Hacker News Reader

Comment by matwood | Hacker News Reader

> We can no longer assume that a book is legit simply because it looks legit.

The problem is that this has been an issue for a long time. My first interactions with the internet in the 90s came along with the warning "don't automatically trust what you read on the internet".

I was speaking to a librarian the other day who teaches incoming freshman how to use LLMs. What was shocking to me is that the librarian said a majority of the kids trust what the computer says by default. Not just LLMs, but generally what they read. That's such a huge shift from my generation. Maybe LLM education will shift people back toward skepticism - unlikely, but I can hope.

View on HN

One of the issues today is the volume of content produced, and that journalism and professional writing is dying. LLMs produce large amounts of "good enough" quality to make a profit.

In the 90s we could reasonably trust that that the major news sites and corporate websites was true, while random forums required a bit more critical reading. Today even formerly trusted sites may be using LLMs to generate content along with automatic translations.

I wouldn't necessarily put the blame on LLMs, this just make it easier. The trolls and spammers was always there, now they just have a more powerful tool. The commercial sites now have a tool they don't understand, which they apply liberally, because it reduces cost, or their staff use it, to get out of work, keep up with deadlines or to cover up incompetence. So, not the fault of the LLMs, but their use is worsening existing trends.

> Today even formerly trusted sites may be using LLMs to generate content along with automatic translations.

Yep - or they're commingling promotional content with their journalism, a la Forbes / CNN / CNET / About.com / etc. There's still quality content online but it's getting harder to find under the tidal wave of garbage.

> I was speaking to a librarian the other day who teaches incoming freshman how to use LLMs. What was shocking to me is that the librarian said a majority of the kids trust what the computer says by default. Not just LLMs, but generally what they read. That's such a huge shift from my generation.

I think that previous generations were not any different. For most people, trusting is the default mode and you need to learn to distrust a source. I know many people who still have not learned that about the internet in general. These are often older people. They believe insane things just because there exists a nicely looking website claiming that thing.

Not sure of the context here is for "previous generation" but I've been around since early in the transition from university/military network to public network, and the reality was the internet just wasn't that big, and it was primarily made up of people who looked, acted and valued the same things.

Now it's not even the website of undetermined providence that is believed; positions are established based on just the headline, shared 2nd or 3rd hand!

> The problem is that this has been an issue for a long time. My first interactions with the internet in the 90s came along with the warning "don't automatically trust what you read on the internet".

I received the same warnings, actually it was more like “don’t trust everything you read on the internet”, but it quickly became apparent that the last three words were redundant, and could have been rephrased more accurately as “don’t trust everything you read and hear and see”.

Our parents and teachers were living with their own fallacious assumptions and we just didn’t know it at the time, but most information is very pliable. If you can’t change what someone sees, then you can probably change how they see it.

I feel like there was also a brief window where "many amateur eyes in public" trumped "private experts"; wikipedia, open source software, etc. This doesn't seem the case in a hyper-partisan and bifurcated society where there is little trust.

> Our parents and teachers were living with their own fallacious assumptions and we just didn’t know it at the time, but most information is very pliable.

Indeed. When I was 14-18 in the UK, the opinion pieces in the news were derogatorily describing "media studies" as "Mickey Mouse studies".

In retrospect such courses were teaching critical analysis of media sources, in much the same way that my history GCSE went into the content of historical media and explored how both primary and secondary sources each had the potential to be either accurate or biased.

Even now, even here, I see people treat the media itself as pure and only the people being reported upon as capable of wrongdoing — e.g. insisting that climate scientists in the 70s generally expected an imminent ice age, because that's what the newspapers were saying.

Believe half of what you see and none of what you hear

True!

Reply on news.ycombinator.com