WTF Does Mythos Mean (to Your Dad)?

Is it a Shibboleth? Is it God? Is it a Rhetorical Device? Is it Anything at All?

Jul 10, 2026

AI Changed Cyber

Cybersecurity is fundamentally a different field because of AI and has been since 2022. AI writing better attacks is not new, nor is AI helping defenders write code faster. What is new is that non-technical laymen can latch on the one word that describes this change: Mythos.

Last May, at a talk at Stanford, a portly Indian man who self-described as a “scout, analyst, and subject-matter expert for a high-net worth individual’s home VC office” asked me one question: What do you think of Mythos’s impact on the cybersecurity industry?

I gave the answer that, were Mythos to become widely available, both attackers and defending software engineers would have the same net advantage, resulting in a net-zero change for the game theory behind attacking infrastructure, and that defenders were already playing at disadvantage since the bottleneck is deploying patches quickly and architecting systems ground-up secure systems.

Often, Bay Area Tech Bros (BATBs) will in-group signal by discussing the latest AI lab drama or bandy about personal knowledge of friends who are employees at these frontier labs. This guy with a tech industry job in Silicon Valley should know about Mythos and should be curious on my thoughts as someone in the field it most affected.

However, it appears Mythos’ story has broken containment. This was probably the 12th or 19th conversation I’ve had on this exact subject, previously with family members or non-technical friends or people at church or randoms on the train. Mythos’ impact and later its very public government-mandated restriction seems to be on everyones’ minds, and has crossed what I call the MEH (the Mom Event Horizon) where even my mother has heard of it.

While my answer to this curious VC scout deserves to be evaluated in its own right, we should examine how he used Mythos in framing his question.

To this man, Mythos wasn’t a model or a chunk of code running on some Nvidia GPUs in Texas or Tennessee. Instead, Mythos was a rhetorical object attached to the field of cybersecurity that allowed laymen to grapple with the impact that widespread AI systems will have on the field of security and how it will relate to civilians’ concerns, which probably can be paraphrased with the following two questions: “Can someone break into my phone with AI?” and “Will AI hackers steal my money?”

Mythos is the English language word for anxiety about AI hackers.

Hackers and Load-Bearing Sports Analogies¹

Non-technical people with no exposure to the security industry can instinctively grasp the dynamic between attackers and defenders, and that attackers suddenly getting a bigger stick is bad news for defenders. The idea that “the attackers just added a star player to their roster and vastly increased their offensive capability” is easy to understand for someone who has likely used ChatGPT or vaguely remembers that time we killed the Iranian nuclear program with a flash drive or has ever watched football.

Patriots Suffer Big Loss: Gronkowski Will Retire - The New York Times — @gronk is this true?

No one in the security fields who pays attention should be surprised about Mythos’ purported abilities. Instead, we should be surprised with how much the world seems to know about this model and its abilities.

If we want to actually evaluate Mythos’ effect on the industry, we have to answer a lot of questions:

Is Mythos actually that good or is it marketing?

Is Mythos hard to replicate?

Is Mythos economical?

Will an American lab have a Mythos-level system in a year? In two? In five?

Will a Chinese lab have a Mythos-level system in a year? In two? In five?

Will there be an open-source Mythos in a year? In two? In five?

No one has the answers to how Mythos impacts cybersecurity, but a lot more people are thinking about these questions, all because the rhetorical object of Mythos made these conversations accessible to the masses. Such is the power of language post-social media.