Hacker Newsnew | past | comments | ask | show | jobs | submit | nautilus12's commentslogin

I was thinking to myself this dismal thought: everyone is so busy because of the acceleration caused by AI that no one will have the time to actually take the five minutes to verify manually that this really is a solution or that the article just generated something plausible looking

Is this assumed to be 2D? I was going to ask if there are any observed examples of 3 body equilibrium observed in nature.

Yes. Because 3 points are coplanar, so every "3D problem" with 3 objects can be turned into a 2D problem on the correspondent plane

(of course in real life your plane would keep changing, and probably some other complicated math I can't think right now)


Put another way, while their positions are one set of 3 points, their momenta are another set of 3 points, and there is no requirement that 6 points will always lay on the same plane.

I wonder what phantom forces would appear when the reference frame changes in some complicated fashion. We get centrifugal "force" when we reconstruct F=dP/dt in a rotating reference frame, what would the 3-body "force" look like?


Oh makes perfect sense, do you have thoughts about real life examples? I did some research using AI and it said there were examples of restricted 3 body problems like the trojan asteroids, but no examples in real life similar to what is in this web app

Why would the plane keep changing? If there are only these three objects, won't the vectors of their gravitational pull to each other all be on this plane too?

If you define the initial conditions such that their relative velocity is zero or parallel to the plane, yes. But that's not the case in general.

Not from an external point of view, as you might have a momentum component perpendicular to that plane

(but yes I think you might be right if we're centered on the CG)


One of the necessary conditions here is that the three objects return to their exact initial position, and so does the centre of mass. Initial conditions with non-zero momentum must trivially be ruled out. But this doesn't stop them from having velocities perpendicular to the initial plane that cancel out perfectly, so this doesn't refute the assertion that the planes keep changing.

Sorry, but this is a word salad.

Most of the solutions always have non-zero momentum, including in the initial conditions. And https://numericaltank.sjtu.edu.cn/three-body/three-body.htm includes periodic solutions that move in all three dimensions.


They presumably meant non-zero total momentum. If the total momentum were non-zero, then the center of mass will be moving in a straight line, and will not return to where it began, and therefore the orbit would not be periodic.

Big picture the center of mass momentum is concerned and it does not matter if the system as a whole is moving up or down or to the right or the left. Like the Earth is basically orbiting the sun in an ellipse [1] so far as the sun is concerned and from the viewpoint of the solar system not care so much that it is moving around the galaxy unless we are interested that orbit being perturbed by other stars that we pass near over millions and milions of years.

[1] ignoring the parameters of that ellipse changing slightly and slowly thanks to the other planets


Oh. That makes sense. Yes. The total momentum has to be zero.

All of the things in the solution can constantly have momentum.


Oh sorry, yeah - I wasn't fully awake yet when I wrote it. As a sibling comment mentioned, I was talking about the total momentum. Anyway, it was in response to GP's assertion that GGP may be correct if "we're centered on the CG", which doesn't make sense because the CG can't move.

Each orbit is a spinning top. You pull on a top from the side, it's spin axis precesses.

It’s an arbitrary plane, chosen at each moment just so you can flatten it

https://numericaltank.sjtu.edu.cn/three-body/three-body.htm shows that there are three dimensional solutions.

We have no observed examples in nature of three body equilibrium. But then again, all places we have looked are either influenced by the chaotic orbits around them of the Solar System, our surrounding galaxy, or nearby galaxies in a cluster.

There aren't a lot of orbiting three bodies without external gravitational influences disturbing them.


> There aren't a lot of orbiting three bodies without external gravitational influences disturbing them.

This is of course a relative statement. Every object affects every other object, subject to the limitations of lightspeed propagation of gravity waves through expanding space.

But as you point out, we still haven't noticed any examples that are stable short-term.


That depends on what you mean by stable short-term.

The Sun, Jupiter, and any small asteroid at L4 of L5 is, by itself, linearly stable. Meaning that any small perturbation will only grow linearly.

But, of course, this system interacts with Saturn. The arrangement of those three objects is still remarkably stable. But the interaction with Saturn makes for a chaotic system again.


This is the real issue, though this article does not really hit on it explicitly just alludes to it.

Yes I think they should be more direct. Not sure why they weren't

Thats not what they are asking. This paper is discussing documents in the training dataset poisoning the LLM for malicious behavior. This person are asking if anyone has deliberately put something in a private chat (presumably with retrain on my data turned off), to see if they can get it to leak across sessions from distinct users. I am positive this happens but I have not seen the proof. I also want to know the answer to this question.

Here are potentially relevant documents?

https://medium.com/secludy/fine-tuning-llm-on-sensitive-data...

https://spylab.ai/blog/non-adversarial-reproduction/

https://arxiv.org/abs/2601.18834


* with train on my data turned ON, yes. Though OFF would of course be even more notable!

Thank you – the non-adversarial reproduction paper ( https://arxiv.org/abs/2411.10242 ) nails it – from chat, to training corpus, to subsequent model. Though in my hasty read, it is not entirely clear whether the snippets it finds are nonces, i.e. present exactly once in the internet.


Did you miss my last paragraph?

I presented the research that I knew was somewhat relevant. Then made it clear that that wasn't what was being asked, and why my expectation is what it is.


Ah i see what you were saying now, you were saying session specific poisoning is more likely to work if it was in tandem with training document specific poisoning.

This should be a watershed moment for everyone using a commercial LLM. EVERY PIECE OF INTELLECTUAL PROPERTY YOU DISCUSS WITH IT BECOMES THEIR PROPERTY. FULL STOP.

It is impossible to fight against this. It is statistically obfuscated, they can just argue it is a derived work.


I get jealous of people that have the luxury of time to do things like this


I just used AI(antigravity with gemini 3.6 flash, but i'm sure many others would work) to implement this and integrate it with my security cameras after reading the blog post 26 minutes ago. I suspect the author did the same thing. I also had it steal the idea from the AvianVisitors link in this thread and it displays a generated collage of the recently detected birds.


My setup took maybe ~4-8h of initial work. All with the help of Codex: code, configuration, even figuring out which equipment to buy.


My issue is that taking a strictly GUI driven approach makes you build an application (in the browser presumably) that will require more permissions than you would typically want in the terminal so TUIs win out by being in proximity to those tools (cat, ls, rm, etc) for development use cases


I think the hope of the business end is that coding expertise will at some point no longer be needed at all.


Can they try to prove that he directed someone to direct the agent to do that?


I call it "yegge sweats"


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

Search: