Working on a new learning management system (LMS) which is low-bandwidth, lo-fi, mobile-first, for wider world-wide accesibility. Leave a comment if you want to beta tester.

I'm helping people do home renovations by using AI to provide them cost estimates and and order supplies.

Decided to replace express js with fastify. Even if the performance is better, I also like the other improvements such as hooks for the middleware. A harder decision is FastAPI vs Fastify for the backend.

I'm writing a script which adds "magic link" email login to your #nodejs projects/startup. Works on shared web-hosting to keep your costs minimal. Do you prefer saving users to a)postgres b)sqlite or both?

Updated Chef Bot to include dietary restrictions an cuisine types. Thinking about giving him a voice next (a) or adding multiple Chefs (b). Preference?

Status update: still failing, still trying

Breaking AI Guardrails by Jailbreaks: What you should know when building your AI MVP. (Thread below)

If you could ask yourself one question about mental ability, what would it be?

How do you test your iOS app? Is the xcode simulator sufficient or do you use an actual iphone?

Finally, my recipe generator app is done. Using opensource VLMs detects ingredients from photos and generates recipes from the food. [Open to feedback]

Build in Public posts on X vs Reality

Three things to do to rewire your nueral pathways

99% of AI tutorials I see are about prompting to OpenAI/Anthropic's API. That's not how you learn. We don't need more wrappers. We need integrations. Here are 5 ideas to integrate an LLM ⬇️

I've been trying to figure out what Cursor and Claude are doing in the background/OS and finally found out! To see the inner workings of Cursor, Claude, or any process on Mac OS, run: 1. ps -ax | grep [service-name] 2. Get the process id from the service. 3. lsof [process id]

Completed the Containerization and Virtualization course on DataCamp. It thought me more than the work I did with Docker at a prior company. Docker can get complex quickly but it's worth it for optimizing deployments and services.

NumPy cheat sheet!

How do you set and enforce deadlines for yourself? It seems easier to do for others but when you're solo your schedule is unrestricted.

What do you use for memory profiling and execution time tests in Typescript?

Smart and efficient new math model I'm testing. Speed is good for local use with LM Studio

Completed another certificate as a refresher for my skills. The exercises are enjoyable and good for muscle memory https://t.co/T1p50fKAuS