Absolutely. People on X keep making the mistake of assuming cloud / network latency is the problem here. The vast majority of America is within 10ms of a data

Absolutely. People on X keep making the mistake of assuming cloud / network latency is the problem here. The vast majority of America is within 10ms of a data center. That's nothing. The current challenge for most interaction is ASR -> prompt processing latency. This will be improved with multimodal models on specialized hardware like Groq.