Here's a silly app that I created using Vercel called Disaster Day. I planned to submit it to a hackathon but not sure yet.
Here is the basic premise of the game.
Design an AWS architecture on a strict hourly budget, then defend it against a randomized barrage of escalating disasters. The Cloud World Model simulates every outcome. How long can your stack stay up?
A few days ago I commented on Flo s discussion about confidential transfers and AI agents:
https://www.producthunt.com/p/so...
The conversation wasn t about adding another blockchain. It was about something simpler: if an AI agent is spending my money, I want guardrails and visibility.
We re building Cloud World Model, a platform that lets people simulate cloud infrastructure and architecture decisions before spending money on real cloud resources.
Think:
AWS, GCP, Azure, OCI, and DigitalOcean environments
One thing I ve always wanted as a cloud engineer is the ability to experience major cloud incidents instead of just reading the postmortem.
We ve started adding real-world cloud incidents as scenarios in Cloud World Model so engineers and AI agents can safely explore how systems behave during major outages.
I created Canvas Cloud AI in 2025 to facilitate hands-on learning in the cloud. Although I did not launch it on Product Hunt, thousands of users have utilized it to generate cloud architectures for learning. Over the past year, while speaking with developers, engineers, and architects about Canvas Cloud AI, I consistently heard two main concerns: one, that cloud services can be expensive, and two, that connecting to cloud providers can create bottlenecks for hands-on learning.
After hearing this feedback repeatedly, I began to explore the possibility of simulating the cloud without relying on connections to cloud providers. This idea eventually led to the creation of Cloud World Model. I have now decided to launch Cloud World Model on Product Hunt, something I did not do with our earlier product, Canvas Cloud AI.
