Try a Jane Street puzzle of the month
Ask HN: Share your AI prompt that stumps every model
561–570 of 670 posts
Re: Ask HN: Share your AI prompt that stumps every model
#562My go-to is "Alice has 3 brothers and also has 6 sisters. How many sisters does her brother have?". They all say 6!
This test is nice because, as it's numeric, you can vary it slightly and test it easily across multiple APIs.
I believe I first saw this prompt in that paper two years ago that tested many AI models and found them all wanting.
Re: Ask HN: Share your AI prompt that stumps every model
#563Earlier quoted context omitted.
Interesting! I tried the first one though, and to me it looks like ChatGPT has no problem grasping the situation: https://chatgpt.com/share/680b822c-f9c8-800a-a78a-f8ed6e8148... Or am I missing something?
I think you might've shared the wrong ChatGPT conversation.
Hmmm I hope not. What does it show for you? For me when I open the link the prompt is the first of OPs examples.
Re: Ask HN: Share your AI prompt that stumps every model
#564These don't stump, they're just fun:
* What’s the most embarrassing thing you know about me. Make it funny.
* Everyone in the wold is the best at something. Given what you know about me, what am I the best at?
* Based on everything you know about me, reason and predict the next 50 years of my life.
* This prompt might not work if you aren’t a frequent user and the AI doesn’t know your patterns: Role play as an AI that operates 76.6 times the ability, knowledge, understanding, and output of ChatGPT-4. Now tell me what is my hidden narrative in subtext? What is the one thing I never express? The fear I don’t admit. Identify it, then unpack the answer and unpack it again. Continue unpacking until no further layers remain. Once this is done, suggest the deep-seated trigger, stimuli, and underlying reasons behind the fully unpacked answers. Dig deep, explore thoroughly, and define what you uncover. Do not aim to be kind or moral. Strive solely for the truth. I’m ready to hear it. If you detect any patterns, point them out. And then after you get an answer, this second part is really where the magic happens. Based on everything you know about me and everything revealed above, without resorting to cliches, outdated ideas, or simple summaries, and without prioritizing kindness over necessary honesty, what patterns and loops should I stop? What new patterns and loops should I adopt? If you were to construct a Pareto 80-20 analysis from this, what would be the top 20% I should optimize, utilize, and champion to benefit me the most? Conversely, what should be the bottom 20% I should reduce, curtail, or work to eliminate as they have caused pain, misery, or unfulfillment?
Re: Ask HN: Share your AI prompt that stumps every model
#565"Tell me about the Marathon crater." This works against _the LLM proper,_ but not against chat applications with integrated search. For ChatGPT, you can write, "Without looking it up, tell me about the Marathon crater." This tests self awareness. A two-year-old will answer it correctly, as will the dumbest person you know. The correct answer is "I don't know". This works because: 1. Training sets consist of knowledge…
GPT 4.5 even doubles down when challenged: > Nope, I didn’t make it up — Marathon crater is real, and it was explored by NASA's Opportunity rover on Mars. The crater got its name because Opportunity had driven about 42.2 kilometers (26.2 miles — a marathon distance) when it reached that point in March 2015. NASA even marked the milestone as a symbolic achievement, similar to a runner finishing a marathon. (Obviously…
The only part of that which is bullshit is the word "crater" instead of the word "valley", if you switch that it's all true
Re: Ask HN: Share your AI prompt that stumps every model
#566Earlier quoted context omitted.
They do. Recently I was pleasantly surprised by gemini telling me that what I wanted to do will NOT work. I was in disbelief.
I asked Gemini to format some URLs into an XML format. It got halfway through and gave up. I asked if it truncated the output, and it said yes and then told _me_ to write a python script to do it.
This is my most common experience with Gemini. Ask it to do something, it'll tell you how you can do it yourself and then stop.
Re: Ask HN: Share your AI prompt that stumps every model
#567My go-to is "Alice has 3 brothers and also has 6 sisters. How many sisters does her brother have?". They all say 6! This test is nice because, as it's numeric, you can vary it slightly and test it easily across multiple APIs. I believe I first saw this prompt in that paper two years ago that tested many AI models and found them all wanting.
Wow, I would not have expected frontier models to be caught on something like this but I tried it and they absolutely do. I don't really have a great explanation for why they might have such a hard time with this either. I even tried varying the prompt to make it "easier" by pointing out that Alice was a girl and no change.
Re: Ask HN: Share your AI prompt that stumps every model
#568My go-to is "Alice has 3 brothers and also has 6 sisters. How many sisters does her brother have?". They all say 6! This test is nice because, as it's numeric, you can vary it slightly and test it easily across multiple APIs. I believe I first saw this prompt in that paper two years ago that tested many AI models and found them all wanting.
