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

AI Police Reports: Year in Review

eff.org

Post reply on HN
Print view View original

101–110 of 224 posts

Re: AI Police Reports: Year in Review

#101

Post by antonvs »

throwuxiytayq wrote: post #54
opan wrote: post #42

Earlier quoted context omitted.

It's pretty similar to looking something up with a search engine, mashing together some top results + hallucinating a bit, isn't it? The psychological effects of the chat-like interface + the lower friction of posting in said chat again vs reading 6 tabs and redoing your search, seems to be the big killer feature. The main "new" info is often incorrect info. If you could get the full page text of every url on the fir…

> If you could get the full page text of every url on the first page of ddg results and dump it into vim/emacs where you can move/search around quickly, that would probably be similarly as good, and without the hallucinations. Curiously, literally nobody on earth uses this workflow. People must be in complete denial to pretend that LLM (re)search engines can’t be used to trivially save hours or days of work. The accu…

> People must be in complete denial

That seems to be a big part of it, yes. I think in part it’s a reaction to perceived competition.

Re: AI Police Reports: Year in Review

#102

Post by antonvs »

gambiting wrote: post #44
krainboltgreene wrote: post #39

Earlier quoted context omitted.

Man, what are we supposed to do with people who think the above?

I don't know, it's kinda terrifying how this line of thinking is spreading even on HN. AI as we have it now is just a turbocharged autocomplete, with a really good information access. It's not smart, or dumb, or anything "human" .

Do you think your own language processing abilities are significantly different from autocomplete with information access? If so, why?

Re: AI Police Reports: Year in Review

#103

Post by cortic »

krainboltgreene wrote: post #39
charcircuit wrote: post #37

Earlier quoted context omitted.

AI is smarter than everyone already. Seriously, the breadth of knowledge the AI possesses has no human counterpart.

Man, what are we supposed to do with people who think the above?

>ChatGPT (o3): Scored 136 on the Mensa Norway IQ test in April 2025

If you don't want to believe it, you need to change the goal posts; Create a test for intelligence that we can pass better than AI.. since AI is also better at creating test than us maybe we could ask AI to do it, hang on..

>Is there a test that in some way measures intelligence, but that humans generally test better than AI?

Answer:Thinking, Something went wrong and an AI response wasn't generated.

Edit, i managed to get one to answer me; the Abstraction and Reasoning Corpus for Artificial General Intelligence (ARC-AGI). Created by AI researcher François Chollet, this test consists of visual puzzles that require inferring a rule from a few examples and applying it to a new situation.

So we do have A test which is specifically designed for us to pass and AI to fail, where we can currently pass better than AI... hurrah we're smarter!

Re: AI Police Reports: Year in Review

#104

Post by andrepd »

logicprog wrote: post #47
simonjgreen wrote: post #34

Earlier quoted context omitted.

Do you have any links you could share to content you found especially insightful about AI use in China?

I don't know if it supports their particular point, but Machine Decision is Not Final seems like a very cool and interesting look at China's culture around AI: https://www.urbanomic.com/book/machine-decision-is-not-final...

In the West we have autonomous systems to commit genocide, detecting and murdering "enemy combatants" at scale, where "enemy combatant" is defined as "male between the ages of 15 and 55".

Sometimes I'm not so sure about any so-called moral superiority.

Re: AI Police Reports: Year in Review

#105

Post by latexr »

hombre_fatal wrote: post #94
latexr wrote: post #73

Earlier quoted context omitted.

> I'm guessing someone is gonna compare this to the old Dropbox post, but whatever. If they do, you’ll be in good company. That post is about the exact opposite of what people usually link it for. I’ll let Dan explain: https://news.ycombinator.com/item?id=27067281

Dan makes a case for being charitable to the commenter and how lame it is to neener-neener into the past, not that it has some opposite meaning everyone is missing out on.

Dan clearly references how people misunderstand not only the comment (“he didn't mean the software. He meant their YC application”) but also the whole interaction (“He wasn't being a petty nitpicker—he was earnestly trying to help, and you can see in how sweetly he replied to Drew there that he genuinely wanted them to succeed”).

So yes, it is the opposite of why people link to it (which is a judgement I’m making, I’m not arguing Dan has that exact sentiment), which is to mock an attitude (which wasn’t there) of hubris and lack of understanding of what makes a good product.

Re: AI Police Reports: Year in Review

#106

Post by cortic »

vid wrote: post #100
cortic wrote: post #98

Earlier quoted context omitted.

> ChatGPT (o3): Scored 136 on the Mensa Norway test in April 2025 So yes, most people are right in that assumption, at least by the metric of how we generally measure intelligence.

Court reports should as much be about human sensibility. I have met plenty of high IQ people who were insensitive.

Having listened to some the new AI generated songs on utube, looks like they might be better at being sensitive humans than we are as well..

Re: AI Police Reports: Year in Review

#107

Post by closewith »

impossiblefork wrote: post #99
roenxi wrote: post #60

Earlier quoted context omitted.

As far as I can tell from poking people on HN about what "AGI" means, there might be a general belief that the median human is not intelligent. Given that the current batch of models apparently isn't AGI I'm struggling to see a clean test of what AGI might be that a human can pass.

LLMs may appear to do well on certain programming tasks on which they are trained intensively, but they are incredibly weak. If you try to use an LLM to generate, for example, a story, you will find that it will make unimaginable mistakes. If you ask an LLM to analyze a conversation from the internet it will misrepresent the positions of the participants, often restating things so that they mean something different o…

This seems distant from my experience. Modern LLMs are superb at summarisation, far better than most people.

Re: AI Police Reports: Year in Review

#108

Post by ptx »

hombre_fatal wrote: post #97
ptx wrote: post #78

Earlier quoted context omitted.

How is verification faster and easier? Normally you would check an article's citations to verify its claims, which still takes a lot of work, but an LLM can't cite its sources (it can fabricate a plausible list of fake citations, but this is not the same thing), so verification would have to involve searching from scratch anyway.

Because it gives you an answer and all you have to do is check its source. Often you don’t have to do that since you have jogged your memory. Versus finding the answer by clicking into the first few search results links and scanning text that might not have the answer.

As I said, how are you going to check the source when LLMs can't provide sources? The models, as far as I know, don't store links to sources along with each piece of knowledge. At best they can plagiarize a list of references from the same sources as the rest of the text, which will by coincidence be somewhat accurate.

Re: AI Police Reports: Year in Review

#109

Post by dataflow »

futuraperdita wrote: post #16

What worries me is that _a lot of people seem to see LLMs as smarter than themselves_ and anthropmorphize them into a sort of human-exact intelligence. The worst-case scenario of Utah's law is that when the disclaimer is added that the report is generated by AI, enough jurists begin to associate that with "likely more correct than not".

One problem here is "smarter" is an ambiguous word. I have no problem believing the average LLM has more knowledge than my brain; if that's what "smarter" means, them I'm happy to believe I'm stupid. But I sure doubt an LLM's ability to deduce or infer things, or to understand its own doubts and lack of knowledge or understanding, better than a human like me.

Re: AI Police Reports: Year in Review

#110

Post by ehnto »

cortic wrote: post #98
Marha01 wrote: post #45

Earlier quoted context omitted.

> a lot of people seem to see LLMs as smarter than themselves Well, in many cases they might be right..

> ChatGPT (o3): Scored 136 on the Mensa Norway test in April 2025 So yes, most people are right in that assumption, at least by the metric of how we generally measure intelligence.

Does an LLM scoring well on the Mensa test translate to it doing excellent and factual police reporting? It is probably not true of humans doing well on the Mensa, why would it be true of an LLM?

We should probably rigorously verify that, for a role that itself is about rigorous verification without reasonable doubt.

I can immediately, and reasonably, doubt the output of an LLM, pending verification.

Post reply on HN