Moritz Hardt thinks that OpenAI and Anthropic’s claims about benchmarks, which are used to test AI models’ capabilities, are misleading. But if the tests aren’t accurate, then how do we know how good the models really are? Read the interview: www.quantamagazine.org/he-wants-ai-...

Show Post

When people keep saying that AI seems to be slowing teams down, it _is_ true in a way, but in a very peculiar way. Teams are getting slower coz they are suddenly moving mountains with previously unthinkable work quality, and this new kind work is so hard to plan and coordinate

Show Post

How is this different than a normal bad paper? Well I review bad papers to give constructive feedback to the authors so they write better papers in the future. If they’re not actually writing the papers, my feedback is worthless and I don’t need to waste time on it

Show Post

well, if the logic of supra-abundance is basically communist then the logic of labor supremacy at the expense of abundance+redistribution = fundamentally fashy, a mindset that will inevitably drift to fashy ideas about labor, value, humanity, work, society, etc.

Show Post

What kinds of aesthetic approaches could make reading into something people specifically seek out on the web? Like enjoyable as a slow-paced experience. Books are persisting as a medium in part because they offer that. What’s the aesthetic of sitting down to read an in-depth blog post in 2003?

Show Post