The metaphor is a bit stretched for the purposes of content marketing a startup. The major difference between vulnerability researchers and the speedrunning community is that speedrunning is highly collaborative and open. There are massive speedrunning Discord communities for each game, and even before Discord existed, tricks and hacks were discovered iteratively just by many people watching other people do them often unintentionally and trying to figure out how they work (a common trend in every Summoning Salt video).
Nintendo doesn't care if people find ACE in decade-old games (usually) and post decompiled versions of games on GitHub so people can find out how they tick, but vulnerability researchers can't do that unless they want to risk causing a legal shitstorm.
Yes, the differences are substantial.
It's also worth noting that although some speedrunning may be akin to vulnerability research, the vast majority of speedrunners are "only" practicing and replicating exploits demonstrated by others. They're in different columns.
Also speedrunners are really bad historians. Their documentation is usually loose google docs links, placed on a discord channel. IF that sounds like there's little to none versioning and little to none searchability - precisely.
I am reminded of the top Super Mario players all congregating for a run on GDQ, only for a complete random fan to come to them and ask "So... if you were all stuck trying to optimize this one pipelfor over half a decade... Why didn't any of you just do the Devil's Spell?"
To which the speedrunner's reaction is: "Wtf is the Devil's Spell".
It happened to be a trick that was documented nearly 2 decades prior.
I think if they're active in the speedrunning community, then they're already well aware of this! And for a fun additional example to add to this article, you can often find TAS'ers talking about arbitrary code execution. The legendary GDQ run of TASBot's alternate ending to OoT[0] utiziling an ACE exploit they found in that game absolutely blew me away.
I love Ocarina of Time speedruns. The sheer level of love that went into that specific run was sooooo beautiful, and like the fact they made it internet live.. via an N64...?
I want to shout out ZFG if ppl arent aware cause he has IMO done the most technically impressive real time speedrun of any game - specifically the 100% SRM run he did is inscrutably insane. But it wasn't just about him - it was an effort by so many people. The number of glitches and exploits that have been found by the community, as well as the NP hard routing and tools created for finding angle perfect setups by various people..
It's straight up community driven exploit art. And it's like yeah, the fastest way to beat the game is to practically manually manipulate memory to redirect specific function calls to give you stuff you need and float around and purposely void out facing exactly a 1/65536 perfect angle setup a hundred separate times to randomly jump around to various rooms in the game?? Wowwwww
And the community around it is so wholesome. The sheer amount of collective curiosity, ingenuity, and effort to dismantle and exploit a 20+ year old game for no other purpose than going fast.. idk. Love it.
I forgot the details, but I think I saw a YouTube upload of a streamer who wrote flappy bird into super Mario by like… jumping at apples at specific times. Or some weird thing like that lol. I’ll try to find it later on my computer
TAS: Tool-Assisted-Speedrun. A kind of speedrunning, where control inputs aren't given by humans, but are carefully pre-programmed into a bot that will replay them. This allows to do things that would otherwise be veeeeery difficult (and sometimes impossible) for humans.
GDQ: Games Done Quick, the name of the Youtube channel.
OoT: Ocarina of Time, a beloved Zelda game from the 90s.
ACE: Arbitrary Code Execution. A vulnerability that lets you run whatever you want. You can use it to skip huge parts of the game, therefore achieving the fast speedrun
I've wondered myself why there's so little overlap between these two closely related interests of mine. Some of it seems to be the "But I don't want to cure cancer. I want to turn people into dinosaurs." effect, where some of the people working on exploiting games ONLY care about what can be done in their one game of interest - it doesn't always generalize to interest in using the same techniques against everything else.
Of course there's also the fact that exploiting 20-30 year old games is just vastly easier than modern software, due to the total lack of mitigations in them. And that's on top of the fact that with popular games, you're building on decades of reverse engineering work rather than (potentially) starting from scratch. And the arguably superior toolset (savestates etc).
But I think a very big factor is the one this blogpost is trying to address - most people just don't know anything at all about the vuln research industry, which is not exactly searching for attention in the ways that speedruns broadcast to hundreds of thousands of viewers for charity are.
Because actual gaming vulnerability researchers that do know who they are are called cheaters and are mostly active in cutthroat PvP games, not single player ones. Just ask the developers of Rust (the game, not the language), they know everything about it. They were one of the very few devs to ask the community to do what all communities in such games always do anyway - find exploits and glitches, and publish them on Youtube. As a result, they ended up with a game that is pretty robust to item duplication and general exploits.
