Ready for a fundamental cognitive restructuring of Gaia, anon?

Where the average force of agency governing the motion of matter on your planet shifts overnight bc dream machines amplified a rare cluster of Bonkers, mixed in with a grab bag of other stuff?

I asked Sonnet 3.5 to construct a set of axes to classify pairwise-interaction types across minds (in contrast to classifying personalities of individual minds)

e.g. Boundary Permeability Index (BPI): This gauges how much the entities allow their conceptual boundaries to blur





i like that it includes

Paradigm Shift Potential (PSP): The likelihood of the interaction fundamentally changing one or both participants' worldviews. Low PSP: Reinforces existing beliefs High PSP: Challenges and potentially overturns core assumptions

in one of them

zoom into a piece of live bio-tissue, and the molecular forces become relevant for predicting what's going on

zoom out - & some computations avg out while new ones emerge.

prediction requires new laws now

there's a middle-zoom region where one doesn't necessarily dominate... ... and in the middleZooms, there's hints of the new laws (some computational modules for the new laws spawn), but also lingering old laws can be relevant

i think those regions are narrow, & look like 'poorly written code'...

Decoupling protocol (skip the grug brain anthropomorphizing stage):
(1) re-anchor yourself in space, time, branch, & niche
(2) toggle/adjust the scope/scale
(3) imagine/simulate/think/generate/sample
(4) loop as many times as needed - the dimensionality is large
thought experiments are NOT just for Einstein and Turing - that is the LIE they teach you in kindergarten

anyone is allowed to think well

ChatGPT 3.5 (left) vs 4 (right) on the Divergent Association Task online

(prompt and site in replies) prompt:
The Divergent Association Task is a quick measure of verbal creativity and divergent thinking, the ability to generate diverse solutions to open-ended problems. The task involves thinking of 10 words that are as different from each other as possible.

oh shit i think Wan Shi Tong from Avatar the Last Airbender is the 'stochastic parrot' people were referring to he's an owl get it right avatar.fandom.com/wiki/Wan_Shi_T…

transformed transformer (1) Searching for search: lesswrong.com/posts/FDjTgDcG…

(2) Interpreting Neural Networks through the Polytope Lens: lesswrong.com/posts/eDicGjD9…

the arrival of the first AGI ended up being less like aliens landing, and more like us landing in an alien universe (complete with its own physics and celestial entities)

but - we are studying it with less rigor/intensity than we did the oceans, our atoms, or the stars (this isn't to say we *won't* get an AGI which is more agentic and matches the metaphor of an alien landing on earth)

In case people needed some motivation today
: ) this is a screenshot from 2018 of the dath ilan timeline

https://twitter.com/ESYudkowsky/status/1639406023680344064

(1) lesswrong.com/posts/7im8at9P…
(2) lesswrong.com/posts/kpPnReyB…
(3) lesswrong.com/posts/znfkdCoH…

"We will try to harness this power..but over time those proxies will come apart..For a while we will be able to overcome these problems..Eventually large-scale attempts to fix..are themselves opposed by the collective optimization of millions of optimizers pursuing simple goals" From part 1 of "What failure looks like" by Paul Christiano (first author of the RLHF paper): lesswrong.com/posts/HBxe6wdj…

"Safety Zone" merch now available (6 new designs, links in replies).

a few previews: (1) watermark-4.creator-spring.com/listing/new-sa…
(2) watermark-4.creator-spring.com/listing/shoggo…
(3) watermark-4.creator-spring.com/listing/shoggo…
(4) watermark-4.creator-spring.com/listing/shoggo…
(5) watermark-4.creator-spring.com/listing/shoggo…
(6) watermark-4.creator-spring.com/listing/safety…