461–470 of 643 posts
Earlier quoted context omitted.
I'd be interested to know why those comments were flagged actually. They don't scream AI and no-one has replied calling them out as AI, etc. But the vast majority are dead.
> four bland messages That's why. Boring, bland, etc. That account's M.O. is basically "write a paragraph that says nothing." Fwiw, I do think AI can be indistinguishable from dumb, boring people, but usually those kinds of people won't be on HN.
I'm still salty that I can't use em-dashes anymore for fear of my writing being flagged as AI generated. Been using them for years—it's just `alt+shift+-` on a Mac keyboard and I find them more legible in many fonts compared to the simple dash on the typical numpad. It's so sad to me that good typographical conventions have been co-opted by the zeitgeist of LLMs.
Em dashes, semicolons, deftly delving. It’s all just so…facile. We might as well tell ourselves we can tell it’s shopped from the pixels, having seen some shops in our day.
Earlier quoted context omitted.
Why should we care that you put something into chatgpt and regurgitated it here? How does that make the conversation more interesting
I think my point was that AI actually ate the original comments which were jokeful and literally showed all the classic AI symptoms on that again while showing the classic issue itself. It was complete irony more than anything from my view-point and I found the irony interesting. The "interesting" thing about this is that you can give any rudiculous idea to AI and say autocomplete after this: "You are absolutely righ…
Edit: I realize that sounds harsh. Not trying to be. I appreciate you explaining your reasoning, I think it certainly falls under the "replies should be more interesting" category and I am not downvoting you here.
So much so that I've started a Wikipedia project to replace the dashes and other abysmal characters with proper typography.
All of my devices replace hyphens with emdashes, ascii typography with glorious unicode, etc.
Prior to the rise of LLM-written posts and the natural reaction of hair-trigger suspicion, I used to em and en dash fairly often in posts on HN. No reason really other than being a bit of a typography geek who happens to have always used dashes in casual writing instead of semicolons. So when I was setting up a modifier-key keyboard layer with AHK many years ago I put the em dash on modifier+dash just because I could…
Were you using them as a replacement for a comma--without spaces on both sides of the em-dash--like how I did just now? If no, you are safe from being mistaken for an LLM program. Honestly, while it is a legitimate punctuation rule, I've never seen a human on the internet to write like that. But LLMs do it constantly, whenever they generate long enough sentences.
This feels like an existential threat to HN, and to the general concept of anonymous online discourse. Trust in the platform is foundational, and without it the whole thing falls down. Requiring proof of identity is the only solution I can think of, despite how unappealing it is. And even then, you'll still have people handing their account over to an LLM. I really struggle to imagine a way around it. It could be tha…
I don't think that's true at all. One of the things HN does is not let you interact in certain ways until you've earned sufficient karma. This is a basic proof-of-work. If your bot can't average a positive karma, then it'll never get certain privileges. Not to say the system is perfectly tuned for bots, because it's not. The point is that proof of identity is not the only option.
Many of them sound and look completely normal and have others on here interacting with them. They don't use em dashes, sometimes they'll use all lowercase text, sometimes the owner of the bot will come out and start commenting to throw you off.
All examples I've witnessed here.
HN should immediately start implementing at least some basic bot detection methods without requiring us to email them every time. I've discovered multiple bots make detailed comments within 30 seconds of each other in different threads, something a normal human wouldn't be able to do. That should be at least flagging the account for review. Obviously they'll get smarter and not do that soon but it would help in the short term.
I'd say it's not an issue but everything I described above has happened in less than a month and every day now I'm discovering bots here.
Fwiw I did some more comparisons, looking for words disproportionately favored by noob comments: word noob new p-value ---------------------------- ai 14.93% 7.87% p=0.00016 actually 12.53% 5.34% p=1.1e-05 code 11.47% 6.04% p=0.00081 real 10.93% 2.95% p=2.6e-08 built 10.93% 2.11% p=2.1e-10 data 8.93% 3.51% p=6.1e-05 tools 7.6% 2.67% p=5.5e-05 agent 7.47% 2.95% p=0.00024 app 7.2% 3.09% p=0.00078 tool 6.8% 1.83% p=8.5e…
You've built an interesting statistic from gathering data across the project. The real answer: ai models and agentic apps make building spam tools more simple than ever. All you actually need is just some trivial api automation code.
Do all the models have this style of talking? Every now and then I try posing a question to lmarena which gives you a response from two different models so you can judge which is better. I feel like transitions like "The real answer...", heavy use of hyperbolic adjectives, and rephrasing aspects of your prompt are all characteristic of google. Most other models are much more to the point
Fwiw I did some more comparisons, looking for words disproportionately favored by noob comments: word noob new p-value ---------------------------- ai 14.93% 7.87% p=0.00016 actually 12.53% 5.34% p=1.1e-05 code 11.47% 6.04% p=0.00081 real 10.93% 2.95% p=2.6e-08 built 10.93% 2.11% p=2.1e-10 data 8.93% 3.51% p=6.1e-05 tools 7.6% 2.67% p=5.5e-05 agent 7.47% 2.95% p=0.00024 app 7.2% 3.09% p=0.00078 tool 6.8% 1.83% p=8.5e…
Can you articulate on the column meanings more? Noob new means nothing to me.