It seems a bit silly since OpenAI LLMs already can output structured data.

Jev is orders of magnitudes cheaper and faster, its a very interesting new paradigm. (Jev pioneered this about 2-3 weeks ago).

	▲	OutOfHere 3 hours ago \| parent [-]
		But is one approach better calibrated and more accurate than the other? And are you implying that OpenAI will use a Jev-like approach?