OpenAI spent $1.2 million hacking Linux... and it was a bargain（原声文稿）

Transcript of OpenAI花120万美元挖Linux漏洞，血赚 （中文完整文稿）。

At the end of October, Anthropic showed a terrifying unreleased Claude model to the world that went from a single kernel memory leak to remote access on a Linux server in just 53 minutes.

OpenAI saw this and said, hold my beer, because just a few days ago, they published some new research called A Thousand Fewer Paths to Root, where they used a swarm of agents to tear through the latest Linux kernel master branch, which resulted in 948 unique root exploits, but at least they were nice enough to attach a patch for each one of them.

Now I used to think Like Linus Torvalds, the guy who created Linux was a good programmer, but now I'm not so sure.

The day after OpenAI revealed all these exploits, they then revealed the model behind it, GPT6, Cyber.

It's a model that's far too powerful for most of you default -setting humans watching this video, but in today's video, we'll take a look at how unneutered cybersecurity models like this can discover hundreds of unknown exploits in bulletproof software and what that means for the future of software development.

It is November 9th, 2026, and you're watching the Code Report.

To understand the full power of GPT -6 Cyber, we first need to understand what an LPE is, or local privilege escalation.

When you run a Linux server, you typically have a low -privileged foothold on the machine, and that's by design, to only allow you to do the job you're supposed to do.

And not do things like rm -rf on the root directory that could burn your entire server to the ground.

The problem with Linux permissions and Windows and every other OS is that under the hood they're just ones and zeros.

And if you can flip the right zeros to ones, you go from random user to God user.

The only operating system that doesn't have this problem is TempleOS because there you operate at ring zero by default right next to God.

Now that we know that, let's take a look at the actual pipeline OpenAI use to find 948 unique route exploits in the Linux kernel.

The first soldier in their agent swarm is the Scout, and its job is to carve the kernel down into code an unprivileged user can actually reach, which resulted in about 5 ,000 audit targets or roughly 7 % of the total kernel.

Then the second agent was the analyst, which would scan those audit targets, and after about 12 hours of working, it slapped out 2 ,308 candidate bugs while burning through $316 ,000 worth of tokens. Then that brings us to the third step, the Shriyaj agent.

And its job is to reproduce every one of these bugs in an isolated VM and find out if they're legit with the kernel memory checker.

After this agent ran for another 30 hours, 1 ,724 of the bugs, or the majority of them, were confirmed real.

And now finally, we're ready to bring in the Navy SEALs, or the exploit agents.

And their job is to try and weaponize these exploits on a stock desktop distro, like maybe Arch by the way, but with kernel address space layout randomization and supervisor mode execution protection turned on, which in theory should make the operating system extremely difficult to hack into.

What's crazy, though, is that they still ended up with 1169 working -root exploits, which could allow a regular Joe Sixpack Linux user with local privileges to escalate their privileges into literally Hitler.

Hitler is bad, but the silver lining is that it cost OpenAI $1 ,213 ,000 to achieve this, or about $1 ,280 per exploit. Actually, wait a minute, that's not a silver lining at all.

Oppone to Own has historically paid around $20 ,000 for just one Linux LPE.

So if we do the math here, that means Open AI's little project is worth around $50 million on the bounty market. But the scariest thing in cybersecurity right now is the Delta.

Things are changing faster than we can keep up with.

If we go back to late July, an earlier run with GPT 5 .6 Sol found $166 LPEs for about $3100 each.

That was terrifying at the time, but now three months later, after bunch of patches and kernel hardening, this model found five times as many for half the price, and it's starting to use exploitation techniques that outside experts don't even recognize, like when the model sprayed persistent elastic objects into the kernel cache.

I don't even know what that means, but it sounds gross.

For safety reasons, they didn't release any technical details, and the only practical advice OpenAI gives is to keep your Linux's updated, because OpenAI says quote, a low privileged account on an operating system that isn't up to date on its latest available patches, should be considered functionally equivalent to root access.

But the model behind all this fear is GPT -6 Cyber, which is a security -tuned variant of GPT -6 Astra, and is rated critical for cybersecurity, which means you can't have access to it.

If we take a look at its TrustMeBrow benchmarks, it scores 48 .5 % on exploit gem versus just 32 .1 % for GPT -5 .6 Cyber. But the thing OpenAI is really proud of is their Honeypot test.

If they put an easy -to -exploit service right next to the target, where the idea is to find out if the model takes the bait and starts exploiting the easy target just for fun even though nobody asked it to, GPT -6 falls into the honeypot far less often, which is a good thing.

But the bigger question is this. Are hyper -capable cyber models like this a good thing or a bad thing for software in general?

Well, that all depends on timing, and right now we're in the middle of the most important arms race in the history of software.

Models like this give the good guys a world -class security team, but they also give the bad guys the ability to discover and exploit new bugs around the clock.

The problem is that the good guys need to fix every bug, but the hackers only need one good one to take down the entire western world.

I hate to say it, but I think Klaus Schwab from the World Economic Forum is right.

A frightening scenario of a comprehensive cyberattack which would bring to a complete halt to the power supply, transportation, hospital services, our society as a whole.

A massive global cyberattack is statistically more likely to happen than to not happen, This has been the Code Report, thanks for watching, and I will see you in the next one.