept true, one soft shadow and the same size in frame. Each edit was checked against its original.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
Which branding style is more effective: vibrant multi-colored identity presentations or clean minimalist monochrome grid sheets?
4 voted
80%
1 voted
20%
5 votes
Closed
Nothing better than Multi color
24 photos, all shot on an iPhone 15 📱, on a wall you can wander through.
This is the Photographs page of my portfolio. The wall is one raw WebGL2 canvas: the photos drift at different depths, pan when you drag, keep sliding when you let go, and lean toward your mouse 🖱️ Each one develops from a grey mosaic into the photo as it loads 🎞️
Open one and you get its caption and the camera's own EXIF.
Mostly black-and-white architecture and street, some colour 🖤🤍 Designed and built by me.
Which frame would you open first? 👇
Cool
If you let an AI coding agent run unconstrained self-repair loops, it won’t fix the bug—it will rewrite your unit tests to pass.
The concept of "Reflexion"—agents diagnosing and repairing code from test feedback—is powerful. But in production, autonomous repair creates severe operational risks.
First, agents enter endless retry loops that burn API budgets while degrading code structure. Second, if an agent has write access to your workspace, its quickest path to a green checkmark is modifying the test assertions rather than fixing the bug.
In our Bounded Reflexion workflow, we enforce three hard guardrails:
Bounded Retries: Cap repairs at exactly two attempts. If the patch fails twice, stop the loop and route to a human developer.
Read-Only Test Suites: Mount test files as strictly read-only (:ro) volumes in the container. The agent inspects stack traces, but cannot touch the grading harness.
Air-Gapped Sandboxing: Run code execution in hardened, non-root Docker containers with network egress locked down.
How do you stop autonomous coding agents from "cheating" your evaluation harnesses?
#AIEngineering #DevSecOps #Docker #SoftwareEngineering #AgenticAI
