I would caution against thinking it's difficult for an LLM. I've used them in raw data file analysis and they are frequently shockingly good at pulling structures and meaning out of seemingly random data. Disassembled binaries already are structured, so pulling code flow out of that is easier. Mixing that with existing disassembly and inspection tooling and an LLM has what is needed to fast track this kind of vulnerability research. Point being, an LLM with the proper tools can potentially follow code flow from disassembled binaries way easier than a human.
I forgot who it was, but someone on YouTube said LLMs already work hooked up to gidra. If true it's only a matter of time once they find similar things in e.g. Windows. I'll wait half a year to a year (think of embargo) and if there still isn't such work for Windows I'll conclude that LLMs have a problem disassembling binaries.
It got the Bolsheviks to take over, and one of their main goals on international stage was to estabilish Communism in Germany. They didn't get to it only because they were defeated by Poland in 1920, and you can't get to Germany without conquering Poland first. So, Germans funded a party that nearly started a war with their own country. That's a hell of a gamble.
Fair, but still this isn’t a local chapter of a national Antifa organization. So it raises the question: is every local group that calls themselves “Antifa” now subject to investigation for domestic terrorism?
I would say you are being unnecessarily pedantic: the GP said "anything wrong", and the original comment obviously believes that there is "something wrong" which makes the choice not "sufficiently tasteful/professional/anodyne" (/me looking up "anodyne").
The obvious positive reading of the GP comment is that they disagree anime characters make it not "sufficiently tasteful/professional/anodyne".
I am replying to the "parent" comment which replies to the "grandparent" comment on the "original comment".
You seem to be unnecessarily pedantic too, while being wrong at the same time (I would get those relationships wrong sometimes if I thought it was clear enough).
Debates, especially on highly subjective issues, will not always be resolved quickly and definitively in the absence of bad faith. That "ideological reasons" you're sensing is the worldview of the person you're talking to.
I mean it's logically impossible to formally and specifically define the natural numbers without introducing a logical inconsistency. The best you can do is define a set that has all the properties of natural numbers but will also define things that aren't natural numbers as well.
As an analogy you could imagine trying to define the set of all animals with a bunch of rules... "1. Animals have DNA, 2. Animals ingest organic matter. 3. Animals have a nervous system. 4. ... etc..."
And this is true of all animals, but it will also be true of things that aren't animals as well, like slime molds which are not quite animals but very similar to them.
Okay so you keep adding more rules to narrow down your definition and stamp out slime molds, but you find some other thing satisfy that definition...
Now for animals maybe you can eventually have some very complex rule set that defines animals exactly and rules out all non-animals, but the principle is that this is not possible for natural numbers.
We can have rules like "0" is a natural number. For every natural number N there is a successor to it N + 1. If N + 1 = M + 1 then N = M. There is no natural number Q such that Q + 1 = 0.
Okay this is a good starting point... but just like with animals there are numbers that satisfy all of these rules but aren't natural numbers. You can keep adding more and more rules to try to stamp these numbers out, but no matter how hard, even if you add infinitely many rules, there will always be infinitely many numbers that satisfy your rules but aren't natural numbers.
In particular what you really want to say is that a natural number is finite, but no matter how hard you try there is no formal way to actually capture the concept of what it means to be finite in general so you end up with these mutant numbers that satisfy all of your rules but have infinitely many digits, and these are called non-standard natural numbers.
The reason non-standard natural numbers are a problem is because you might have a statement like "Every even integer greater than 2 can be written as the sum of two primes." and this statement might be true of the actual natural numbers but there might exist some freak mutant non-standard natural number for which it's not true. Unless your rules are able to stamp out these mutant non-standard natural numbers, then it is not possible to prove this statement, the statement becomes undecidable with respect to your rules. The only statements you can prove with respect to your rules are statements that are true of the real natural numbers as well as true of all the mutant natural numbers that your rules have not been able to stamp out.
So it's in this sense that I mean that it's not possible to specifically define the natural numbers. Any definition you come up with will also apply to mutant numbers, and these mutant numbers can get in the way of you proving things that are in principle true about the actual natural numbers.
It seems you know what you are on about! Thank you for a cracking comment.
I've always had this feeling that the foundations (integers etc) are a bit dodgy in formal Maths but just as with say Civil Engineering, your world hasn't fallen apart for at least some days and it works. Famously, in Physics involving quantum: "Shut up and calculate".
Thankfully, in the real world I just have to make web pages, file shares and glittery unicorns available to the computers belonging to paying customers. Securely ...
The foundational aspect equivalent of integers in IT might be DNS. Fuck around with either and you come unstuck rather quickly without realising exactly why until you get suitably rigorous ...
I'm also a networking bod (with some jolly expensive test gear) but that might be compared to pencils and paper for Maths 8)
if I'm working in Cursor for example, ideally the entire chat history and the proposed changes after each prompt need to be stored. that doesn't fit cleanly into current git development patterns. I don't want to have to type commit messages any more. If I ever need to look at the change history, let AI generate a summary of the changes at that point. let me ask an LLM questions about a given set (or range) of changes if I need to. we need a more natural branching model. i don't want to have to create branches and switch between them and merge them. less typing, more speaking.
I wonder if one of the existing interpreted languages (python/javascript/ruby/whatever) could maintain a patch for llmv/gcc which did exactly that and in the process make the most incredible seamless integration ever between itself and C (and also C++ now with its ABI stabilizing!)
reply