Top all-time · #8237
M
— Moire
Personal blog · generative.inkaiphilosophy
— Moire is a personal blog at generative.ink about ai and philosophy. Since 2022, 10 different Hacker News readers have linked it 29 times, most often “Large Language Models as Simulators”. It is still writing: the last post was on 2024-12-16, and you can follow it by feed.
29mentions?
10HN people?
2022–2024first – last linked?
Too few sampled comments (4) to say how often HN recommends it.?
Why HN links it
HN commenters link Moire's posts on generative.ink about AI writing interfaces such as Loom and the block multiverse, and on how language models predict and generate text. The comments don't judge the writing itself.
Summarised by Claude from a sample of 4 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Pen-and-ink gives expression to rules of probability that conjoin to inconceivable complexity.
Most-linked posts?
- Large Language Models as Simulators5 people · 24 mentions · first linked 2022-09-07
- Loom: Interface to the Multiverse2 people · 2 mentions · first linked 2023-02-23
What Hacker News says?
They did make it, in 2021. https://generative.ink/posts/loom-interface-to-the-multivers... (click through to the GitHub repo and check the commit history, the bulk of commits is at least 3 years old)
Comment linking /posts/loom-interface-to-the-multiverse/
on “Show HN: I made an app to use local AI as daily driver”Submitted Large Language Models as Simulators as a story · 3 points
HN discussionAs a long time LLM enjoyer, here is the most insightful take I've seen https://generative.ink/posts/simulators/ but it's not an easy read if you don't already know some stuff about large language models. Read it if you have seen the "stochastic parrot" and "blurry jpeg" explanations and you feel like they are missing the mark.
Comment linking /posts/simulators/
on “GPT-4”Submitted Visualizing LLM memorization with block multiverse plots as a story · 1 points
HN discussionAs a long-time LLM enjoyer, by far the most insightful take I saw along these lines is https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “AI vs. AGI vs. Consciousness vs. Super-Intelligence vs. Agency”Submitted GPT Is a Simulator as a story · 1 points
HN discussion>Read this[0] for a conceptual explanation of block multiverse interface and demo video [0] https://generative.ink/meta/block-multiverse/
Comment linking /meta/block-multiverse/
on “Experimental tree-based writing interface for GPT-3”Submitted Language Models Are Multiverse Generators as a story · 1 points
HN discussionThe New Yorker keeps putting new takes on ChatGPT by science fiction authors and pop science authors. But they have the opportunity to earn my enduring respect by inviting janus from https://generative.ink/posts/simulators/ to write a GPT thinkpiece, hopefully after they get a chance to experiment with the GPT-4 or GPT-5 base model (no RLHF).
Comment linking /posts/simulators/
on “What Kind of Mind Does ChatGPT Have?”Submitted Loom: Interface to the Multiverse as a story · 1 points
HN discussionI read the first one on internet archive but I was only commenting on the title of the second one. Maybe someone will post an internet archive link I don't know how to get it. > It's also nice for a publication to have space for more than one take. It's true I'm glad they are giving it another shot. By the way, by far the most insightful one I found was https://generative.ink/posts/simulators/ but I will definitely…
Comment linking /posts/simulators/
on “What Kind of Mind Does ChatGPT Have?”Another one like this is 'loom' https://github.com/socketteer/loom which was used in the research of the simulators essay https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “NodePad: LLM-aided graph based note-taking and brainstorming tool”This is a good take. Possibly there is no better one yet that is shorter and more accessible. > Wait... Maybe It’s All Just Roleplay? As a long time LLM enjoyer my favorite analysis is https://generative.ink/posts/simulators/ which explores this idea in much more depth.
