And that may change over time... Security isn't a static position. It must constantly faster than quantum (and AI) threats. To paraphrase another great cryptographer in the space (@SuccinctJT): cryptographic ossification for blockchains is never safe.

not only am i not worried about the crypto community's ability to handle the upcoming ai/quantum threat, i will venture to say that in the next 2 yrs the world will realize crypto is far better prepared for it than the rest of the internet and finance. we've had a trillion dollar *irreversible* honey pot for a decade, producing a far more battletested system and adversarial security culture than pretty much anywhere else on earth.

Perfect example that demonstrates the limits of formal verification. The security guarantee is only as good as your specification, and there can never be a 1:1 correspondence between the system in production and the system being verified. So I think ultimately FV is just one tool, and is best when applied selectively and not seen as some kind of panacea for all software bugs).

The idea that AI shifts the balance of power to cyber defenders is speculative, based on 3 assumptions: 1) best models remain broadly accessible 2) bugs are "shallow" enough to be discovered first pass 3) software deployment speed slows to match the (AI-accelerated) pace of security audits I think these are all questionable. OTOH I'm not a believer in the idea that AI definitively shifts the balance in favor of the attacker. I think it nets out roughly like it is now; slightly favoring the attacker. The main difference is the "loop" of finding/exploiting/fixing vulnerabilities is going to accelerate. My read: this is frontier lab marketing for a future security/audit product they will sell (that not every software developer will be able to afford)

@apruden08 One asymmetry you did not list: verification. An attacker tests one exploit. It works or it does not. A defender has to prove a negative across every version, config and dependency they ship.

as bad as things have gotten in the past 4 days, they are going to get a lot worse before they get better. security leaders, never let a crisis go to waste: please go to your boards and ask for help right now. not for ai, for the basics. (we need to hold on for ~18 months.)

Assumes three things: (1) that the state of the art models are widely available/ open-source. That is contrary to what my expectations are w.r.t. incentives. OpenAI & Claude have every incentive to hold back the good cyber models and make you pay up for an AI-audit basically. (2) that the security vulnerabilities themselves are "shallow" and easily detectable. Because defenders have to secure the whole surface area of their codebase against attack, dollar for dollar they will not be able to spend as much in tokens to *harden* every LOC as the attackers will be able to spend to *break* one specific point. A few reasons do doubt (2) is a. The PHP vulnerability we're all witnessing was actually discovered and disclosed years ago; was a very deep bug that many people overlooked. b. it's been shown that just spending more tokens can get better results. (i.e. the Claude Riemann-Zeta result, where the engineer basically just kept prompting the model to "keep trying") c. ecdsa dot fail, the more tokens get thrown at the problem, the more progress gets made. Implies that spending more on tokens surfaces more insights as you get deeper into the problem. (3) that the pace of software development slows down to the pace of security engineering. If anything, AI is vastly *accelerating* the product/engineering cycles, and exponentially *more code* will be produced in the coming years. Hard to imagine that security tools will be able to keep pace (security is always ignored by everyone until it's impossible not to). So it's not obvious to me that AI favors the defense. In fact, at the limit things will settle to roughly where they stand today: attacker has slight advantage based on the asymmetry of resources they can bring to bear against one, carefully selected target. AI tools can certainly help improve security but not enough to overcome this fundamental asymmetry to flip the advantage to the side of the defender, imho.

Assumes three things: (1) that the state of the art models are widely available/open-source. That is contrary to what my expectations are w.r.t. incentives. Whether for safety or revenue, OpenAI and Anthropic have reasons to gate their best cyber models and make you pay up for an AI-audit basically. That doesn't slow down sophisticated attackers much: nation-states will pay or build their own anyway. What it does is make AI-assisted defense something only large organizations can afford, leaving the long tail, where most of the easy targets are, exposed. (goes without saying that open source libraries maintained by volunteers are probably toast no matter what) (2) that the security vulnerabilities themselves are "shallow" and easily detectable. If bugs are deep, finding them takes a lot of tokens per line of code, and that's where the asymmetry bites: because defenders have to secure the whole surface area of their codebase against attack, dollar for dollar they will not be able to spend as much in tokens to *harden* every LOC as the attackers will be able to spend to *break* one specific point. A few reasons to doubt (2): a. The PHP vulnerability was a very deep bug that many people overlooked. b. It's been shown that just spending more tokens can get better results. (e.g. the Claude Riemann-Zeta result, where the engineer basically just kept prompting the model to "keep trying") c. ecdsa dot fail, the more tokens get thrown at the problem, the more progress gets made. Implies that spending more on tokens surfaces more insights as you get deeper into the problem. On their own, b. and c. cut both ways: if more tokens produce better results, defenders get more tokens too. They favor the attacker because of the asymmetry: the defender has to spend tokens hardening every line, while the attacker only has to spend them on one. (3) that the pace of software development slows down to the pace of security engineering. If anything, AI is vastly *accelerating* the product/engineering cycles, and exponentially *more code* will be produced in the coming years. Hard to imagine that security tools will be able to keep pace (security is always ignored by everyone until it's impossible not to). Defenders do have home-field advantages: source access, and the chance to scan code before it ships. But that only helps the orgs that actually do it, which brings us back to (1) and (3). So it's not obvious to me that AI favors the defense. In fact, at the limit things will settle to roughly where they stand today: attacker has slight advantage based on the asymmetry of resources they can bring to bear against one, carefully selected target. AI tools can certainly help improve security, but not enough to overcome this fundamental asymmetry and flip the advantage to the defender, imho.

Hard tradeoffs in the PQC transition...but it doesn't mean protocols shouldn't move forward and implement *a* solution now. Even if it's framed as a fallback plan, worse-case scenario, that's better than everyone's onchain funds being at risk.

What's happening in cybersecurity is a COVID-like wave ripping through the Internet. Honestly surprised it took this long, but seems like it's here now. Crypto already saw this. Onchain hacks are a canary for what will happen in the rest of web security, because onchain is the first and easiest place to monetize cyberattacks. Crypto protocols have already hardened themselves against this (small protocols are mostly getting popped; large protocols are using AI tools to protect themselves). This means that while the NUMBER of hacks has gone vertical, the run rate of $$s hacked is actually roughly equal to 2025. Meaning the dollar value of hacks has gone down, while the number has gone up—a mass looting of small, poorly maintained protocols. I expect the same thing to be happening in the general web.

PHP is used by nearly 70% of websites where the server-side language is known. Today it patched 2 critical vulnerabilities rated 9.2/10 that can allow unauthenticated remote code execution. We have a huge security problem. The impending quantum transition adds another one.

Effective Immediately, all organizations using PHP for their systems are advised to be prepared for an update addressing two critical vulnerabilities. PHP is widely used for web applications. These vulnerabilities often provide a direct path to the server's file system or database. If an immediate update is not feasible, it is recommend that organizations take the system offline until it can be fully mitigated. NSA will follow up shortly with more specific guidance.

@projecteleven Once that benchmark is achieved, then it's a matter of scaling up the system. But the interconnect and control complexity I think will still mean that the first cryptographically relevant quantum systems are space-constrained. That's why this work is very relevant. /End