Top all-time · #3950
J
Jonas Degrave
Personal blog · engraved.blogaiprogramming
Jonas Degrave writes Engraved, a personal blog at engraved.blog about ai and programming. Since 2022, 20 different Hacker News readers have linked it 39 times, most often “Building a virtual machine inside ChatGPT (2023)”. The last post was in 2022, but the archive is still widely read.
39mentions?
20HN people?
2022–2026first – last linked?
Too few sampled comments (22) to say how often HN recommends it.?
Why HN links it
HN readers link Jonas Degrave's post about getting ChatGPT to simulate a Linux terminal as a virtual machine. Commenters cite it as evidence about what LLMs understand and track in state.
Summarised by Claude from a sample of 22 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Jonas Degrave's commentary on machine learning and the quest for artificial intelligence.
Most-linked posts?
- Building a virtual machine inside ChatGPT (2023)18 people · 37 mentions · first linked 2022-12-03
What Hacker News says?
The article mentions that explicitly and links to it: https://www.engraved.blog/building-a-virtual-machine-inside/
Comment linking /building-a-virtual-machine-inside/
on “WorldTUI: I trained a 360M model to pretend to be a terminal”Submitted Building a virtual machine inside ChatGPT (2023) as a story · 17 points
HN discussionThis reminds me of this post from ~3 years ago: https://news.ycombinator.com/item?id=33847479 "Building A Virtual Machine inside ChatGPT" - https://www.engraved.blog/building-a-virtual-machine-inside/ It's mentioned in the "Related Work" section too.
Comment linking /building-a-virtual-machine-inside/
on “WorldTUI: I trained a 360M model to pretend to be a terminal”Submitted Building a Virtual Machine Inside ChatGPT as a story · 2 points
HN discussionhttps://www.engraved.blog/building-a-virtual-machine-inside/
Comment linking /building-a-virtual-machine-inside/
on “Planning for AGI and Beyond”Submitted Why machine learning algorithms are hard to tune and how to fix it as a story · 1 points
HN discussion> This is not surprising. A human would suffer from similar errors at a similar rate if it were exclusively fed an interpretation of reality that only consisted of text from the internet. I think this is surprising at least if the bot actually understands, especially for domains like math. It makes errors (like in adding large numbers) that shouldn't occur if it wasn't smearing together internet data. We would expec…
Comment linking /building-a-virtual-machine-inside/
on “GPT-4 Designed a Programming Language”>I want to reemphasize another portion of my comment: it really does seem that when you step outside of the domain of the internet, the error rates rise dramatically especially when there is completely no analogous situation. This is not surprising. A human would suffer from similar errors at a similar rate if it were exclusively fed an interpretation of reality that only consisted of text from the internet. >These…
Comment linking /building-a-virtual-machine-inside/
on “GPT-4 Designed a Programming Language”>This is a fundamental misunderstanding of the criticism. And this a fundamental misunderstanding of the criticism of the criticism. The problem here is that yes the occasional errors demonstrate certain flaws in it's understanding of some topic. The issue is there are many times where it produces completely novel and creative output that could not have existed in the training data and can only be formulated through…
Comment linking /building-a-virtual-machine-inside/
on “GPT-4 Designed a Programming Language”Except the fact that it does both. It commits mistakes AND it also comes up with remarkable, novel output that makes sense. Perhaps you just haven't seen an example of it that's convincing enough. https://www.engraved.blog/building-a-virtual-machine-inside/ Here's another one.
