Work in Progress
Getting AI-written prose to sound really human and match a person's individual writing style—or even a company's brand style—remains a serious challenge. In my experience, it takes a lot of time, effort, and careful prompting to coax a mediocre first draft from most large
ChatGPT Plus
As you may recall from several previous posts, I've been working on creating audio recordings of beginner Mandarin lessons for English speakers using text-to-speech (TTS) technology. I used OpenAI's TTS model, which handles both English and Mandarin well, and also tried ElevenLabs, which offers many voice
Work in Progress
It was a good morning. The sun was peeking over the horizon. I had a steaming mug of coffee, and my MacOS scripting-and-voiceover app was behaving as expected. I'd even added voices from the ElevenLabs text-to-speech (TTS) model by adapting this Swift package. While it doesn't
Work in Progress
Remember my quest to create Mandarin lessons for beginners and record them with AI? I wanted to use a OpenAI's GPT-4 model to generate the lessons and then record them with AI voices from OpenAI and ElevenLabs. How hard could it be? Spoiler alert: It was actually pretty
Text-to-Speech (TTS)
After some long conversations with my Mandarin for Beginners GPT, I decided that allowing it to improvise lessons and deliver them through voice chat was not the best approach. Users could easily burn through their 40 messages per three hours just with some basic pronunciation and conversation practice. I know
Work in Progress
I love the idea of learning and speaking other languages, but my enthusiasm exceeds my ability. While I can manage small talk in French and Spanish, my attempt to learn Mandarin ended with dropping the class to preserve my GPA. So, of course, I am trying to build an infinitely