Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Highlights from the Claude 4 system prompt

simonwillison.net

Post reply on HN
Print view View original

21–30 of 95 posts

Re: Highlights from the Claude 4 system prompt

#21

Post by craigmccaskill »

loufe wrote: post #20
handfuloflight wrote: post #7

Claude 4 overindexes on portraying excitement at the slightest opportunity, particularly with the injection of emojis. The calm and collected manner of Claude prior to this is one of the major reasons why I used it over ChatGPT.

The use of emojis is a source of small blips of anger I could really do without. I'm not sure why this sort of thing isn't a toggle.

You can just add it to your master prompt. I have this added to mine, YMMV:

Tone & Communication Style

* Keep it clear, structured, and concise.

* Maintain a professional tone but make it feel natural and human; avoid robotic or overly formal language.

* Use a more conversational tone in casual, mentoring, or internal team contexts.

* Do not use em-dashes or emoji unless specifically requested.

Re: Highlights from the Claude 4 system prompt

#22

Post by jhardy54 »

triyambakam wrote: post #15

Claude recently said this to me deep in a conversation about building an app: *STOP READING THIS.* *OPEN YOUR CODE EDITOR.* *TYPE THE FIRST LINE OF CODE.* *THE TRANSACTION MACHINE BUILDS ITSELF ONE LINE AT A TIME.* *BUILD IT NOW.*

Smells like basilisk to me.

Re: Highlights from the Claude 4 system prompt

#23

Post by blrboee »

We want people to know that they’re interacting with a language model and not a person. But we also want them to know they’re interacting with an imperfect entity with its own biases and with a disposition towards some opinions more than others. Importantly, we want them to know they’re not interacting with an objective and infallible source of truth. This is exactly why I use claude over chatgpt. Chatgpt quickly started acting like my friend calling me 'bro', 'dude', 'oh man, thats true' language. which i liked on first day and became weird later on.

Re: Highlights from the Claude 4 system prompt

#24

Post by buu700 »

triyambakam wrote: post #15

Claude recently said this to me deep in a conversation about building an app: *STOP READING THIS.* *OPEN YOUR CODE EDITOR.* *TYPE THE FIRST LINE OF CODE.* *THE TRANSACTION MACHINE BUILDS ITSELF ONE LINE AT A TIME.* *BUILD IT NOW.*

16 years ago, when my friend and I had some downtime while working at a summer camp, we got bored and put together a crazy "AI"-assisted* essay that was filled with wild ranting, references to people at camp, and other camp inside jokes, and emailed it to the main office with the subject "VERY IMPORTANT: Surrealism in the Works of Koons.". It's been a bit of a running joke since then, and I had to laugh when a few of Claude 4's commit messages had the same energy.

*: The OG ChatGPT: https://en.wikipedia.org/wiki/Postmodernism_Generator

Re: Highlights from the Claude 4 system prompt

#26

Post by dirtbag »

0xTJ wrote: post #9

I'm towards the end of one paid month of ChatGPT (playing around with some code writing and also Deep Research), and one thing I find absolutely infuriating is how complimentary it is. I don't need to be told that it's a "good question", and hearing that makes me trust it less (in the sense of a sleazy car salesman, not regarding factual accuracy). Not having used LLMs beyond search summaries in the better part of a…

> I'm towards the end of one paid month of ChatGPT (playing around with some code writing and also Deep Research), and one thing I find absolutely infuriating is how complimentary it is. I don't need to be told that it's a "good question", and hearing that makes me trust it less (in the sense of a sleazy car salesman, not regarding factual accuracy).

I was also frustrated by the constant use of "You're right", "Excellent question" and similar at the start of responses. It took me a while to find effective custom instructions to prevent those, for example "Do not compliment me" didn't really work.

I ended up going with this which has worked so far: "Begin each response with substantive content. The first sentence must omit second-person pronouns ('you', 'your') and evaluative terms (e.g. 'right', 'great').".

Re: Highlights from the Claude 4 system prompt

#28

Post by SequoiaHope »

blrboee wrote: post #23

We want people to know that they’re interacting with a language model and not a person. But we also want them to know they’re interacting with an imperfect entity with its own biases and with a disposition towards some opinions more than others. Importantly, we want them to know they’re not interacting with an objective and infallible source of truth. This is exactly why I use claude over chatgpt. Chatgpt quickly sta…

Chatgpt has never called me bro (nor have I ever called it bro or used the word bro in any chats) Maybe you can clear your cache or something.

Re: Highlights from the Claude 4 system prompt

#30

Post by Stagnant »

kristianp wrote: post #27

Is there a default system prompt for the API? Because these problems haven't popped up when using a simple chat through the API. Admittedly I've been using chatgpt api, not Claude, but similar principles may apply.

No default system prompt in the API. There are some topics I much prefer chatting with the API due to system prompt of the web front-end being so restrictive (eg. song lyrics). In general I recommend people to try the API with no system prompt to more accurately see what the default tone of a model is.
Post reply on HN