GraemeStuart @ggstuart

The time for feel-good posts about brave protestors throwing off their hijabs in Iran seems long over. We appear to be in the “mass murder of civilians” phase of this revolution. Iran is ruled by psychopathic theocrats. I am not egging anyone on from the sunshine and safety of California. This is life and death struggle in the streets. But to those brave souls who know the risks and STILL return to the streets: I am in awe of your courage amidst your staggering losses. I can’t imagine what you are going through. Yet it is the inspiration of the world to all those everywhere who are still awake and alive. May the great powers outside your borders be inspired to use their tools and levers to help you in your bid to throw off your own murderous tyrants. Humanity First. Good luck.

I’m going to throw a bit of a long ball here, but it seems timely. The AI discourse has stalled (and festered) because of at least two category errors. 1. AI cannot be properly governed by either the Market or the State. There is a lot necessary to establish this, but for now, I’ll simply propose it. Fortunately, we have been operating around half a millennium with an increasing blind spot that has led us to believe that these two were the only choices. They are not. The third category (and, as it turns out, the more fundamental) is sometimes called the Commons. It is this that is proper to governance of AI. (You might note the connection to AI unemployment). Another name for the Commons is the Church. Again, a lot is behind this, but here I’ll just say that unless and until we recognize this fact and, thereby, revive the proper form and responsibility of the Church, things will continue to unravel. So, yes, I’m a very much saying that the Church is the (only) possible way to govern AI. Look forward to this discussion! 2. AI cannot be aligned with humanity. What I’m about to say will be esoteric for those who are not familiar with the notions, but they are quite precise. AI cannot be aligned with “humanity” because alignment must be at the level of the Soul and “humanity” has no Soul. Fortunately, each individual human has a Soul, and so a fully peer to peer AI can be aligned. Intimacy and care at the very personal level.

"Move 37" is the word-of-day - it's when an AI, trained via the trial-and-error process of reinforcement learning, discovers actions that are new, surprising, and secretly brilliant even to expert humans. It is a magical, just slightly unnerving, emergent phenomenon only achievable by large-scale reinforcement learning. You can't get there by expert imitation. It's when AlphaGo played move 37 in Game 2 against Lee Sedol, a weird move that was estimated to only have 1 in 10,000 chance to be played by a human, but one that was creative and brilliant in retrospect, leading to a win in that game. We've seen Move 37 in a closed, game-like environment like Go, but with the latest crop of "thinking" LLM models (e.g. OpenAI-o1, DeepSeek-R1, Gemini 2.0 Flash Thinking), we are seeing the first very early glimmers of things like it in open world domains. The models discover, in the process of trying to solve many diverse math/code/etc. problems, strategies that resemble the internal monologue of humans, which are very hard (/impossible) to directly program into the models. I call these "cognitive strategies" - things like approaching a problem from different angles, trying out different ideas, finding analogies, backtracking, re-examining, etc. Weird as it sounds, it's plausible that LLMs can discover better ways of thinking, of solving problems, of connecting ideas across disciplines, and do so in a way we will find surprising, puzzling, but creative and brilliant in retrospect. It could get plenty weirder too - it's plausible (even likely, if it's done well) that the optimization invents its own language that is inscrutable to us, but that is more efficient or effective at problem solving. The weirdness of reinforcement learning is in principle unbounded. I don't think we've seen equivalents of Move 37 yet. I don't know what it will look like. I think we're still quite early and that there is a lot of work ahead, both engineering and research. But the technology feels on track to find them. youtube.com/watch?v=HT-UZk…

Have you ever had a "gut feeling" that was right? Or just knew something without knowing how you know? That’s your intuition—and most people either trust it blindly or dismiss it completely—because they don’t know how it works. But it’s too powerful not to understand it…🧵

Flying back from Mexico tonight after a huge family holiday culminating in the eclipse at Mazatlan. We drove 2,647km and had many adventures along the way.

I think a lot of our problems today are fundamentally semantical. People mean very different things when they say words like woman, or capitalism. If we could agree on our terminology I think there would be a lot less fighting.

Sometimes the quick simple solution is best.

Focus on the green dot You will see one or more of the yellow dots vanish and reappear. The dots themselves haven't gone, but what you experience is a struggle for supremacy between the left and right side of the brain. Source whyevolutionistrue.com/2014/09/02/why…

Makes sense to me. #Python #programming

World map according to fish

Young punks in Communist Hungary - Budapest, 1982

Fractals ggstuart.github.io/lsystems/#

I done a website. #js #htmlcanvas #fractal #lsystems ggstuart.github.io/lsystems

@snapey1979 I see what you're saying. Perhaps responsibility is unrelated to covid testing? It seems the strategy is to go back to normal and keep an eye on it in case something changes. Seems broadly reasonable.

@snapey1979 What's absurd about it? I'm no fan of the prime minister but the metaphor seems perfectly reasonable to me. The concern is a new variant may arise. There is no sign of it on the horizon, but it is a potential danger. Scanning the horizon is literally what the crows nest is for.

Web design 1990s nostalgia McDonald's in 1996 KFC in 1999 Pepsi in 1996 Pizza Hut in 1996 #WebDesignHistory

keyboard_arrow_left Previous keyboard_arrow_right Next

Year/Month/Day makes sense because it narrows down from large time unit to small. Day/Month/Year makes sense because it broadens out from small time unit to large. The US convention, Month/Day/Year, makes no sense at all, is indefensible, and should be abandoned.

The James Webb telescope is heading to orbit at 1 mile per sec. I hope I’m not the only person who rates this the most important event of 2021. Despite all our justified misgivings, there really are good reasons to take pride in being human. go.nasa.gov/3mwONba

@classpert this link is broken

“To do evil, a human being must first of all believe that what he’s doing is good.” #Solzhenitsyn