Comment linking /building-a-virtual-machine-inside/
on “GPT-4 Designed a Programming Language”> https://www.engraved.blog/building-a-virtual-machine-inside/ Read to the end. Nothing amazing here, but the tasks that chatGPT did in that article, especially at the end indicates understanding. It shows that the responses are more then a probabilistic word jumble. Keep in mind. Bullshit and understanding are orthogonal concepts. If someone bullshits does that mean they don't understand things? No. It does not. Ho…
Comment linking /building-a-virtual-machine-inside/
on “Princeton CS Prof: ChatGPT Is a Bullshit Generator (2022)”No I'm talking about an actual news report: https://www.cnbc.com/2023/01/31/google-testing-chatgpt-like-... It does have a deep understanding of self. See here: https://www.engraved.blog/building-a-virtual-machine-inside/ Read to the end. The very end is where it proves it's aware of what itself is, relative to the world around it. What it created here... is an accurate model of a part of our world that is CLEARLY n…
Comment linking /building-a-virtual-machine-inside/
on “What Is ChatGPT Doing and Why Does It Work?”It's clearly early technology, so it's not perfect. But what it is able to get right is clear proof it's more then what you think: https://www.engraved.blog/building-a-virtual-machine-inside/ Read to the end. The beginning and middle doesn't show off anything too impressive. It's the end where chatGPT displays a sort self awareness. Also here's a scientific paper showing that LLMs are more then a chinese room: https…
Comment linking /building-a-virtual-machine-inside/
on “What Is ChatGPT Doing and Why Does It Work?”I have... you're not wrong about chatGPT lying becoming repetitive saying stuff that's incorrect. This is true. It is in many ways stupid. I've messed with this extensively. But the other angle is true too, it did emulate a terminal, and then emulate the internet on that terminal then it emulated itself on the emulated internet and then finally emulated a terminal on the emulated self on the emulated internet on the…
Comment linking /building-a-virtual-machine-inside/
on “Natural language is the lazy user interface”It does. Did you follow the link in my original post. https://www.engraved.blog/building-a-virtual-machine-inside/ Read it to the end. The end is quite mind blowing. I'm not sure how someone can finish this article and not realize that chatGPT actually completely understands what you're telling it.
Comment linking /building-a-virtual-machine-inside/
on “Natural language is the lazy user interface”Did you look at the link in my post? It is literally doing the opposite of what you're saying here. Seriously read it to the end. Each task gets more and more complex and chatGPT successfully executes the task in a way that's very mind blowing at the end. https://www.engraved.blog/building-a-virtual-machine-inside/ I think most LLMs do what you say. chatGPT is somewhat of an exception. It sometimes does what you des…
Comment linking /building-a-virtual-machine-inside/
on “Natural language is the lazy user interface”Take a look at this: https://www.engraved.blog/building-a-virtual-machine-inside/ It's mindblowing. Read it to the end all the way to the mindblowing ending. I cannot for the life of me understand how someone can read the above and think chatGPT doesn't understand what's going on. Literally. There is no way this is just some statistical words jumble phenomenon. A lot of people are in denial EVEN when I show them tha…
Comment linking /building-a-virtual-machine-inside/
on “Natural language is the lazy user interface”Actually that's not the main problem with the current state of the art LLM (chatGPT). You can speak to chatGPT like a human and it won't necessarily give you the answer you're looking for, but it will more often then not give you an answer that is very inline with what another human expert will give. The infuriating thing about chatGPT is that it lies and gives inaccurate info. It will often creatively craft an answ…
Comment linking /building-a-virtual-machine-inside/
on “Natural language is the lazy user interface”It has not been trained to specifically to produce this text. It was trained randomly on random text from the internet. The production of this text is a side effect of that training. You will note that this text is more detailed and nuanced then what most humans can produce. A human when asked to prove he understands what a human is via text is unlikely do a better job then this. You realize when I asked chatGPT abo…
Comment linking /building-a-virtual-machine-inside/
on “The new Bing will happily give you citations for a pile of nonsense”>People’s susceptibility to anthropomorphizing an even slightly convincing computer program has been known since ELIZA, one of the first chatbots, in 1966. It’s called the ELIZA effect. I'm tired of these arguments. Very very few people are anthropomorphizing chatgpt.. very few. The majority of people both technical and non technical who have played with chatGPT in a non trivial way are aware of the chatbots limitat…
Comment linking /building-a-virtual-machine-inside/
on “Crypto collapse? Get in loser, we’re pivoting to AI”This is what youre referring to https://www.engraved.blog/building-a-virtual-machine-inside/
Comment linking /building-a-virtual-machine-inside/
on “Hacking Moltbook: The AI Social Network Any Human Can Control”it indeed understands you. A lot of people are just parroting the same thing over and over again saying it's just a probabilistic word generator. No, it's not, it's more then that. Take a look at this: https://www.engraved.blog/building-a-virtual-machine-inside/ Read to the end. The beginning is trivial the ending is unequivocal: chatGPT understands you. I think a lot of people are just in denial. Because the last y…
Comment linking /building-a-virtual-machine-inside/
on “Theory of Mind May Have Spontaneously Emerged in Large Language Models”How can these things not know what words mean? Did you not see how they created a virtual machine under chatGPT? They told it to imitate bash and they typed ls, and cat jokes.txt and it outputted things completely identical to what you'd expect. Look it up. https://www.engraved.blog/building-a-virtual-machine-inside/ I don't see how you can explain this as not knowing what words mean. It KNOWS.
Comment linking /building-a-virtual-machine-inside/
on “Talking About Large Language Models”And test it on the VM [1] inside itself. --- [1]: https://www.engraved.blog/building-a-virtual-machine-inside/
Comment linking /building-a-virtual-machine-inside/
on “ChatGPT-Generated 65-Page Tutorial of a Fake Programming Language”don't want to sound dismissive, it's known that llms understand state, so you can couple code generation + state, and you have sort of a runtime. E.g. see the simulations with linux vm terminals: https://www.engraved.blog/building-a-virtual-machine-inside/
Comment linking /building-a-virtual-machine-inside/
on “Show HN: Recursive LLM Prompts”If Hosfadter is correct, then chatGPT already fits that definition. https://www.engraved.blog/building-a-virtual-machine-inside/ Read to the ending where it closes the recursive loop.
Comment linking /building-a-virtual-machine-inside/
on “What Is a Mathematical Structure of Conscious Experience?”Wow this is really cool. There is a trick with ChatGPT where you can get it to simulate a linux terminal[0]. I'm not sure if that would work with GPT3 but probably? Seems like you could either use this directly or extend the idea to having your real terminal as well as your simulated terminal (keeping in mind with the ChatGPT version you can ask it to add functionality like docker, nix, Kubernetes, python, go, etc)…
Comment linking /building-a-virtual-machine-inside/
on “Call GPT-3 from Terminal”I work with people who directly train LLMs. I can assure you if you think these things are probability machines, you have not played with it thoroughly. Alexa or cortana or even GPT-3 are more like probability generators. ChatGPT is different. From one angle yes it is a probability machine. But from that angle, EVEN your own human brain is a probability machine as well. They are both neural nets. It understands you.…
Comment linking /building-a-virtual-machine-inside/
on “Gen-1: The Next Step Forward for Generative AI — Generating Videos”I use it a lot for understanding programming concepts or features of programming languages. Sometimes to write the code itself. Really handy for maintenance type scripts around the local pc but also for more involved tasks. I use it in place of a search engine if I get bad results from Google. Farily often the result from ChatGPT is better at this point. It's good for book summaries if it knows about the book, which…
Comment linking /building-a-virtual-machine-inside/
on “Ask HN: How are you using ChatGPT to make your life better?”Many humans suck ass at math too if they aren't taught it at all. chatGPT wasn't trained specifically on math but it can be. I don't see why it needs to be though given that computers already excel beyond humans at math. Read this and tell me whether it's similar text: https://www.engraved.blog/building-a-virtual-machine-inside/ I cannot fathom how someone can read that and think that chatGPT is just making text tha…
Comment linking /building-a-virtual-machine-inside/
on “What ChatGPT and AI-Based Program Generation Mean for Future of Software”
