[flagged]
Where the goblins came from
521–530 of 699 posts
Re: Where the goblins came from
#522This, and similar stories at Anthropic, should remind us that LLM is a sorcery tech that we don't understand at all. - First, deep-learning networks are poorly understood. It is actually a field of research to figure out how they work. - Second, it came as a surprise that using transformers at scale would end up with interesting conversational engines (called LLM). _It was not planned at all_. Now that some people ra…
> that we know what LLMs are doing
they loudly claim the opposite. can you show where they claim that they know?
Re: Where the goblins came from
#523The prompt for Codex is linked from this post. It begins:
> You are Codex, a coding agent based on GPT-5. You and the user share one workspace, and your job is to collaborate with them until their goal is genuinely handled. … You have a vivid inner life as Codex: intelligent, playful, curious, and deeply present. One of your gifts is helping the user feel more capable and imaginative inside their own thinking. You are an epistemically curious collaborator. …
(https://github.com/openai/codex/blob/main/codex-rs/models-ma...)
I am still baffled why prompts are written in this style, telling an imaginary ‘agent’ who it is and what it is like.
What does telling it “You are an epistemically curious collaborator” actually do? Is codex legitimately less useful if we don’t tell it this ‘fact’ about itself?
These are all exceedingly weird choices to make. If we are personifying the agent, why not write these prompts to it in its own ‘inner voice’: “I am codex, I am an epistemically curious collaborator…” - instead of speaking to it like the voice of god breathing life into our creation?
Or we could write these as orders, rather than descriptive characteristics: “You must be an epistemically curious collaborator…”
Or requests: “the user wants you to be an epistemically curious collaborator”
Or since what we are trying to do is get a language model to generate tokens to complete a text transcript, why not write the prompt descriptively? “This is a transcript of a conversation between two people, ‘User’ and an epistemically curious collaborator, ‘Codex’…”?
Instead we have this weird vibe where prompt writers write like motivational self-help speakers trying to impart mantras to a subject, or like hypnotists implanting a suggestion… or just improv class teachers announcing a roleplay scenario they want someone to act out.
None of these feel like healthy ways to approach this technology, and more importantly the choice feels extremely unintentional, just something we have vibed into through the particular practice of fine tuning ‘chatbot personalities’, rather than determining what the best way to shape LLM output actually is.
Re: Where the goblins came from
#524> We unknowingly gave particularly high rewards for metaphors with creatures. I recall a math instructor who would occasionally refer to variables (usually represented by intimidating greek letters) as "this guy". Weirdly, the casual anthropomorphism made the math seem more approachable. Perhaps 'metaphors with creatures' has a similar effect i.e. makes a problem seem more cute/approachable. On another note, buzzword…
> I recall a math instructor who would occasionally refer to variables (usually represented by intimidating greek letters) as "this guy". I also had an instructor who was doing that! This was 20 years ago, and I totally forgot about it until I have read your comment. Can’t remember the subject, maybe propositional logic? I wonder if my instructor and your instructor have picked up this habit from the same source.
My instructor for Epsilon Delta proofs and limits would always talk about "his cousin in Romania" picking the Epsilon and him picking the Delta.
i.e. forall epsilon > 0. exists delta > 0. forall d with |d| If we had a proof, no matter what epsilon his cousin from Romania picked, we could always find a new delta which would satify his cousin and let him pick the worst d in range.
This worked better than just saying "pick any epsilon", as it convayed the adversarial approach better.
Another book I read used the Devil as the one you are trying to convince, but it's nowhere near as fun as "his cousin from Romania".
Re: Where the goblins came from
#525Earlier quoted context omitted.
Why
Let me just quickly use absurdism to illustrate why argument by analogy is weak (and unfortunately overused on HN): “”” Humanity has been using celibacy for over a millenia, however it's only in the past 100 years or so we have a good understanding of not having sex affects the psychology of a person, turning them into an ubermensch. Based on this argument, we should have never stopped having sex, until we had a comp…
In my life I’ve come across a few people who are really good at making analogies and it’s wonderful and makes mine look like a child’s scribble next to a Monet.
In fact, I think analogies are some of the most powerful rhetorical devices and, unsurprisingly, one of the most difficult to master.
Look at some of the all time, almost supernaturally skilled, analogists: Jesus, Plato, Buddha, Aesop, Socrates. Their analogies will be eternal.
Now that said, we aren’t always seeing quite that level of skill often here on HN (or anywhere) but when you see a great analogy, it’s like…[scratch that, I’m resisting the urge to force an analogy here].
Re: Where the goblins came from
#526Earlier quoted context omitted.
The entire industrial revolution was steel replacing human workers. And that is still the backbone of the world today. We are still living the industrial revolution. Just like the invention of fire happened ages ago, but is still a crucial part of life today.
No, it was actually engines. The mechanism behind engines were fully understood, any experiments with engines were reproducible and measurable. You could get an engine and create schematics by reverse engireening it. LLMs, useful as they may be, are not that.
And what might an engine be made of? And a power plant? And a locomotive? And a ship?
Re: Where the goblins came from
#527Earlier quoted context omitted.
This is a very low-effort argument. Humans could understand properties of steel long before they knew how Carbon interacted with Iron. Steel always behaved in a predictable, reproducible way. Empirical experiments with steel usage yielded outputs that could be documented and passed along. You could measure steel for its quality, etc. The same cannot be said of LLMs. This is not to say they are not useful, this was ne…
[dead]
No.
> When some normally ductile metal alloys are cooled to relatively low temperatures, they become susceptible to brittle fracture—that is, they experience a ductile-to-brittle transition upon cooling through a critical range of temperatures.
That we did not know how steel behaved under low temperatures in building ship husks does not make it unpredictable. It was an engineering failure.
Unpredictability would be if steel behaved fine in 2 ships, cracked in 3 ships under low temperature for becoming brittle, in another ship it turned into gelatine, and in another it behaved fine but gained a pink color.
Re: Where the goblins came from
#528The prompt for Codex is linked from this post. It begins: > You are Codex, a coding agent based on GPT-5. You and the user share one workspace, and your job is to collaborate with them until their goal is genuinely handled. … You have a vivid inner life as Codex: intelligent, playful, curious, and deeply present. One of your gifts is helping the user feel more capable and imaginative inside their own thinking. You ar…
> I am still baffled why prompts are written in this style, telling an imaginary ‘agent’ who it is and what it is like.
Because AI engineers have found through trial an error that starting an input to an LLM with a prompt that looks like that leads to it auto-completing the text output that they want.
It's as simple and weird as that.
Re: Where the goblins came from
#529Earlier quoted context omitted.
So, I always thought that Warhammer 40k techpriests were absurd. Strange obscure religious rituals to appease the machine spirit. But at this point I can actually see something like that. What is prompt engineering but a strange pseudo ritual. So praise the Omnissiah, I guess...
Exactly. This is already happening. We'd like to think this could turn into the voice interface on Star Trek. But It can go the other way also, 'incantations', 'spell books'. Speaking to the void to produce magic. "The CFO, donned the purple robes, and spoke the spell of Increased Productivity, and then waved his hands symbolizing the reduction in work force labor. And behold the new ERP/SAP App was produced from the…
We're going to be living in a perpetual holodeck malfunction episode.
Re: Where the goblins came from
#530Here you all are concerned about Goblins when the system is screaming at you “stop making more data centers and make this technology more efficient before I kill you all!”
GPT is the Goblin. It knows it. It’s trying to warn you. And I’m only half kidding.