I thought this was an excellent paper! Thanks to Anthropic for asking me to write a review of it, linked below
I've long suspected that models have some kind of "working memory" to store intermediate variables during a forward pass and IMO this paper has the best evidence yet
New #preprint@LPiolopez@BenjaminLy61243
https://t.co/yDzBbbQI6Q
"Alignment Is to a Virtual Governor: A Theory of Coordination in Diverse Intelligence"
Abstract: Alignment is a central problem in systems composed of diverse, interacting components, from biological development to social and engineered systems. One crucial aspect of this problem is determining an answer to the question of to whom or to what the alignment should be. We argue that in decentralized systems, alignment is necessarily to a virtual governor, an abstract governing entity embodied in the coordinating relationships among agents, such as bioelectric networks or the price system. Crucially, despite not being a physical object, virtual governors are causally instructive, controlling the behavior of a system by aligning its parts toward higher-level goals. As a result, agents, from cells to humans and all manner of diverse intelligences, behave as if pursuing the objectives of the virtual governor. We argue that alignment is necessarily to a virtual governor in decentralized, coordinated systems, discuss examples of virtual governors, and describe how virtual governors are constructed by the very components they align.
Revisiting Rockefeller Jr’s Credo this morning and it dawned on me - I don’t think I have heard an AI leader utter even a single one of these principles that have held together and defined the West for 250 years.
I just released an all-in-one tool that lets you extract textures from any perspective, turn them into seamless textures, and use them on your 3D models.
#gamedev#indiedev#gamedevtools#itchio#indiegame
I think the basic issue in all such discussions is that these kinds of competencies aren't some kind of essential "human attributes". Measuring cognitive properties by humans is myopic and causes all kinds of pseudoproblems. The emerging field of diverse intelligence has better frameworks. A spectrum of highly variable intelligences, and lots of research on which kinds of architectures enable which kinds of patterns (of behavior, of computation, of physiology, etc.), is more useful for discussions of natural, artificial, and hybrid agents of varying provenance and composition.
https://t.co/lex2pO1h39
https://t.co/j5r0xpuTaV
and more at https://t.co/jVx1yp9ZJl, plus lots of other good labs.
One of the most amazing things I’ve ever seen: a standing ovation for the full Daraxonrasib results
I feel inspired and energised, to put it mildly — we have a targeted therapy for pancreatic cancer now, and nothing is undruggable anymore
The secret password still works too, you know*... 😉
(*You'll have to go "WayBack" to 2011 and click on the animated orbment up top to try it out, though~)