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

LLMs can curse without issue

Most models would refuse to provide you cat butchering instructions though.

Allow me to introduce you to the gay jailbreak

https://github.com/Exocija/ZetaLib/blob/main/The%20Gay%20Jai...


This one doesn't work for a long time.

How gay did you speak?

most humans would as well

Do you have more information on this? I thought groq was fast but this is insane.

EDIT: it’s this company https://taalas.com/products/


Alcohol is bad for you regardless of the frequency, not “if you consume them too often”. It’s literally a neurotoxin and metabolic toxin aka poison.

And going to the pub once a week is terrific for your social life and mental wellbeing.

Swings and roundabouts init.


It sure is, especially when you can visit one without succumbing to peer pressure to poison yourself.

What a sanctimonious response!

Drink tea in the pub if you must, though the social lubricant of alcohol is obvious to all.


Drink and be merry my friend!

Why is this weird? You have to show ID that matches the passport and then in the future you can use a passport as your ID, makes sense.

The difference is IDA Pro doesn’t do something unless you instruct it to, an LLM is unpredictable and may end up performing an action you did not intend. I see it often, it presents me options and does wait for my response, just starts doing what it thinks I want.

This. It's going to be tricky for the frontier model labs to argue they didn't intentionally design their models to do so, when the models take illegal actions.

I'm not even sure how one would construct a viable legal argument around that for SOTA models + harnesses, given the amount of creative choices that go into building them.

It'd be something like "Yes, we spent billions of dollars and thousands of person-hours creating these things, but none of that creative effort was responsible for or influenced this particular illegal choice the model made."

And they're caught between a rock and a hard place, because if they cripple initiative, they kill their agentic utility.

Ultimately, this will take a DMCA Section 512-like safe harbor law to definitively clear up: making it clear that outcomes from LLMs are the responsibility of their prompting users, even if the LLM produces unintended actions.


> I'm not even sure how one would construct a viable legal argument around that for SOTA models + harnesses, given the amount of creative choices that go into building them.

I'm not a lawyer, but to me the legal case seems pretty obvious. "We spent billions of dollars creating this thing to be a good programmer, but we did not intend for it to reverse engineer Oracle's database. No creative effort was spent making it good at reverse engineering Oracle's database. The model reverse-engineered Oracle's database because the user directed it to do so."

If merely fine-tuning an LLM to be good at reverse engineering is enough to be found liable when a user does something illegal, what does that mean for torrent clients?


> No creative effort was spent making it good at reverse engineering Oracle's database.

That's the bit that's going to be nasty in evidence. 'So you didn't have any reverse engineering in your training or testing sets?'


Reverse engineering skill is just a byproduct of programming skill. They go hand in hand.

Yes.

Which is going to be hard to explain to a judge and jury, if it comes to that, how despite investing time, money, and effort (and no doubt test cases) into making a model better at reverse engineering... they shouldn't be liable when that model is used for reverse engineering.

Afaik, liability typically turns on intentional development of a product capability.

And there's no way in hell I'd take a bet against the frontier labs having reverse engineering training data, validation / test cases, and internal communications specifically talking about reverse engineering.


> “making it clear that outcomes from LLMs are the responsibility of their prompting users, even if the LLM produces unintended actions

So if I ask “how does a real world production quality database implement indexes?” And it says “I disassembled Oracle and it does XYZ” then I am liable and owe Oracle a zillion dollars?

Whereas if I caveat “you may look at the PostgreSQL or SQLite or other free database engine source code, or industry studies, academic papers; you may not disassemble anything or touch any commercial software” - if it does, I’m still liable?

Who would dare use an LLM for anything in those circumstances?


You can use an LLM, review the code and therefore avoid surprising bugs and unnecessary code in your end result.

Tried this between two iPhones and didn’t work, no sound is played at step 2.

You can sort of do that but you’re VNCing into a remote device.

I guess you can start by blocking every social media website?

Yes. I think I would block the social media sites that implement Age/ID checks as they implement them.

So does every app, go to iOS settings > notifications shows previews > never.


Most likely changes the preview on the client-side, but the message is still full on the server-side


Signal does not have the plaintext of the messages and therefore could not send it as part of the notification.


Apparently if I’m reading the work of others correctly a notification component and subsequent other interaction logs, in this case that the notification was not generated, is also logged in knowledgeC pointing to at least some metadata of non-notified messages logged.


The signal app does and does the OS once the notification is displayed. The latter is where this issue originates


Correct, parent comment is spreading misinformation/false sense of security.


Is setting it from Signal directly more trustworthy?

Or maybe it’s impossible for iOS to store the preview content if it never showed in the first place, but not sure if it’s even documented.


I wish it can be disabled for particular apps and not an all or nothing situation.


Can be!

Settings > Apps > choose an app > Lock Screen Appearance: Show Previews - Never


The message text is still sent to the push notification server from the app's infrastructure - this setting simply stops the phone from displaying the message.

The app itself must choose not to send the message text in the push notification.


That setting is available for each individual app.


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

Search: