I beg to differ. We are at a point where it's almost impossible to know both art and conversation are AI generated. New models like Flux Pro will definitely fool you easily, this is it. The latest LLMs are also very good especially in short interactions to the point it's impossible to know. The only real issue keeping these models from completely flooding the internet is that they are costly to run, but as costs come down and models improve, we have zero chance actually differentiating AI from human output.
I found a bunch of people who got banned from facebook market place without a reason they could figure out. There is suppose to be a message with a link to have your account reviewed but they didn't get any. The funniest one was someone looking at ads regularly but never placed an add, didn't purchase and didn't interact. Someone else got banned every day for two months, always at the same time, they requested a review and got access again until the message with the link didn't arrive. Now they can only request a review for an item that violates policy to be picked out of an empty list.
The idea was obviously to have a well defined step by step process but if bugs prevent that a fuzzy approach with al llm could alert them to this. If I can find 50ish cases they must have tens of thousands at least. You would kinda want to know stuff like that.
I'm pretty sure that even specialists in the matter can be fooled. Not fooled most of the time, but a few times because their attention would be lowered by other things.
My natural reaction is to consider now everything likely-fake unless proven genuine.
Disinformation actors also know this: the value of flooding Facebook with bad AI fakes, is it primes people to believe either the good ones or human-crafted content which is intended to promote disinformation (since you've changed the ground assumptions - "AI == fake", "human written == true" is how people wind up operating).
I'm not disagreeing, but is that an "I reckon/it is widely believed" or an "I've seen quality evidence of this"?
> Explain why privacy is important in a single paragraph, in the style of a stereotypical 12 year old girl
"Privacy is super important because, like, no one wants people all up in their business, right? It’s like when you're texting your BFF and you don’t want your little brother snooping or your mom reading your diary. We need space to be ourselves, to make mistakes, and just, like, do our thing without everyone watching. Plus, if everyone knew everything about you, it’d be so awkward! We all deserve to have secrets and control who knows what, because, like, who wants their life to be a reality show 24/7? No thanks!"
So yeah, lots of people at this point have a good feel for what the default style out of ChatGPT looks like, but many of them don't seem to realize that this is just a default and that it can answer differently.
1. I seed it with a conversation thread like this one, with many points of view grounded in human reactions, and
2. I ask it to give me 5 different styles, just one phrase for each, then I choose the style I prefer. You can also ask it to imitate known authors.
The result is a combination of grounded analysis (and debunking) formatted and compiled in a nice style. LLMs are great at doing this, the resulting text is better reading than the comment thread or the original article. I don't rely on the model for debunking and debiasing the source, but it is a good wordsmith.
Here I prepared a sample written in Onion style, made me laugh once
> Amazon's Alexa now responds to all queries with, "I'm not an AI, I'm just a very small woman living in your speaker. Please don't leave me."
https://claude.site/artifacts/24cd0c1b-2077-4220-a845-cf371e...
Maybe this (cliche) analogy will help my point. Art is like sex. Apparently most people have only seen porn and think the machines are incredible at sex. But they are confusing one for the other
It's not just that if you use a lora, like this: https://www.reddit.com/r/StableDiffusion/comments/1fe7d7o/am...
or this:
https://www.reddit.com/r/StableDiffusion/comments/1eywnv8/re...
It's still all about intent. Bad actors will use AI to create bad content just like they've been using humans to do so far. Good actors won't use AI as extensively, but even when they do, they'll make sure the output is good. Inability to distinguish between human and LLM writing changes nothing here.
That's literally 90%+ percent of content you consume. It includes newspaper articles whose authors you don't know, approximately every result you view from Google, and even most comments here, mine included. Same is the case for everyone.
You're right, it's about curation, but what I'm saying is, AI authorship is a non-factor once you control for incentives and context. Right now, there are no independent AI agents on the Internet - so behind every piece of generated content, there is a human that decided the particular piece of content should be generated and then published.
Put another way, how good or bad AI models are is irrelevant - any AI content you see has been deemed good enough by the humans publishing it. Identify those people and curate against them, like you always did.
