It would be interesting to get a sense of what OpenAI is thinking in terms of dialogue agents and if they are a priority or not. Even if it’s a non-multimodal agent, it still seems as if a “next-gen Siri” which can prompt/respond/remember/etc (via audio) would be profoundly useful to humans across the board.
DALL-E’s use cases are mind-blowing but it remains a fact that not everyone “cares” about image generation/art/etc. I think it’s safe to assume a sufficiently advanced voice assistant would be instantly useable to most, if not all humans. Voice-driven dialogue would almost certainly be the most impactful/generalizable tool.
Surely OpenAI knows this… so I wonder why they don’t directly address this topic.