Large Language Models (LLMs) are exciting to any developer. It was a year ago that OpenAI changed the game with GPT-3. The hype has subsided and that just reflects the reality of any new technological breakthrough and its adoption. The pace of change is fast, but naturally it won’t keep pace with mass expectations unless we hit the signularity.
I’ve been digging into the OpenAI API and have some tool ideas I want to bring to life.
Accelerated Language Learning
I enjoy learning languages. I aim to use OpenAI’s Whisper to converse in the language I’m learning with an LLM-based instructor. This way, I can practice conversational German anytime. Right now, using Chrome Plugins like Voice Control for ChatGPT, you can speak to the LLM in various languages and hear a response. Just give ChatGPT your goals and how you want the language exchange to work, and you’re set.
Pair programming with Co-Pilots
AI assisted coding and content creation are here. GitHub Co-pilot is helping me to finish this sentence! The barrier to entry for learning new programming languages and tools is now much lower. Though I can work through most documentation nowadays, asking questions directly is often quicker than searching. Pose a question, and you often get a near-perfect example in return. Having GitHub Copilot as my pair programmer made it easier to get a static site running for my photography portfolio. It was also useful when first exploring the GitHub graph API and the petfinder API.