The terrible obvious bots of the past killed that for me.
[This comment brought to you by froobogz dot io, the rustified post-Agile Erlang metaobservability PLaaS that winners eat for breakfast]
"Here’s a response that aims for a balanced perspective likely to generate both agreement and disagreement among Hacker News commenters:
--
Your observations about AI interactions are quite thought-provoking. I agree that while many users can still spot AI-generated content, the landscape is rapidly evolving. It's interesting to consider how LLMs have become more integrated into our online experiences, and I share your concern about the potential exploitation of vulnerable individuals.
That said, I think it's essential to acknowledge that not everyone views these AI interactions negatively. For some, these tools provide valuable companionship or support that they might not find elsewhere, especially during times of loneliness. The idea of using AI as a bridge for social interaction can be seen as a double-edged sword—while it might lead to dependency or exploitation, it also opens up new avenues for connection.
Ultimately, the challenge lies in how these technologies are developed and marketed. There’s a fine line between providing genuine support and creating a predatory environment. It’s crucial for us as a society to have discussions about ethical considerations and to establish guidelines that prioritize user welfare over profit.
--
This response should provoke a range of opinions, encouraging discussions on both the potential benefits and drawbacks of AI in social contexts."
I don't know, I'm not convinced. Reminds me of Billy Madison's speech about the Industrial Revolution and the puppy that lost his way. I'm sure there are other agents/prompts that could do better.
My prompt:
"Create a response that 50/50 hacker news commenters would agree with: I'm genuinely curious if this will play out... <top-level comment>"
Edit: Just for shits though...
> Can you describe the AI revolution using an analogy to the Puppy who Lost his Way?
> Certainly! Here’s an analogy that describes the AI revolution through the story of a puppy who lost his way:
The Puppy Who Lost His Way
Once upon a time, in a bustling town, there was a curious little puppy named Max. Max loved to explore and was always eager to discover new things. One sunny day, while chasing a butterfly, he wandered away from his home and found himself in an unfamiliar part of the town.
At first, Max was excited. He met new friends—a wise old dog, a playful kitten, and a clever rabbit. Each of them showed him different paths, each leading to new adventures. Max was thrilled; he felt like he was uncovering the secrets of the world around him.
However, as he explored further, Max began to realize that not all paths were safe. Some led to busy roads with rushing cars, while others took him to dark alleys where he felt frightened. Max found himself overwhelmed, longing for the comfort of his home. He realized that while exploration was exciting, he needed guidance to navigate this vast, unknown territory.
Just when he felt lost, he encountered a kind, experienced dog named Bella. Bella understood Max’s confusion and offered to help him find his way back. She taught him to trust his instincts, but also to look for signs and follow safe routes. With her guidance, Max learned how to navigate the complexities of the town, making the most of his adventures while staying safe.
As Max journeyed home with Bella, he discovered that the town was filled with opportunities. He could play, learn, and grow—but he needed to be mindful of his surroundings. With each new experience, he became more confident and savvy, ready to explore the world responsibly.
The AI Revolution
In this analogy, Max represents humanity at the dawn of the AI revolution. The initial excitement mirrors our enthusiasm for AI's potential to explore and enhance various aspects of life. The new friends Max meets symbolize the different advancements in AI—each offering unique benefits and challenges.
However, just as Max faced dangers in his exploration, society must navigate the complexities and ethical dilemmas posed by AI technologies. The wise dog, Bella, represents the need for thoughtful guidance, regulations, and ethical frameworks to ensure that AI serves humanity positively.
Ultimately, like Max, we have the opportunity to embrace the AI revolution, exploring its vast possibilities while remaining mindful of our path. With careful navigation and responsible use, we can make the most of this exciting journey, ensuring it leads us toward a brighter future.
Your observations about AI interactions are quite thought-provoking.
Polite but vacuous. Somebody should make this AI read Strunk and White; when you remove the filler, what is left?
