Join 📚 J'S Highlights
A batch of the best highlights from what J's read, .
In deep learning’s forward–backward sweep of prediction–correction, it seems like the process of weaving back and forth has followed the history of programmable systems, from the first Jacquard weaving looms of the early nineteenth century, which were controlled by punched cards, to the deep learning systems of the twenty-first century.
Resisting AI
Dan McQuillan
For example, when a community loses its memory, its members no longer know one another. How can they know one another if they have forgotten or have never learned one another’s stories? If they do not know one another’s stories, how can they know whether or not to trust one another? People who do not trust one another do not help one another, and moreover they fear one another. And this is our predicament now. Because of a general distrust and suspicion, we not only lose one another’s help and companionship, but we are all now living in jeopardy of being sued.
The World-Ending Fire
Wendell Berry
The content serves as a description of the form, with all its doublings and chinese-box secrets, thus foregrounding the notion of the secret itself - what Deleuze describes as something the content of which is too big for its form. Such metonymy works by appeal to the suggestion that the (hidden) functional blueprint of the two disparate items - the house and the mind, or the form and the content - is identical.
...catch up on these, and many more highlights
Readwise