Maker packs an opinionated, googly-eyed AI chatbot into a mobile suitcase, powered by an Nvidia Jetson — entirely local machine entity runs Gemma 4 E4B and can respond in 200ms
… A roughly 200 ms Time To First Token TTFT means Sparky can start formulating responses very fast, and then runs at about 14-15 tokens per second, according to the LLM enthusiast. What’s more, the response is natural for a robot , using SenseVoiceSmall for speech-to-text and Piper for text-to-speech. …
