I mean the Zero Data reasoning directly cites previous work on the same principle. That in particular seems like a significant step forwards though - one of the main critiques of current methods is "humans don't learn by ingesting terabytes of common crawl, they learn from experience".
Of course some works are indeed genuinely good ideas. But sometimes things just don’t work out quite as well. Many nature papers in my field are dead ends applications-wise.
I’m not saying any of those works specifically are, just that research should be approached with a healthy dose of skepticism.