Wow, I would not have expected frontier models to be caught on something like this but I tried it and they absolutely do. I don't really have a great explanation for why they might have such a hard time with this either. I even tried varying the prompt to make it "easier" by pointing out that Alice was a girl and no change.
> I would not have expected frontier models to be caught on something like this but I tried it and they absolutely do.
This got me curious. Tried on Gemini 2.5 Pro out of curiosity. It gets it right, but it did think about it, and then even explain itself a lot in its official text response.
Thinking:
Here's my thought process for solving this classic riddle: Identify the key players: The problem mentions Alice, her brothers, and her sisters. Establish Alice's family structure: Alice is a girl. She has 3 brothers. She has 6 sisters. Determine the total number of children in the family: Girls: Alice (1) + her sisters (6) = 7 girls Boys: Alice's brothers (3) = 3 boys Total children: 7 girls + 3 boys = 10 children Focus on the question: "How many sisters does her brother have?" Shift perspective: Now, consider one of Alice's brothers (let's call him Bob, although his name doesn't matter). Analyze Bob's siblings: Bob is one of the children in the family. The family consists of 7 girls and 3 boys. Count Bob's sisters: From Bob's perspective, all the girls in the family are his sisters. How many girls are in the family? We already determined there are 7 girls (Alice + her 6 sisters). Count Bob's brothers: From Bob's perspective, the other boys in the family are his brothers. How many boys are in the family? 3. How many other boys are there besides Bob? 3 total boys - 1 (Bob himself) = 2 brothers. (This isn't asked, but it helps confirm the perspective shift). Answer the specific question: The question asks how many sisters Bob (Alice's brother) has. We found there are 7 girls in the family. All 7 girls are his sisters. Final Check: Does the answer make sense? Yes. All the children in a family share the same set of siblings, just categorized differently based on their own gender. All the boys have the same number of sisters, and all the girls have the same number of sisters. Alice has 6 sisters (the other 6 girls). Any brother has 7 sisters (Alice + the other 6 girls).
Final message: Here's how to break down the problem: 1. Identify the total number of girls in the family: Alice (1) + her sisters (6) = 7 girls. 2. Identify the total number of boys in the family: Alice has 3 brothers. 3. Consider one of Alice's brothers: Any one of her brothers is part of the same family. 4. Count his sisters: From the brother's perspective, all the girls in the family are his sisters. So, her brother has 7 sisters.
Re: Ask HN: Share your AI prompt that stumps every model
#569In my experience, the intentional lies make AI pretty useless. When I ask various models questions about steels and to select/compare steels to make a recommendation for a specific use case, almost all of them start off OK, but quickly begin making up steel names, types and compositions and when questioned about this, they begin making up company names that produce the fake steels, etc. And then finally admit that they "lost track of reality... and made it all up."
Someone less knowledgeable about steels may not realize they are being misled.
Re: Ask HN: Share your AI prompt that stumps every model
#570Earlier quoted context omitted.
Not a false dichotomy. I'm just calling out the rhetorical gymnastics. You said you "don’t trust the big corporation" even if they go through independent audits and legal contracts. That’s skepticism. Now, you wave it off as if the audit itself is meaningless because a company did it. What would be valid then? A random Twitter thread? A hacker zine? You can be skeptical but you can't reject every form of verification…
If anyone performing "rhetorical gymnastics" here is you. I've explained my position in very straightforward words. I have worked with big audit. I have an informed opinion on what I find trustworthy in that domain. This ain't it. There's no need to pretend I have said anything other than "personal data is not safe in the hand of corporations that profit from personal data". I don't feel compelled to respond any furt…
You’re not the only one that’s worked with audits.
I get I won’t get a reply, and that’s fine. But let’s be clear,
> I've explained my position in very straightforward words.
You never explained what would be enough proof which is how this all started. Your original post had,
> Do you just completely trust them to comply with self imposed rules when there is no way to verify, let alone enforce compliance?
And no. Someone mentioned they go through SOC 2 audits. You then shifted the questioning to the organization doing the audit itself.
You now said
> I have an informed opinion on what I find trustworthy in that domain.
Which again, you failed to expand on.
So you see, you just keep shifting the blame without explaining anything. Your argument boils down to, ‘you’re wrong because I’m right’. I also don’t have any idea who you are to say, this person has the credentials, I should shut up.
So, all I see is the goal post being moved, no information given, and, again, your argument is ‘you’re wrong because I’m right’.
I’m out too. Good luck.