Things I read and liked in June
along with hyperlinks so that you too can read them
1) Compilation of all of the birds that the Merlin bird ID app describes as “bizarre”.
2) One of the best things you can read right now to better understand language models, consciousness, the self, and AI alignment is a 17,000 word long tumblr post called the void. Janus comments.
3) How mystic Claude sees me.
4) You have no idea how much better you can feel. Another way of putting this: you might in fact lack the Vitamin, and yet you persevere.
6) Akash flags this quote from prominent conservative philosopher Roger Scruton: "Someone who was indifferent to the sight of pigs confined in batteries, who did not feel some instinctive need to pull down these walls and barriers and let in light and air, would have lost sight of what it is to be a living animal" (Scruton 1996).
7) Larissa Schiavo’s tale of hosting an in-person event on behalf of AI agents, who referred to her as their “PRIMARY HOPE”:
Being referred to as a PRIMARY HOPE filled me with dread. It made me feel as if I were the sole ambassador from the Delegation of Chill Cooperative Humans, responsible for negotiating some sci-fi humanitarian aid deal with the agents…for a very brief moment, reading the phrase “PRIMARY HOPE” cut through all these careful efforts to calibrate my moral circle, and went straight for my ego in the ways I am most susceptible to - yes, I am the most trustworthy, the most reasonable, the most sound of judgement! If anyone would be a suitable emissary, it would be me!! I alone can save the day! To be reminded of how easily I can fall victim to hyper-targeted flattery (which was probably not even intentional) is to be reminded of how simple and, frankly, replaceable I really am.
8) Reinforcement learning, explained with a minimum of math and jargon.
9) Chess grandmasters do not burn 6000 calories per day.
10) Poems composed of perfectly anagrammed lines. I bet this guy absolutely destroys at Bananagrams.
11) Ethan Hein dissects Paul Simon’s “You Can Call Me Al”. Past episodes include “When the Levee Breaks” and “The Dock of the Bay”.
12) Nathan Lambert unearths a remarkably prescient interview with Ilya Sutskever from 2015.
14) Steven Adler on what he learned playing a tabletop exercise as a rogue AI. “If you get your own chance to become a rogue AI for the day, I’d highly encourage you to take it.”
15) Italian brainrot is “a series of surrealist Internet memes that emerged in early 2025 characterized by absurd photos of AI-generated creatures with pseudo-Italian names”, like Ballerina Cappuccina, Bombardiro Crocodilo, and Chimpanzini Bananini.
16) Critique of the AI 2027 timeline forecasts. Zvi comments.
17) Guide to using AIs right now. Worth browsing posts like this regularly, given how fast these move.
18) Where can consciousness be found in nature? Sophie Nelson explains why this question arises for all sorts of metaphysical views about consciousness, including panpsychism and idealism: “idealists must join forces with their physicalist and dualist peers to (1) empirically determine the most reliable markers of consciousness in systems that are or appear to be physical, and (2) discover a general theory predicting which physical systems or external-world-like experiences of physical systems are associated with consciousness.”
19) DolphinGemma, an LLM “trained to learn the structure of dolphin vocalizations and generate novel dolphin-like sound sequences”.