Unless things change dramatically, or there is some special sauce that is required & which the labs keep secret, frontier AI capabilities are likely to be available through open models all the way to any sort if possible AGI (with the added implication that no guardrails will hold in open models)

The 18 Reasons Why Complex Systems fail is a classic. They fail for complex reasons, and only adaption & learning by organizations, as well as individual actions keep us safe.

I asked Claude to read this and come up with more rules. They are surprisingly wise (with black background), especially 25.

The big AI companies appear to have bumped up against the limits of the so-called scaling laws — but it may not matter as much as you'd guess. I wrote about the debate that has captivated Silicon Valley this week: www.platformer.news/openai-googl...

Netflix needs to stop asking if I’m still watching and start asking if I moved the laundry to the dryer