$ cat ai.txt
## AI
I am deeply interested in the practical applications of Artificial Intelligence.
My recent AI work focuses on two key approaches to modern AI: secure on-device processing and context-aware cloud integration.
To explore local inference, I recently built this macOS app (free here on the Mac App Store) that demonstrates the power of the Apple Neural Engine. 'AI Color Assistant' allows users to describe a design vision in plain English and receive an AI-curated six-color palette. By leveraging on-device Apple Intelligence, this tool represents a fundamental shift in AI software: Designed with a privacy-first architecture, every palette generation happens entirely on your Mac, ensuring your creative concepts never leave your machine. This reliance on local processing via Apple Silicon delivers near-instant results with zero latency, and you never have to worry about API costs or usage limits as the AI processing happens on your device.
While my macOS app focuses on private, local processing, my web-based voice assistant explores Context Window Injection. To use a non-technical analogy, I built a system that injects my latest real-time media consumption data (Specifically, the music I’m listening to lately via the Spotify API and the books I’ve read recently via the Hardcover App API) into the AI’s “subconscious mind” in the moment just before each conversation begins.
You can explore this functionality yourself for free by initiating a conversation with my AI Assistant and guiding the discussion toward questions about my taste in music or novels.
If you'd like to get in touch for a chat about AI in Boston, send me an email here.
To explore local inference, I recently built this macOS app (free here on the Mac App Store) that demonstrates the power of the Apple Neural Engine. 'AI Color Assistant' allows users to describe a design vision in plain English and receive an AI-curated six-color palette. By leveraging on-device Apple Intelligence, this tool represents a fundamental shift in AI software: Designed with a privacy-first architecture, every palette generation happens entirely on your Mac, ensuring your creative concepts never leave your machine. This reliance on local processing via Apple Silicon delivers near-instant results with zero latency, and you never have to worry about API costs or usage limits as the AI processing happens on your device.
While my macOS app focuses on private, local processing, my web-based voice assistant explores Context Window Injection. To use a non-technical analogy, I built a system that injects my latest real-time media consumption data (Specifically, the music I’m listening to lately via the Spotify API and the books I’ve read recently via the Hardcover App API) into the AI’s “subconscious mind” in the moment just before each conversation begins.
You can explore this functionality yourself for free by initiating a conversation with my AI Assistant and guiding the discussion toward questions about my taste in music or novels.
If you'd like to get in touch for a chat about AI in Boston, send me an email here.
