Rationality Feed: Last Month's Best Posts

by deluks917 2y12th Feb 20182 min read1 comment

64


I write a daily rational feed. I write up summaries/teasers for the previous day's article that I found interesting and/or enjoyable. I follow most rationalist blogs as well as LW2.0 and the EA Forum on RSS. The daily feed is posted in the SSC Discord and on my Wordpress Blog . However the rational feed includes quite alot of articles and many people have requested a more heavily pruned feed. In the past my best-of lists have been well received. I am going to try a monthly best-of list. Lets get to it:

=== The Babble and Prune Sequence by alkjash

This series is probably my favorite concept posted on lesswrong 2.0 so far.

Babble - Babble and Prune model of thought generation: Babble with a weak heuristic to generate many more possibilities than necessary, Prune with a strong heuristic to find a best, or the satisfactory one.

More Babble - Modeling Babble as a random walk on a graph where nodes can represent concepts or words and edges are mental connections. You can improve your babel/prune by increasing connections between different areas or by reducing connections to unproductive clusters. An Analogy between babble/prune perspective on writing and Generative Adversarial Neural Nets.

Prune - People’s filters on their babble are far too strong. The early filters: What you think consciously, what you speak, and what you write down. These filters eliminate all but a trickle of babble. Specific mental techniques to weaken the gates.

NP is the God of Babble. His law is: humans will always be much better at verifying wisdom than producing it. Therefore, go forth and Babble!

=== Politics and Social Dynamics:

An Apology Is A Surrender by Zachary Jacobi – [Recommended largely based on the comments]. If you apologize you should surrender. If you keep fighting its not a real apology. However fake apologies are common. Many commenters point out this implies you often should not surrender to the people demanding an apology. Hence its often rational to ‘fake apologize’.

Nice-manning by The Unit of Caring – How to put yourself in a situation where you are equipped to be nice. Niceness advice: Have a wide audience in mind, When someone is a dick, you can actually just pretend they weren’t, Vary topics a lot.

Whence Comes Nhilism by samzdat – Samzdat wrote a long series of posts on modernity, power and nhilism. References include Nietzche and Seeing Like a State. This peice compresses the series down to 1200 words. However its quite clear and easy to read.

=== Thoughts on Applying Mental Models

Be The Baby With A Hammer by Brian Lui – “Conventional wisdom warns against being fascinated by a new idea and applying it to everything. But the reverse is true; we should deliberately apply new ideas more than it seems necessary.”

Hammers And Nails by alkjash – Hammers are people who apply the same techniques to a variety of problems. Nails are people who apply many techniques to the same few problems. Why its better to either a dedicated nail or a dedicated hammer. Hammers can be underrated, powerful hammers have honed their simple tricks into superweapons.

=== AI Safety

AI: Racing Toward The Brink by Waking Up with Sam Harris – “Sam Harris speaks with Eliezer Yudkowsky about the nature of intelligence, different types of AI, the “alignment problem,” IS vs OUGHT, the possibility that future AI might deceive us, the AI arms race, conscious AI, coordination problems, and other topics.”

Honest Organizations by Paul Christiano – Goal: Create an honest AI organization. Strategy: The company makes public statements that it promises are not misleading. Employees commit to publicly dispute the statements if they find them misleading. This strategy cannot create trust but it can dramatically amplify initial levels of trust.

=== In Depth Articles: These are recommended but they are extremely long and/or dense

Arbital Postmortem – Arbital was a rationalist startup designed to solve online explanation. Arbital is being shut down, though the site is still online. The founder and main coder of arbital talks about why the site failed, the need to get users more aggressively, attempted pivots and working closely with Eliezer.

Announcement Ai Alignment Prize Winners And Next Round by cousin_it et all – The six winning essays in the AI alignment writing contest. The top prize was given to an essay on Goodhart Taxonomies. “Detailing the possible failures that can arise when optimizing for a proxy instead of the actual goal. Goodhart’s Law is simple to understand, impossible to forget once learned, and applies equally to AI alignment and everyday life. While Goodhart’s Law is widely known, breaking it down in this new way seems very valuable.”

This Review Is Not About Reviewing The Elephant In The Brain by Artir – Very, very detailed review of Robin Hanson’s book on hidden motives. Discussion of the core thesis: What exactly is the elephant in the brain? Topics: Signaling Model of Education. Medicine - The key facts to explain and which treatments empirically work. Charity - how people actually donate, the drowning child argument. Diverse thoughts on the underlying theory and which ideas count as confused concepts. [21K words - thats very long even relative to the long articles section]

Rationality Abridged by Quaerendo – “I present to you: Rationality Abridged — a 120-page nearly 50,000-word summary of “Rationality: From AI to Zombies”. Yes, it’s almost a short book. But it is also true that it’s less than 1/10th the length of the original.”

64