Actually this implies there's probably an opening in YouTube for someone to make vulnerability videos in the style of speed run videos. And then poaching some of that audience, and riding the sponsorship opportunities. Not my skillset but yeah, I could definitely see that working.
For HN reference, MrCheeze is well known and has done quite a lot of work over the years glitch-hunting in older games. (and is cited in the SethBling video posted several times in this thread)
Since speedrunners who find glitches are obviously very technical, do they usually already have some sort of day job in tech? I imagine it might be easier and just as lucrative to work on some CRUD app 9-5 and devote the rest of their time to research/streaming, and may be preferable to overloading their brain with even more of the same kind of research.
I know a speedrunner who turned down a promotion beyond their data job because they were in a role that they already had automated a large chunk of, and wanted to stay in it so they could keep pretending to be busy at work while instead practicing speedruns.
As an n=1 data point, that was my exact situation for a while. Also a lot of the people who put out high effort stuff are college students, which works for the same reason.
More interestingly and more surprisingly, some of the people who work on exploiting games _don't_ do any sort of tech work and have no background in compsci - they're purely self educated just for the sole purpose of breaking the one game they're interested in. This was the case for some of the biggest contributors to ACE in Zelda Ocarina of Time.
This is absolutely and obviously true. Vulnerability researchers watch tool-assisted speedrun videos with jealousy. Side-note: when we did Microcorruption, game devs outperformed everybody but elite vuln researchers.
Microcorruption is basically just a Zachtronics game, if you squint, which I always thought was a fun framing. Reading the blog posts about Starfighter/Stockfighter definitely made me think of video game style exploits, too, if not the same type of glitches. Video game players love to find ways to sell items to NPCs and then buy them back at a lower price for infinite money...
That rules. That last level was Nick Carlini and Hans Nielsen, both of whom have done awesome things since then; we just interviewed Nicholas on cryptanalytic attacks against LLMs:
No doubt about this. Game-devs, modders, map-maker/hackers get hands
on such a breadth of skills (graphics, sound, scripts, network) inside
a contested environment that means a lot to them they naturally feel
at home in cybersec.
I watched the world record speedrun of Subnautica the other day and someone was kind enough to have posted a comment with a full list of all the bugs he exploited to beat the game in 28 minutes.
It was quite mind boggling. When I played the game I barely encountered a single bug or glitch - it seemed pretty polished! - but in actual fact there were 100’s of outstanding bugs, years after the game’s release and multiple updates.
If you look in the top-right corner at 0:19, the build being played is "Sep-2018 61056", strange for a Dec 2024 speedrun. Presumably that specific old version was used because the glitches it relies on have been fixed in current versions.
A lot of speedruns will not only use specific versions of online patchable games, but old games will have players use specifically Japanese or European physical copies for the same reason: there are glitches that are only present on the Japanese version of Pokemon Red/Blue that were fixed for the NA release, for example. Some of the time it gets really weird, like people specifically using the Wii Virtual Console re-release of a game in order to take advantage of its emulator being different from physical hardware - which is (usually) allowed, because it's still an "official release".
It's doubtful that there is one piece of software that does anything non trivial that has bugs, even short command line programs. The best we can hope for is that bugs don't seem to cause too many problems.
> It's doubtful that there is one piece of software that does anything non trivial that has bugs, even short command line programs.
TeX and Metafont and other Knuth programs come pretty close.
There's also programs that have been proven correct. (And the proof systems themselves are usually fairly small and can have multiple independent implementations. Or at least the core of these systems that everything hinges on.)
The important distinction, and where the comparison might fall short as the job-advertisement purpose of this post, is motivation. Speedrunners enjoy games because games are fun. Speedrunners get to actually use these vulnerabilities in a way that is meaningful in their lives, whereas vulnerability researchers typically don't.
This is an observation about cyber security in general, but in my experience, bug hunting and reverse engineering require a lot of tenacity at a level that writing software and other areas of IT do not. I think tenacity is a difficult thing to summon if your only tangible motivation is a salary, the target software is intrinsically boring, and you know that you'll be rewarded whether or not you find the bugs.
Though it's too bad that cyber security is not as intrinsically fun and interesting to a lot of speed runners as video games. A large part of what allows speedrunners to spend hours searching for glitches and exploits in these games is that they're having an absolute blast while doing it! Also exploiting glitches in decades old games is generally pretty accessible and doesn't have a high barrier to entry like cyber security.
