Comment by jaccola
10 days ago
All of the latest models I've tried actually pass this test. What I found interesting was all of the success cases were similar to:
e.g. "Drive. Most car washes require the car to be present to wash,..."
Only most?!
They have an inability to have a strong "opinion" probably because their post training, and maybe the internet in general, prefer hedged answers....
Here’s my take: boldness requires the risk of being wrong sometimes. If we decide being wrong is very bad (which I think we generally have agreed is the case for AIs) then we are discouraging strong opinions. We can’t have it both ways.
Last year's models were bolder. Eg. Sonnet-3.7(thinking), 10 times got it right without hedging:
>You should drive your car to the car wash. Even though it's only 50 meters away (which is very close), you'll need your car physically present at the car wash to get it washed. If you walk there, you'll arrive without your car, which wouldn't accomplish your goal of getting it washed.
>You'll need to drive your car to the car wash. While 50 meters is a very short distance (just a minute's walk), you need your car to actually be at the car wash to get it washed. Walking there without your car wouldn't accomplish your goal!
etc. The reasoning never second-guesses it either.
A shame they're turning it of in 2 days.
[flagged]
You know what they mean by opinions. Policing speech like this is always counter productive.
yet the llms seem to be extremely bold when they are completely wrong (two Rs in strawberry and so on)
> They have an inability to have a strong "opinion" probably
What opinion? It's evaluation function simply returned the word "Most" as being the most likely first word in similar sentences it was trained on. It's a perfect example showing how dangerous this tech could be in a scenario where the prompter is less competent in the domain they are looking an answer for. Let's not do the work of filling in the gaps for the snake oil salesmen of the "AI" industry by trying to explain its inherent weaknesses.
Presumably the OP scare quoted "opinion" precisely to avoid having to get into this tedious discussion.
this example worked in 2021, it's 2026. wake up. these models are not just "finding the most likely next word based on what they've seen on the internet".
Well, yes, definitionally they are doing exactly that.
It just turns out that there's quite a bit of knowledge and understanding baked into the relationships of words to one another.
LLMs are heavily influenced by preceding words. It's very hard for them to backtrack on an earlier branch. This is why all the reasoning models use "stop phrases" like "wait" "however" "hold on..." It's literally just text injected in order to make the auto complete more likely to revise previous bad branches.
The person above was being a bit pedantic, and zealous in their anti-anthropomorphism.
But they are literally predicting the next token. They do nothing else.
Also if you think they were just predicting the next token in 2021, there has been no fundamental architecture change since then. All gains have been via scale and efficiency optimisations (not to discount that, an awful lot of complexity in both of these)
5 replies →
Unless LLMs architecture have changed, that is exactly what they are doing. You might need to learn more how LLMs work.
10 replies →
Did you try several times per model? In my experience it's luck of the draw. All the models I tried managed to get it wrong at least once.
The models that had access to search got ot right.But, then were just dealing with an indirect version of Google.
(And they got it right for the wrong reasons... I.e this is a known question designed to confuse LLMs)
I guess it didn’t want to rule out the existence of ultra-powerful water jets that can wash a car in sniper mode.
I enjoyed the Deepseek response that said “If you walk there, you'll have to walk back anyway to drive the car to the wash.”
There’s a level of earnestness here that tickles my brain.
Opus 4.6 answered with "Drive." Opus 4.6 in incognito mode (or whatever they call it) answered with "Walk."
They pass it because it went viral a week ago and has been patched
I tried with Opus 4.6 Extended and it failed. LLMs are non deterministic so I'm guessing if I try a couple of times it might succeed.
>Only most?!
There is such a thing as "mobile car wash" where they come to you, so "most" does seem appropriate.
Right, I use it all the time.
There are car wash services that will come to where your car is and wash it. It’s not wrong!
Kind of like this: https://xkcd.com/1368/
And it is the kind of things a (cautious) human would say.
For example, that could be my reasoning: It sounds like a stupid question, but the guy looked serious, so maybe there are some types of car washes that don't require you to bring your car. Maybe you hand out the keys and they pick your car, wash it, and put it back to its parking spot while you are doing your groceries or something. I am going to say "most" just to be sure.
Of course, if I expected trick questions, I would have reacted accordingly, but LLMs are most likely trained to take everything at face value, as it is more useful this way. Usually, when people ask questions to LLMs they want an factual answer, not the LLM to be witty. Furthermore, LLMs are known to hallucinate very convincingly, and hedged answers may be a way to counteract this.
> Most car washes... I read it as slight-sarcasm answer
There are mobile car washes that come to your house.
Do they involve you walking to them first?
You could, but presumably most people call. I know of such a place. They wash cars on the premises but you could walk in and arrange to have a mobile detailing appointment later on at some other location.
That still requires a car present to be washed though.
but you can walk over to them and tell them to go wash the car that is 50 meters away. no driving involved.
1 reply →
opus 4.6 extended still fails.
> Only most?!
I mean I can imagine a scenario where they have pipe of 50m which is readily available commercially?
> Only most?!
What if AI developed sarcasm without us knowing… xD
That's the problem with sarcasm...
Sure it did.
Once I asked ChatGPT "it takes 9 months for a woman to make one baby. How long does it take 9 women to make one baby?". The response was "it takes 1 month".
I guess it gives the correct answer now. I also guess that these silly mistakes are patched and these patches compensate for the lack of a comprehensive world model.
These "trap" questions dont prove that the model is silly. They only prove that the user is a smartass. I asked the question about pregnancy only to to show a friend that his opinion that LLMs have phd level intelligence is naive and anthropomorphic. LLMs are great tools regardless of their ability to understand the physical reality. I don't expect my wrenches to solve puzzles or show emotions.