Live data from Hacker News

Frontier AI has broken the open CTF format

kabir.au

391–400 of 502 posts

Re: Frontier AI has broken the open CTF format

#391
post #223
post #6

https://en.wikipedia.org/wiki/Capture_the_flag_(cybersecurit... still has no mention of AI, but that will likely change as they increasingly dominate competition.

Using AI on CTF is like using a car to get better at the 100 yard dash

Except some people want to win and don't care about their own personal performance.

Re: Frontier AI has broken the open CTF format

#392
post #380

Must I beg to have an acronym spelled out a least once, the first time it's used? Even if you assume 90% of readers already know, the other 10% (including me, in this case) will thank you, it doesn't take much effort, and it expands the reach of your communication or idea. Exceptions for cases where the acronym is just so well known that a lot of people don't even know what it stands for even though they know the con…

I didn't know what BGP is, but I did know CTF. YMMV

Hah, same. Always entertaining how people map their tendencies to the world at large.

Re: Frontier AI has broken the open CTF format

#393

I feel the post. For me AI has ruined both, playing CTFs and also building CTFs challenges. The most annoying thing to me is the "yeah idk but here is the flag" mentality. Before when playing CTFs with my mates was usually sitting there for hours tackling a challenge until some other mate joined, had some look together and solved it with you together in 30 minutes which is the most rewarding learning experience. Nowa…

I don’t know like chess engines didn’t kill chess. You could just play with people that don’t use the “engine”

It’s different, unfortunately. I wish you were right. The problem is that creating interesting and fun CTF challenges is a very active, time consuming, creativity-heavy task. A chess board is always the same, and always will be, but every CTF competition is unique. There is little to no incentive anymore to spend time creating the challenges. You might say “well create the challenges and share them with people who care and who want to play honestly” which is probably the right answer here, and might happen at a smaller scale. I picture CTF in the future almost like a tabletop RPG experience, one where a small amount of people will share with close friends who they trust. But the usual “open” CTF scene (as mentioned by op) is probably over for good, if we’re being honest.

Re: Frontier AI has broken the open CTF format

#394
post #223
post #6

https://en.wikipedia.org/wiki/Capture_the_flag_(cybersecurit... still has no mention of AI, but that will likely change as they increasingly dominate competition.

Using AI on CTF is like using a car to get better at the 100 yard dash

Yeah but for the brief time window (which is probably coming to an end right now) where getting your name on the leaderboards was still worth it anyway, because people had not yet realized that the game is over, players will use the car anyway. Now that the game is over, leaderboards are meaningless, so we will figure out ways to move past that (like playing with people who care). But that will change the game in unavoidable ways: the tiny, fragmented scale will give less incentive for creators to come up with massive intricate complex challenges (it would be like hoping Bethesda would make Skyrim for a handful of die hard players). And soon, maybe on a few years, people will invariably start questioning whether it even makes sense to waste their time with this hobby if learning CTF skills is basically useless in an AI dominated world. There are still people learning Assembly for fun, but almost nobody does Assembly programming challenges as a hobby.

Re: Frontier AI has broken the open CTF format

#395

>The competition is turning into "who can afford to run enough agents, with enough context, for long enough." This will basically become true for everything.

It’s unavoidable, but really sad, isn’t it? Thinking about the incredible creativity and hard work that went into creating such challenges and now it’s probably history. I feel something similar with free-to-play gacha video games. The gacha mechanics are slowly creeping into every type of game, and where you once had very clear and obvious slot machines, now you get them transparently mixed in with beautiful and fun games (e.g. ZZZ, Where Winds Meet, Genshin, etc) in a way that sounds like in the near future no company will have any incentive to not have gacha mechanics in any game.

Re: Frontier AI has broken the open CTF format

#396

Earlier quoted context omitted.

Which acronym do you mean? CTF? I think that acronym, just like BGP, is more well known by itself than what it stands for. More generally, not every piece of writing is meant for every audience. Like if someone writes a blog post about CTFs aimed at people who like CTFs, nobody in the target audience needs to have CTF explained to them. Ultimately HN is a link aggregator, but sometimes its a bit like eavesdropping on…

I dont know what CTF stands for so I dont know if I am interested in this article or learning anything about it. Maybe I am. Are you really arguing for not just typing out whatever 3 words this stands for once in the name of clarity?

[dead]

Re: Frontier AI has broken the open CTF format

#397

Earlier quoted context omitted.

Best practice in writing about technical concepts is to spell out acronyms like this on their first use. There is a ton of stuff I learn about here on HN that I didn't know anything about before. It doesn't help that the linked article never bothers to explain this either.

For a general audience this is good advice. This article was written for a specific audience who follows this blog because they know the term. If you start spelling out fundamental acronyms it makes the content look more basic and general. This always upsets the general audience who stumble upon the article (like this) but it wasn’t meant for a general audience. CTF is extremely well known and the people who would be…

It's the common practice in technical writing. Even when you are writing for other experts in the same field, your target audience never shares as much context as you would prefer. The world is much weirder and more varied than you would think.

Informal writing about technical topics is another story. There you can assume a lot more shared context, as you are only writing for a specific subculture within the field. It doesn't matter much if other people in the field fail to understand you.

Re: Frontier AI has broken the open CTF format

#398

Earlier quoted context omitted.

Since this is the top comment at the moment: CTF stands for Capture The Flag. Personally I have never, ever heard that concept referred to by the initialism. Granted, it's almost never come up in my circles, so... shrug

Just to give the actual answer, CTF in this context means a computer security competition. Generally the way they work, is you get some programs, and you have to hack them to get some string called the flag (e.g. maybe the server has a root owned file called flag, so you have to get root somehow to read the file). Team with the most flags at the end wins. In this context, CTF is almost exclusively referred to by the…

THANK YOU

This has been the most annoying HN so far until your comment appeared.

Re: Frontier AI has broken the open CTF format

#399

Kinda FUD article... the reality is that common problems are going to be easy because the solution is probably inside the training dataset, the challenge should be adapted to make LLM's useless for example once at Defcon CTF the problems were for an unknown CPU architecture based on octal that required to write even your own disassembler... this are the kind of things that will probably be hard for frontier LLM's

Speaking from experience, the LLM agents adapt fairly well to these contexts too. It's not at all FUD, you're at a significant disadvantage if you don't compete with AI now. I went to a CTF recently against teams I have won against every year, and within 10 minutes of the event starting they had solved every challenge. They have an agent loop and it solves everything immediately, so they won. Anyone attempting to solve the challenges on their own has no chance, even if you think "maybe this is too out of the box for LLMs". Furthermore, the DEFCON CTF you're referring to has quals, and if you don't qualify you don't get those challenges in the finals. Quals has mainly binary exploitation challenges which Opus (and others) solve as long as you hold the gas pedal down on your API bill. I don't believe it's hyperbole to say CTF is dead, as a competitor.

Re: Frontier AI has broken the open CTF format

#400
post #171

Earlier quoted context omitted.

Fizzbuzz is such an incredibly simple problem if you can’t do it I struggle to see how you’d be able to complete any task that requires very basic reasoning and very basic coding knowledge. And if an AI system can do those parts, what am I getting for spending tens of thousands of pounds per year by hiring a person who can’t? Wouldn’t I just tag codex on the tickets? I’m not talking about gotcha level stuff here wher…

I feel the same way about inverting a binary tree, but a lot of people act like it's an arduous request. I am guessing it's because they've never read the description of what inverting a binary tree is, but maybe people are just that bad at recursion.

Yeah, you read it and expect the inversion to be inside-out or upside-down, defined in some hopelessly arcane way.
Post reply on HN