Readit News
Nintendo doesn't care if people find ACE in decade-old games (usually) and post decompiled versions of games on GitHub so people can find out how they tick, but vulnerability researchers can't do that unless they want to risk causing a legal shitstorm.
I am reminded of the top Super Mario players all congregating for a run on GDQ, only for a complete random fan to come to them and ask "So... if you were all stuck trying to optimize this one pipelfor over half a decade... Why didn't any of you just do the Devil's Spell?"
To which the speedrunner's reaction is: "Wtf is the Devil's Spell".
It happened to be a trick that was documented nearly 2 decades prior.
So they are red teamers :p
Which ironically is more like pentesting than VR.
Just about every project has a bug bounty, but you'll be hard pressed to find any online discussion about works in progress towards a reportable bug.
[0] https://youtu.be/PNbkv_DJ0f0?t=3112
I want to shout out ZFG if ppl arent aware cause he has IMO done the most technically impressive real time speedrun of any game - specifically the 100% SRM run he did is inscrutably insane. But it wasn't just about him - it was an effort by so many people. The number of glitches and exploits that have been found by the community, as well as the NP hard routing and tools created for finding angle perfect setups by various people..
It's straight up community driven exploit art. And it's like yeah, the fastest way to beat the game is to practically manually manipulate memory to redirect specific function calls to give you stuff you need and float around and purposely void out facing exactly a 1/65536 perfect angle setup a hundred separate times to randomly jump around to various rooms in the game?? Wowwwww
And the community around it is so wholesome. The sheer amount of collective curiosity, ingenuity, and effort to dismantle and exploit a 20+ year old game for no other purpose than going fast.. idk. Love it.
Here's a commentated tool assisted human-like run (but not live): https://www.youtube.com/watch?v=R8EE9FXeJnE
And the actual run: https://www.youtube.com/watch?v=Sdxdwnpi-wU
here is the link for anyone curious
GDQ: Games Done Quick, the name of the Youtube channel.
OoT: Ocarina of Time, a beloved Zelda game from the 90s.
ACE: Arbitrary Code Execution. A vulnerability that lets you run whatever you want. You can use it to skip huge parts of the game, therefore achieving the fast speedrun
Of course there's also the fact that exploiting 20-30 year old games is just vastly easier than modern software, due to the total lack of mitigations in them. And that's on top of the fact that with popular games, you're building on decades of reverse engineering work rather than (potentially) starting from scratch. And the arguably superior toolset (savestates etc).
But I think a very big factor is the one this blogpost is trying to address - most people just don't know anything at all about the vuln research industry, which is not exactly searching for attention in the ways that speedruns broadcast to hundreds of thousands of viewers for charity are.
More interestingly and more surprisingly, some of the people who work on exploiting games _don't_ do any sort of tech work and have no background in compsci - they're purely self educated just for the sole purpose of breaking the one game they're interested in. This was the case for some of the biggest contributors to ACE in Zelda Ocarina of Time.
https://securitycryptographywhatever.com/2025/01/28/cryptana...
It was quite mind boggling. When I played the game I barely encountered a single bug or glitch - it seemed pretty polished! - but in actual fact there were 100’s of outstanding bugs, years after the game’s release and multiple updates.
If you look in the top-right corner at 0:19, the build being played is "Sep-2018 61056", strange for a Dec 2024 speedrun. Presumably that specific old version was used because the glitches it relies on have been fixed in current versions.
TeX and Metafont and other Knuth programs come pretty close.
There's also programs that have been proven correct. (And the proof systems themselves are usually fairly small and can have multiple independent implementations. Or at least the core of these systems that everything hinges on.)
This is an observation about cyber security in general, but in my experience, bug hunting and reverse engineering require a lot of tenacity at a level that writing software and other areas of IT do not. I think tenacity is a difficult thing to summon if your only tangible motivation is a salary, the target software is intrinsically boring, and you know that you'll be rewarded whether or not you find the bugs.
Though it's too bad that cyber security is not as intrinsically fun and interesting to a lot of speed runners as video games. A large part of what allows speedrunners to spend hours searching for glitches and exploits in these games is that they're having an absolute blast while doing it! Also exploiting glitches in decades old games is generally pretty accessible and doesn't have a high barrier to entry like cyber security.
EDIT: There was supposed to be a link here. https://www.youtube.com/watch?v=jnZ2NNYySuE