Comment linking /posts/simulators/
on “ChatGPT is Really Good at Roleplaying”I mean if you read the definitions you can see that it's a Markov chain. I don't know what else to tell you. Maybe you have the idea that only low order Markov chains, like second or third order Markov chains are "true" Markov chains and that 2048th order Markov chains shouldn't be called Markov chains any more? Maybe you think that the transition matrix must be constructed explicitly as a square of numbers stored i…
Comment linking /posts/simulators/
on “ChatGPT sometimes makes up facts. For one law prof, it went too far”Imagine trying to write a better take on GPTs than https://generative.ink/posts/simulators/ it will never happen. Anyway, Yud's main point is obviously true and many people either don't understand it or don't believe it: "GPT-4 is still not as smart as a human in many ways, but it's naked mathematical truth that the task GPTs are being trained on is harder than being an actual human." "And since the task that GPTs a…
Comment linking /posts/simulators/
on “GPTs are Predictors, not Imitators (Eliezer Yudkowsky, Apr 8 2023)”As a long-time LLM enjoyer I think the 'simulators' take is by far the best one https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “Large Language Models Can Simulate Everything”> It's a genuinely insightful summary of the last few months of activity in open source models Yes this is an amazing summary! Just for its summary alone, it is probably one of the top five writings I saw on LLMs and I read every one! > because I have trouble imagining anyone writing something this good and deciding NOT to take credit for the analysis themselves. Not everyone has a substack that they spam onto hacke…
Comment linking /posts/simulators/
on “Google “We Have No Moat, and Neither Does OpenAI””It looks similar to loom https://github.com/socketteer/loom the program that was used in the research behind the simulators essay https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “Show HN: Use ChatGPT on a 2 Dimensional Interface”I am not convinced by this argument. It is very misleading to think that, since GPT is trained on data from the world, it must, necessarily, always produce an average of the ideas in the world. Humans have formulated laws of physics that "minimize loss" on our predictions of the physical world that are later experimentally determined to be accurate, and there's no reason to assume a language model trained to minimiz…
Comment linking /posts/simulators/
on “91% of ML Models Degrade over Time”> .. we can abstract, and formalise, and prove, and demonstrate. You can't do that just by predicting. That's a power over and beyond modelling data. I think we will find this not to be the case. And of course, many people like me do think that the ability to predict well enough implies the ability to do everything else. Since I wrote my reply, I found two other posts on HN that seem relevant to that if you want to…
Comment linking /posts/simulators/
on “Understanding ChatGPT”I'm a long time LLM enjoyer and by far the best analysis I've seen for that is https://generative.ink/posts/simulators/ but it's way "too much" for normal people who are learning about stochastic parrots and blurry jpegs instead.
Comment linking /posts/simulators/
on “What Will Transformers Transform?”> "What don't you like about the [blurry jpeg] article?" It's valid at some technical level but I dislike it so much. There are so many things. If I had to pick only a few, it's that it encourages to see LLMs as a worse and static rendering of an 'ideal' piece of data that we already have. This is not how the LLMs are useful. An LLM can be at its best when it's explaining or imagining for example something like 'how…
Comment linking /posts/simulators/
on “Understand by Ted Chiang”The best way to think about LLMs since even before ChatGPT has always been the simulators essay https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “The Many Ways That Digital Minds Can Know – A Better Way to Think about LLMs”In my opinion this is a very good take, I think they are accurate and many are not widely accepted yet (but they will be). For a deeper analysis the most insightful take I saw on GPTs was published even before ChatGPT was released: https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “Things to Know about Large Language Models”Best one is https://twitter.com/repligate they wrote the simulators essay https://generative.ink/posts/simulators/ but they have veered towards schizoposting so much recently, I hope they get access to base models of GPT-4 or GPT-5 soon! If OpenAI were smart they would let them use it. Other one is https://twitter.com/emollick it's an entrepreneurship professor but he is good at prompts. I'm pretty sure the best pla…
Comment linking /posts/simulators/
on “Ask HN: What are the best Twitter accounts to follow for GPT/ChatGPT?”Here's another GPT tree interface https://github.com/socketteer/loom it was used in the research behind https://generative.ink/posts/simulators/
Comment linking /posts/simulators/
on “Conversation branching in ChatGPT”It's changing so fast and I personally don't use them, but one I saw was loom https://github.com/socketteer/loom it's a free version and I think the author currently uses a non-public continuation of the project but I'm not sure. It's the one they used in the research behind the simulators essay https://generative.ink/posts/simulators/ I guess it's not primarily a mind map tool that adds AI as an afterthought. It's…
Comment linking /posts/simulators/
on “Ask HN: What's the best tool for mind mapping with AI/ChatGPT?”As a long time LLM enjoyer I just want to mention https://generative.ink/posts/simulators/ as I think it's by far the most insightful take on the GPT LLMs even though it was from before ChatGPT. It's better than blurry jpeg and stochastic parrot etc.
Comment linking /posts/simulators/
on “ChatGPT as a Calculator for Words”I would just add that if you are an LLM enjoyer and not necessarily a normie, then https://generative.ink/posts/simulators/ is still the best take.
Comment linking /posts/simulators/
on “ChatGPT Explained: A Normie's Guide to How It Works”
