Hacker Newsnew | past | comments | ask | show | jobs | submit | gmzamz's commentslogin

Boss is using ai. 11 is clearly bigger than 5

Looks like AI to me too. Em dashes (albeit nonstandard) and the ‘it’s not just x, it’s y’ ending phrases were everywhere. Harder to put into words but there’s a sense of grandiosity in the article too.

Not saying the article is bad, it seems pretty good. Just that there are indications


It's also strange to suggest readers use ChatGPT or Claude to analyze email headers.

Might as well say "You can tell by the way it is".


I don’t understand this comment. I’ve found AI a great tool for identifying red flags in scam emails and wanted to share that.


The content ChatGPT returns is non-deterministic (you will get different responses on the same day for the same email), and these models change over time. Even if you're an expert in your field and you can assess that the chatbot returned correct information for one entry, that's not guaranteed to be repeated.

You're staking personal reputation in the output of something you can expect to be wrong. When someone gets a suspicious email, they follow your advice, and ChatGPT incorrectly assures them that it's fine, then the scammed person would be correct in thinking you're a person with bad advice.

And if you don't believe my arguments, maybe just ask ChatGPT to generate a persuasive argument against using ChatGPT to identify scam emails.


It's a good point and I should make a distinction on what models are appropriate. I think of chatGPT 4 like a college student and chatGPT 5.1 5 Pro (deep thinking model) more like a seasoned professional. I wouldn't trust non-frontier, non-thinking models with a result for this kind of question. But the determinism of the result does not scare me, the out output may vary but not directionally. The same thing would happen if you asked the foremost security expert in the world, you'd get slightly different answers on different days. One time as a I test I ran a very complex legal analysis through chat GPT pro 10 times to see how the results would vary and it was pretty consistent with ~10% variation in numbers it suggested.


1. They are all scam emails.

2. AI detecting a scam, sure - it’s a scam. AI saying the email is ok… then what? I’d never trust it.


I agree with this, my experience is that a small light weight LLM is a fantastic spam filter.


- em dash

- needless lists

- actual bullet point characters

- first sentence a variation of ‘you’re right’

There’s only 30ish comments on your post and half of them are yours. Are you really so busy that you can’t write your own 4 sentence reply?


Noted. Actually mods wrote me about this, apologies, I won’t use AI in comments here. Dlog is a private journaling app with optional goals and project lists. it’s not a team PM tool. The coach is optional and journals stay on the Mac by default. Concrete product feedback is welcome. I’m tightening the site copy and will keep replies brief, but in some cases provide links for longer. To keep both people who prefer brevity and depth happy. Take care. Johan.


You can't optimize for two things at the same time in the /same system/.

In your example, the best volume for the stereo will change based on how fast you're going. Noise from air going around your car will require a slightly louder volume. It technically goes the other way too, I'm sure louder music will affect your mileage slightly, reducing your MPG due to increased power draw. Realistically it won't matter, but they do exist in the same system.

You can absolutely optimize for two things at the same time in different systems. The ideal speed optimizing for MPG in different cars aren't really linked and can be found independently.


This is an instant buy for me. I've been looking for a 'smart timer' that I can interact with programmatically and via Home Assistant for the longest time. I'd resigned myself to either building a terrible version myself or hiring someone to make a bespoke one. I'm super glad you're making this!


Reader mode saved the day again for me. But it shouldn’t be needed at all.


I wish I liked reader mode, I just can’t get into it. I like how each website looks different. It adds variety to the web. Otherwise everything is just too homogenous.


Yes I agree. The differences from site to site also aid in reading. When things all look the same it is harder to read. That's why justified text is harder to read.


Utilizing fancy gTLDs can still prevent you from receiving mail. Not due to a decision to explicitly block it but because it doesn’t match whatever regular expression they use to validate. Notably, .email fails consistently due to it being >3 characters. I tried to convert to using first@last.email and there is a significant minority of sites that didn’t allow it.


Good point! That's still kind of orthogonal to deliverability though. In fact in line with my point, you're better off finding this out before you start transferring personal correspondence to that domain.

The only similar problem I've experienced is sometimes companies will get uppity if you put their company name in the email address you give them. But it's easy enough to just make up a difference nonce for those cases (or start your scheme based on opaque nonces for everyone). I'm still waiting for the other shoe to drop and surveillance companies to start discriminating against non-surveillance-company email addresses the way they do against VOIP phone numbers.


I've been using gTLDs for an email for several years now (about the time gTLDs came out). It was really rough going for a while but these last couple of years it's been rare for me to have an issue.


They offer a docker you can run locally as well.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: