Is this an example of the bitter lesson? Could diffusion models not be this good with equal scaling/compute as LLMs?

Well the LLMs are much bigger than diffusion models so I think that's the bitter lesson. You could scale compute for diffusion models though.

▲

demibabs a day ago | parent [-]

Well no, the bitter lesson isn’t the scaling laws themselves. It’s that approaches which can take advantage of scaling laws will ultimately beat ones that can’t.

	▲	moojacob a day ago \| parent \| next [-]
		You learn something new every day! thanks
	▲	E-Reverance a day ago \| parent \| prev [-]
		Scale with compute* not explicit to reference scaling laws of training