Eric J Ma's Website

I built a translator using GPT3 for Ark Channel, a daily devotional written in Chinese. The translator parses the text into structured data, translates the devotional and additional texts, and generates cover images.

I also built a UI using Panel and deployed it on a Dokku server on DigitalOcean. The result? My translation time went from 10-15 minutes to just 2! 🚀

I discuss two Twitter threads that outline potential business ideas that could be built on top of ChatGPT3, a chatbot-based language model. What's the tl;dr? As mankind has done over and over, we build machines to solve mundane and repetitive tasks, and ChatGPT3 and other generative models are no exception!

I gave AI coding tools like GitHub Copilot a try and found them subtly valuable. They're not replacements but augmentations for skilled practitioners who can verify their output. Incorrect outputs can spark creative solutions, making these tools powerful. So, no, AI won't take our jobs, but will supercharge our skills. 🚀👩‍💻🤖

With the Twitter-to-Mastodon migration ramping up, I've received questions on which Mastodon instance to go to. Here's my answer!

Just had an insightful 1:1 with a teammate. We discovered how writing project documentation helps with presentations! 📝 It clarifies our mental knowledge graph, gives us practice in creating narratives, and provides a 'hard copy' of project elements for future use. 🧠💡

After many attempts, I've found a solution to code on my iPad using VSCode in Blink! All my settings sync perfectly, and I can even run Jupyter notebooks. It's like having my MacBook Air, but lighter. 📱💻🚀

Today I learned how to solve yet another fairly esoteric issue with apt.

For us data scientists, whose work is what Cal Newport calls "deep work", staying focused amongst a barrage of distractions is a challenge. In this post, I outline several tips that data scientists can use to stay focused at work.

Python has a batteries-included philosophy. This means that when you install Python, you get a lot of basic functionality thrown in for free! Yet, as time progressed, we (the Python language community) gradually figured out how to accomplish that same functionality with fewer lines of code. In other words, we built opinionated tools to streamline common ways of doing things. In this blog post, I'd like to highlight some of these tools, both big and small, and the built-in modules (or otherwise patterns of usage) that they replace.