TL;DR
- Learning the language helps you feel comfortable; building a project is when you genuinely develop the necessary coding skills.
- ChatGPT is revolutionary. It can potentially write the entire code for you, but a basic grasp of the coding concepts involved remains essential.
- Opening up about your struggles instead of viewing everyone as an adversary can lead to finding kindness and support from others.
How it all started
- Just as mastering a foreign language once deepened my understanding of a new culture, I realised that learning a programming language could be my window into the deep world of AI, truly grasping its capabilities and potential. Still, this realisation wasn’t quite the push I needed to dive in, until 👇
- I had such imposter syndrome being a non-coder in SF, feeling out of place and unable to fully participate in hackathons that predominantly centred around tech demos. The pressure intensified with events requesting a showcase of GitHub contributions, something I had none to show.
- Admittedly, I somewhat naively assumed it would be straightforward. Developers seemed to have a myriad of tools and GitHub repositories at their disposal to facilitate their work. Plus, the buzz around ChatGPT suggested it could virtually code on its own.
So I set The Goal
😤 “I will ship my first coding project AS A SOLO HACKER at the GenLab Hackathon”. The event was scheduled roughly a month away.
The Result
🌈 The emotional roller coaster peaked with the elation of code submission, making every moment throughout the journey worthwhile!
The Bonus
- 🏅 Got into the finalists
- 😊 Earned the nickname “solo dev” from a fellow developer
- 👩🏻💻 Reconnected with a friend over the project, and now we’re aiming to top the app store chart in three months. Hello, Xcode!
How I did it
Here’s a snapshot of my journey. From structured courses to binge-watching YouTube tutorials, I embraced a spectrum of learning avenues.
Codecademy
Courses taken:
Review: Absolutely loved the UI/UX — the lessons and the in-browser IDE facilitated a smooth learning curve. Their structured paths are a bonus for anyone eyeing a career in tech.
Note: I gave myself a 7-day target to complete the ‘Learn Python 3’ course. Then Codecademy extended my free trial period for another week, so I signed up for more courses. But I didn’t finish the extra ones as I started to doubt whether this method would effectively lead me to my goal.
Uplimit
Course taken:
Review: The course is only one week long and perfect for beginners keen on assimilating the basics of building an LLM app.
Note: I didn’t go forward with the project because the scope didn’t quite resonate with me. Instead, I channeled my energy into brainstorming ideas for the hackathon.
Deeplearning.ai
Courses taken:
- Generative AI with Large Language Models
- ChatGPT Prompt Engineering for Developers
- LangChain for LLM Application Development
- LangChain: Chat with your data
- Finetuning Large Language Models
Review: My go-to for understanding the AI landscape while on the move. The concise lessons fit perfectly into my daily routine, offering knowledge bites even during grocery runs.
Note: I dabbled in the ‘Generative AI with Large Language Models’ course long before I decided to learn to code. The course is great even for non-coders.
YouTube videos
I’ve watched a plethora of them, too many to list here, but I am immensely grateful to all the creators for their content!
ChatGPT
Projects built:
Initially, I assumed creating a Chrome extension for my first project would be straightforward. But it turned out that building a browser extension involves grappling with JavaScript and HTML — languages with which I was completely unfamiliar. That’s where ChatGPT proved invaluable, guiding me to successfully build a Chrome extension in a mere three days.
Utilising ChatGPT extensively while building these projects significantly steepened my learning curve. It became my indispensable assistant for everything — from understanding broad concepts to nailing down line-by-line coding and tackling debugging issues. It elevated the learning-by-doing approach to an entirely new level.
But does ChatGPT work as a no-code tool? Can a complete beginner simply copy and paste the code generated by ChatGPT to build a functioning project?
The definitive answer is no. Fundamental knowledge remains a prerequisite. The process involves clear and precise communication — akin to interacting with another individual, one cannot simply overload them with information. Constructing a foundational roadmap of the project in one’s mind is essential, and this demands an understanding of how developers initiate their projects.
ChatGPT has its set of constraints. When the tasks became more intricate, it tended to falter, struggling with the nuances. Given it was trained on a dataset that might not include the most recent updates, it had moments of stumbling, especially with newer frameworks. But the silver lining here is that many of these hitches seem to be less prevalent in its newer counterpart, GPT-4.
What I Built
Eli5 Chrome Extension
This project marked the beginning of my coding adventure — my very first project! I decided to open-source it, which means you can use it for free here. Initially, it was meant to be my hackathon submission, but since I completed it ahead of schedule, it became a precursor to the real deal.
What it does: Provides a pop-up explanation ‘as if you were five’ for any highlighted term, facilitating a smoother reading experience.
Development timeline: Set out a week for it but nailed it in 3 days 👀
Interactive Children’s Book
The Hackathon Project
What it does: Transforms any topic into an interactive book for children aged 8–12, encouraging early exploration and development of diverse skills and professions.
Development Timeline: Built the complete LLM web app, with front-end and back-end in … roughly a day! 😳
Hackathon Result: Got into the finalists presenting in front of Jeremiah Owyang (Blitzscaling Ventures), Nicolai Wadstorm (BootstrapLabs), Elana Lian (Intel Capital) and an audience of over 200 hackers.
(Sign up for early access here)
What I Learned in One Week of Building Two Projects
- JavaScript
- HTML
- CSS
- Advanced Python concepts
- Manifest V3 for Chrome extensions
- Flask
- API
- Bash
- Debugging
- Setting up a local server
- How to test code
- How to structure a coding project
How I recommend aspiring coders to start
🎢 FAFO. Leverage the non-judgmental nature of ChatGPT to ask any question, no matter how “stupid” it may seem.
What’s next for me
Driven by the incredible potential of large language models and the vibrant AI development scene in SF, I am setting my sights on revolutionising the distribution of AI dev tools. My goal is a future where these tools are not just widely accessible, but seamlessly integrated into daily life for everyone. Stay tuned! 🫡
