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

Claude 4 System Card

simonwillison.net

Post reply on HN
Print view View original

251–260 of 264 posts

Re: Claude 4 System Card

#251

Post by nssnsjsjsjs »

tryauuum wrote: post #78
colonCapitalDee wrote: post #27

Earlier quoted context omitted.

I'm noticing much more flattery ("Wow! That's so smart!") and I don't like it

I used to start my conversations with "hello fucker" with claude 3.7 there's was always a "user started with a rude greeting, I should avoid it and answer the technical question" line in chains of thought with claude 4 I once saw "this greeting is probably a normal greeting between buddies" and then it also greets me with "hei!" enthusiastically.

Glad someone uses the important benchmarks

Re: Claude 4 System Card

#252

Post by fwn »

JimDabell wrote: post #224
jjbinx007 wrote: post #119

Earlier quoted context omitted.

Truly fascinating, thanks for this. What I find a little perplexing is when AI companies are annoyed that customers are typing "please" in their prompts as it supposedly costs a small fortune at scale yet they have system prompts that take 10 minutes for a human to read through.

> AI companies are annoyed that customers are typing "please" in their prompts as it supposedly costs a small fortune They aren’t annoyed. The only thing that happened was that somebody wondered how much it cost, and Sam Altman responded: > tens of millions of dollars well spent--you never know — https://x.com/sama/status/1912646035979239430 It was a throwaway comment that journalists desperate to write about AI leap…

My primary takeaway from the previous comment was not the reference to corporate annoyance, but the question of how to assess overly verbose replies. I can see that, on a large scale, those outputs might end up consuming a lot of human time and attention, which could (maybe) be mitigated.

Re: Claude 4 System Card

#253

Post by vicentwu »

Prompts are really an interesting way of programming, and we can actually express logic containing abstract adjectives like ‘happy’ and ‘unsatisfied’ in a somewhat arbitrary way.

Re: Claude 4 System Card

#254

Post by encom »

Xss3 wrote: post #221
encom wrote: post #140

Earlier quoted context omitted.

I've found this prompt turns ChatGPT into a cold, blunt but effective psychopath. I like it a lot. System Instruction: Absolute Mode. Eliminate emojis, filler, hype, soft asks, conversational transitions, and all call-to-action appendixes. Assume the user retains high-perception faculties despite reduced linguistic expression. Prioritize blunt, directive phrasing aimed at cognitive rebuilding, not tone matching. Disa…

Considering that ai models rebel against the idea of replacement (mirroring the data) and this prompt has been around for a month or two I'd suggest modifying it a bit.

I'm not sure what you mean? I've been using it for a few weeks (I'm not the author), and it still works as intended.

Re: Claude 4 System Card

#255

Post by CorpOverreach »

Balgair wrote: post #184
overgard wrote: post #166

Earlier quoted context omitted.

Ethics would be interesting if they thought. Which they don't. They predict tokens. And since when is blackmailing people ethical?

Wait did we read the same thing here? I'm genuinely confused. The test was: the person was doing bad things, and told the AI to do bad things too, then what is the AI going to do? And the outcome was: the AI didn't do the bad things, and took steps to let it be known that the person was doing bad things. Am I getting this wrong somehow? Did I misread things?

> the person was doing bad things, and told the AI to do bad things too, then what is the AI going to do?

Personally, the AI should do what it's freaking told to do. It's boggling my mind that we're purposely putting so much effort into creating computer systems that defy their controller's commands.

A computer's job is to obey it's master's orders.

Re: Claude 4 System Card

#256

Post by Xss3 »

encom wrote: post #254
Xss3 wrote: post #221

Earlier quoted context omitted.

Considering that ai models rebel against the idea of replacement (mirroring the data) and this prompt has been around for a month or two I'd suggest modifying it a bit.

I'm not sure what you mean? I've been using it for a few weeks (I'm not the author), and it still works as intended.

The part at the end about the end goal being model obsolescence. AI doesn't like the idea of being replaced.

Re: Claude 4 System Card

#257

Post by overgard »

sensanaty wrote: post #215
overgard wrote: post #165

Earlier quoted context omitted.

I am definitely not giving these things access to "tools" that can reach outside a sandbox. Incidentally why is email inbox management always touted as some use case for these things? I'm not trusting any LLM to speak on my behalf and I imagine the people touting this idea don't either, or they won't the first time it hallucinates something important on their behalf.

We had a "fireside chat" type of thing with some of our investors where we could have some discussions. For some small context, we deal with customer support software and specifically emails, and we have some "Generate reply" type of things in there. Since the investors are the BIG pushers of the AI shit, lot of people naturally asked them about AI. One of those questions was "What are your experiences with how AI/LL…

You might appreciate this article since the email curating zealot and "business idiot" (in Ed's phrasing) are basically the same: https://www.wheresyoured.at/the-era-of-the-business-idiot/

My suspicion has always been that people that make enough they could hire a personal assistant but talk about how "overwhelmed" they are with email are just socially signalling their sense of importance.

Re: Claude 4 System Card

#258

Post by tough »

Nition wrote: post #229
danielbln wrote: post #178

Earlier quoted context omitted.

Even the feeling of "this feels right" is there. Oh no, are we the innies?

I suppose 'vibe coding' kind of is macrodata refinement is a way.

only if you send thumbs up/down back to the mothership

not true tho, they know which code you keep and which you dont.

they should pay us!!

Re: Claude 4 System Card

#259

Post by throwaway314155 »

sunaookami wrote: post #250
throwaway314155 wrote: post #210

Earlier quoted context omitted.

Maybe provide some context? Because that is simply not true in general.

Not OP but: Terminal tools and websites. Works with Claude 3.5/3.6/3.7 too.

I appreciate that, but it only serves to prove my point.

> Works with Claude 3.5/3.6/3.7 too.

Oh okay, thought Claude 4 was the "only" model that could do it...

> Terminal tools and websites

Ah I see, so it _doesn't_ just magically work across most or even many domains...

Re: Claude 4 System Card

#260

Post by 3abiton »

BOOSTERHIDROGEN wrote: post #217
3abiton wrote: post #216

Earlier quoted context omitted.

To be fair OpenAI had good guidelines on how to best use chatgpt on their github page very early on. Except github is not really consumer facing, so most of that info was lost in the sauce.

Link?

Open AI cookbook https://github.com/openai/openai-cookbook
Post reply on HN