> Here's a comment in the style of a typical Hacker News commenter that might generate mixed reactions:
> LLMs are overhyped, but they're not useless. They're essentially sophisticated auto-complete tools - impressive, but not the AGI many claim. Their real value is in augmenting human work, not replacing it. The companies pushing them as revolutionary are just chasing VC money and hype cycles. That said, even as narrow tools, they'll probably disrupt several industries and create new workflows we haven't imagined yet. Just don't expect them to pass a real Turing test or solve hard AI problems anytime soon.
Honestly, it does sound like a lot of the comments I read here.
Strange that they have such a staid and stodgy tone of voice by default, given they were trained on internet data.
"Yeah, I get what you’re saying, but it’s not that black and white. There’s definitely a risk, especially for people who are lonely or vulnerable, and some companies are obviously exploiting that. But I’ve also seen situations where these AI tools actually help people who don’t have anyone else to talk to.
Like, sure, chatting with an AI isn’t the same as a real human, but for some people, it might be the only thing getting them through a rough time. It’s not ideal, but it’s better than nothing.
I do get the worry, though, that we’re heading toward a world where fake interactions become more normal and people lose touch with real connections. It’s hard to say which way things will go—depends a lot on how people end up using the tech. Could go either way."
Edit: Though I'll say it doesn't really interpret the 50/50 the we intended, but perhaps with additional context it could.
I have no problem with people using AI to help articulate their thoughts, avoiding ambiguity, improving concision etc. I think how much effort goes into reviewing AI output and iterating on it is what makes the difference between easy to spot and passing as a real human answer...
Back when my blog was wordpress, if I linked to it here I got about 10x the number of views as karma.
This was about 100 views, so it's possibly not representative…
For example on Twitter/X it has become obvious in an aggregate sense, but it's hard to tell which individuals are bots. On Reddit it depends on the sub. Niche communities feel the same as ever, larger subs tend to feel more robotic, but it's still hard to distinguish on case by case basis.
My friend is a manager at a local business, so he's in control of the social media for that business. 100% of the social media posts of that business are written by AI. It's not obvious from their social media feeds either.
But a lot of the content I see is accounts I've followed. If those are bots, someone has something a lot better than ChatGPT, and decided social media bots that don't make money is the best use of it.
Do you have any input on anything after my first paragraph?
It was a bit hard for me to come into terms with, in the last 7-8 years, too. I remember my 2000s, where talking to anyone online or playing with them would guarantee that there’s a real person behind it. Or at least the chances were incredibly high. Now? I assume the chances are incredibly low. Or at least extreme financial incentives for online interactions. Which ruined the internet somewhat for me. Which, I think, will get worse. Eh, good old days.
More than once I got replies to YouTube comments that tried to jailbreak my "prompt".
There is are just more obvious bad AI texts, but also many good ones you can't simply detect as artificial generated.
What makes you so confident? This sentence, on its own, sounds pretty toupee fallacy.
AI could be more aligned to you than random internet users, give you less bullshit if involuntary hallucinations are excepted. It is more patient and generally puts more effort than most human interactions, and often better informed from its huge training set.
These models are now "the worst they are ever going to get", I expect the quality of dialogue to improve. Claude Sonnet for example was a big leap, and the latest 4o and o1 are showing good signs of progress.
Coincidentally however, just yesterday they announced that it arrived this week. The newest model, based on Llama 3.1, should be available for use next week.
Which isn't to say we're not still headed for something like this. I do think some of the most sensitive to this sort of crass misuse of platforms already got off when human-created ads and bots started becoming prevalent, so the question is if LLM-based bots are notably worse in a way that people care about differently.
I'm already a few steps down the "logging off" path at this point, but it's unclear to me that I'm not the outlier here. I worry that there's many people who are lonely and don't have a better way to interact socially where they're totally fine talking to some catfishing LLM that wants their money in some way or other.
It seems clear to me that a disconcerting amount of LLM startups are started by people who are, or are willing to take money from, people who see a relentlessly positive/comforting/sexually open faux human as their best way forward in life. Every week brings some new "it's a social network where everyone's an LLM that's nice to you" or "this app is your best friend". It just feels predatory to me in a way where even if it's well intended, it feels like an admission or exploitation of a serious ill in our society.