aarong11

Hi HN,

I’m working on an open-source simulation platform called InfraSim. It lets users create companies, build out simulated infrastructure (servers, APIs, networking, blockchains), and then red-team them through fault injection and scenario testing.

The goal is to build a sandbox for learning, experimentation, and security practice, where LLMs and humans collaborate to simulate realistic systems.

InfraSim is still early, but it includes: • An in-browser UI with a component system • Blockchain-backed company creation using Solidity • A “tick-based” simulation clock (inspired by EVE Online) • Human + LLM red-teaming workflows • Persistent state, internal APIs, and networking

It’s meant for ethical hacking, system prototyping, training exercises, and worldbuilding.

The code is here: github.com/infrasimorg/main

I’d love feedback, ideas, contributors, and testers! AMA.

--

Thank you :)

I guess a good example is things like the physical characteristics of a line drawn by the stroke of a paint brush. You can often see they all align generally in the same direction, and have a sort of "fingerprint"

I mean depending on which side of the brain the tumour is, it could be relevant. Don't quote me on this (I'm not a doctor) but the right hemisphere controls the left side of the body (including left hand) and the left controls the right.

I mean it happens every day. When people have a lot to lose and not much left to gain, they tend to roll over a lot more easily.

As far as what would motivate him: look at what's happening to Musk right now as an example. I'm not saying he doesn't comply and cooperate with law enforcement, but he's certainly being made a big example of right now from all sides.

Continuing government contracts, expectation of favours in return, they almost certainly hold some dirt on him (and many others).

And payment processors can also identify these. And most of them you have to purchase with a minimum amount on ($10 or $20 afaik).

Virtual debit cards however are interesting. My bank lets me set up as many of those as I like and I don't even have to use my real name or billing address with them.

And yeah, there's also the stolen credit card / debit card market. I really can't see this adding that much pain for these bot handlers. It might make it a bit easier to identify the patterns at least.

I think this will just end up moving the problem further down the line and end up with twitter accounts being bought / sold.

There's no such thing as "one correct answer" to most tasks in any language. In mathematics, maybe. If you're asking it to recall a specific quote verbatim, yeah. But other than that there are many (infinite?) different ways you can paraphrase differently to get the same point across (at least in English).

Just because its 2 things you'd otherwise accept doesn't mean it isn't a dark pattern. Lots of dark patterns (like the setting up of privacy / data sharing settings on google) do similar things.

Leave the microwave running

With 96gb of VRAM, will you really need to? Personally I think apple servers might be adopted for AI and LLM workloads soon.

Except edge then opens and presents you with some message about setting it to the default browser (which a lot of people will just click ok to make it go away)

The last time this happened with titles. Everyone seemed to think it was hallucinating but it actually turned out to be really leaking other peoples titles.

Ultimately its pretty difficult to distinguish between the two, so these sort of reports do need to be taken seriously. We have no idea what sort of caching mechanisms they have with ChatGPT, nor do we know how they handle memory.