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

I'm becoming AI-blind

cymerys.com

Post reply on HN
Print view View original

391–400 of 533 posts

Re: I'm becoming AI-blind

#391

Post by ACCount37 »

lelanthran wrote: post #353
TalkingCodeMonk wrote: post #59

Earlier quoted context omitted.

The "something deeply wrong" part about AI, that even most technology enthusiasts evidently do not seem to grasp, is that it is still fundamentally a statistical model — an algorithmic construct — and does not possess any real intelligence or critical thought whatsoever. No matter how much investors and tech companies want you to believe that they are on the verge of super intelligence, nothing I've seen to date can…

> The "something deeply wrong" part about AI, that even most technology enthusiasts evidently do not seem to grasp, is that it is still fundamentally a statistical model — an algorithmic construct — and does not possess any real intelligence or critical thought whatsoever. The problem might be that it cannot backtrack. When AI generates output, there is no backspace key for it - it uses "No, but wait!" all over inste…

Reasoning is the first solution to that. Reason, write, reason, rewrite. Backtracking without backtracking.

Re: I'm becoming AI-blind

#392

Post by ACCount37 »

wongarsu wrote: post #98
pyridines wrote: post #72

Earlier quoted context omitted.

I just can't accept that it possesses no intelligence. It is not equivalent to human intelligence, obviously, but how can a system without some semblance of rational thinking solve open math problems? Even composing earlier human work into something novel requires intelligence and understanding on some level.

We couldn't agree on what intelligence means before ChatGPT happened. Now, agreement on the term seems even further away If performing well on an IQ test or performing at a high level on knowledge work is intelligence to you, these models are intelligent. If intelligence requires sentience for you, then ... well, I don't think we really agree what that is either, never mind how to measure it. But LLMs certainly don't…

> But LLMs certainly don't have it right now

Why are you so sure of that? If you say yourself that we can't agree on what it is, and have no trusted measurement tools for it.

LLM sentience is firmly in the realm of "maybe".

Re: I'm becoming AI-blind

#393

Post by scotty79 »

causal wrote: post #30

There's some psychological mechanism by which my brain immediately recognizes AI generated text and just short-circuits to "there is no information here". And when I force myself to read AI-generated text I realize I'm making my brain do creative work to impart meaning to the words. It is exhausting because my brain is literally trying to do a just-in-time rewrite of the text into something valuable. Something is dee…

Didn't you have this mechanism before the AI? The internet is a cesspit filled with empty words and commercial speech for some time now.

Re: I'm becoming AI-blind

#394

Post by scotty79 »

TalkingCodeMonk wrote: post #59
causal wrote: post #30

There's some psychological mechanism by which my brain immediately recognizes AI generated text and just short-circuits to "there is no information here". And when I force myself to read AI-generated text I realize I'm making my brain do creative work to impart meaning to the words. It is exhausting because my brain is literally trying to do a just-in-time rewrite of the text into something valuable. Something is dee…

The "something deeply wrong" part about AI, that even most technology enthusiasts evidently do not seem to grasp, is that it is still fundamentally a statistical model — an algorithmic construct — and does not possess any real intelligence or critical thought whatsoever. No matter how much investors and tech companies want you to believe that they are on the verge of super intelligence, nothing I've seen to date can…

> it is still fundamentally a statistical model — an algorithmic construct — and does not possess any real intelligence or critical thought whatsoever

How could I seriously repeat that prayer, when it builds things I wouldn't be able to build and solves problems that I wouldn't be able to solve? I would have to assume that nothing I did in 25 years for money required any intelligence or critical thought whatsoever and I have higher IQ than 99% of the population. You might be comfortable with that but I'm more comfortable with ascribing at least some intelligence and critical thought to AI.

Re: I'm becoming AI-blind

#396

Post by mixermachine »

I need to support integrators for a mobile app SDK. Bank stuff.

We have one integrator which consistently uses a truly bad AI. My brain skips after two sentences already. There are always +9 questions what should just be 2 max. Redundant info is requested (e.g. please provide a change history of this function and if it will be deprecated) and everything is just unbelievably verbose.

90% of questions we get from them are already answered by the web documentation (which even has a search function) or are just non sense.

I'm truly considering to build an MCP just for them...

Re: I'm becoming AI-blind

#397

Post by pllbnk »

causal wrote: post #30

There's some psychological mechanism by which my brain immediately recognizes AI generated text and just short-circuits to "there is no information here". And when I force myself to read AI-generated text I realize I'm making my brain do creative work to impart meaning to the words. It is exhausting because my brain is literally trying to do a just-in-time rewrite of the text into something valuable. Something is dee…

Back and forths with the LLM are very useful when the person inquiring the LLM is invested in the conversation and wants to uncover the ground truth. Ultimately, some good result can come out of it because the person has a goal and LLM helps them reach that goal by uncovering the layers of knowledge which the person might not have.

When the AI-generated content is presented to a person without any prior investment, it just looks incoherent. An especially great example are these Claude-generated explainer-type pages, which look really nice, even interactive, the information from the first sight looks really well presented. But somehow it all just doesn't make sense to a human. And I think it's because humans are processing information linearly and building an internal story about the information. One could argue that LLM's also consume information linearly but the way this information is processed is a kind of all-at-once approach.

Just some speculation on my part but I have been trying to cope with this way information is presented because I am currently working at a place which is heavily documented by AI. And the only way for me to properly understand the documentation is by inquiring AI to help me.

Re: I'm becoming AI-blind

#398

Post by codethief »

idiotsecant wrote: post #310
codethief wrote: post #264

Earlier quoted context omitted.

Do they? https://longbets.org/1/ has yet to be settled. Either way, I doubt an LLM could fool anyone here who who knows how LLMs work into thinking it is human, at least not for an extended period of time (think about context length/compression, prompt injections, …).

You'll notice those goalposts are substantially stretched from the original test.

How so?

Re: I'm becoming AI-blind

#399

Post by scotty79 »

I think AIs are writing less and less for humans. The fact that reasoning trace is using English and we require them to be performant means the models need to squeeze in more of information into it. More information that is perfectly fine for AI to ingest, but it makes the text less and less comprehensible to humans. It's almost like steganography where model encodes crucial information in phrases meaningless or empty to a human. Human directed messages for LLMs become a rounding error they don't pay that much attention to. Efficacy of reasoning is what dictates their style.

We can no longer assume that the models will create a passable English in their user targeted messages. So in the short term, we need to have a tight guidelines in the harness, but in the long term, we'll probably need separate languages for thinking traces and for human communication. Possibly even not going through tokens at all when doing thinking part.

Alternatively, we could evaluate thinking traces on human readibility, but that's probably infeasible due to sheer volume.

Re: I'm becoming AI-blind

#400

Post by user43928 »

tlkn_bot_praxis wrote: post #258

I've been stuck in loop all week where a coworker asked me to review some methodology documents that are clearly generated by Claude (which is fine in this case) but every time I sit down my brain can't get through 2 sentences without feeling like I'm reading nothing. It actually feels like the content, which isn't empty, has no meaning. So I go on to another assignment and come back later and the same thing happens.…

I don't want to read my own Claude outputs, much less someone else's.

If I had to, I'd process it into a short summary and/or ask an agent questions I have about the methodology.

I would then give my feedback. If they ask for details, I can have my agent update the document directly too.

I would not treat an AI generated artifact, be it documents or code, as something that a human should process fully manually.

Post reply on HN