TextFX is an AI experiment designed to help rappers, writers, and wordsmiths expand their process. It was created in collaboration with Lupe Fiasco, drawing inspiration from the lyrical and linguistic techniques he has developed throughout his career. TextFX consists of 10 tools, each is designed to explore creative possibilities with text and language.
TextFX is powered by Google’s PaLM 2 large language model, via the PaLM API.
We built TextFX as an experiment to demonstrate how generative language technologies can empower the creativity and workflows of artists and creators. This app is also an example of how you can use the PaLM API to build applications that leverage Google’s next-generation large language models (LLMs).
Google Lab Sessions is an ongoing series of collaborations between our latest AI technology and visionaries from all realms of human endeavor—from artists to academics, scientists to scientists, creators to entrepreneurs, and more.
Lab Sessions is part of labs.withgoogle.com—our place to test early stage experiments and shape the future of technology, together.
The PaLM API is an entry point to Google’s large language models, and it enables developers to build AI-powered applications for a variety of use cases. If you’d like to learn more about the PaLM API, head here.
If you’d like to build your own experiments like this, you’ll need an API key. An API key is a unique identifier that lets you access the PaLM API. If you have been granted access to the PaLM API, you can get an API key by following the instructions here. Otherwise, you can join the waitlist.
MakerSuite is a platform that lets users easily experiment with LLMs. We used MakerSuite to prototype all of the tools in TextFX.
To get started using MakerSuite, head here.
Wasalu Jaco, professionally known as Lupe Fiasco, is a Grammy award-winning rapper, professor, entrepreneur, and community advocate. Rising to fame in 2006, Lupe has released eight critically acclaimed studio albums. His latest is Drill Music In Zion, released in June of 2022.
Most recently, Lupe’s passion for cognitive science, linguistics, semiotics and computing has landed him an MLK Visiting Professors and Scholars fellowship at MIT, where he teaches a class on rap theory and practice.
Part of what makes large language models so useful is that they’re creative tools that can address many different language tasks. Unfortunately, this also means that LLMs can generate outputs that you don’t expect, including text that’s offensive, insensitive, or factually incorrect. Although the PaLM API filters these kinds of outputs, we cannot guarantee that they will be removed completely.
Visit Official Website