The debate over how quickly artificial intelligence should advance is increasingly moving from AI laboratories into the security operations center. A recent LinkedIn post by Asaf Wiener, CEO and co-founder of Mate Security, has become a focal point for that conversation, arguing that efforts to slow frontier AI development could leave cybersecurity defenders operating at a disadvantage.
Wiener was responding to a recent letter from Anthropic CEO Dario Amodei, who has advocated for “pacing the frontier” as AI systems become more capable. Wiener’s counterargument is that cybersecurity does not operate according to the same rules as AI development. Companies can coordinate around responsible development, he argues, but attackers have no obligation to participate.
“Pace the frontier, and you pace exactly one side of this fight, the defense side,” Wiener wrote in the LinkedIn post. “The answer isn’t to slow defenders down. It’s getting trusted, verifiable AI into their hands faster, with the right guardrails in place.”
The argument struck a chord among cybersecurity professionals on LinkedIn. The discussion includes comments from professionals outside Mate Security, turning the post into a broader industry conversation rather than a company-centered announcement.
The Other AI Frontier
Wiener’s central argument is that there are effectively two AI races happening at once.
The first is the familiar race among frontier AI developers to build increasingly capable systems. That is the frontier addressed by Amodei’s proposal.
The second is happening among attackers who are already using available AI capabilities to automate reconnaissance, generate malicious content, accelerate analysis and potentially compress the time between intrusion and impact.
That second frontier is what Wiener believes deserves greater attention.
“Attackers didn’t sign the OpenAI letter, didn’t answer to a three-step plan, and were never going to wait for a ‘so-called decision’ to slow down,” he wrote in Mate Security’s accompanying analysis.
The distinction has resonated with commenters. Omer Karny wrote, “Exactly. Thinking that everyone will pace themselves because someone said it- just won’t work…” Inbal Argov took the discussion further, writing that organizations should focus on “the mechanisms around the models that make them trusted and operationalized, and hunting down the frontier-powered attackers.”
A Question Of Speed
The issue is ultimately less about whether AI should have safeguards and more about where those safeguards should be applied.
Wiener agrees with the premise that increasingly capable AI systems require verification, accountability and controls. His disagreement is with applying restraint primarily to the development of frontier capabilities while defensive teams remain slower to adopt comparable technology.
For security teams, speed has become a central part of the equation. An attack that can be automated and executed continuously does not necessarily give analysts the same amount of time to investigate, validate and respond.
That changes the economics of defense. Security organizations have historically relied heavily on human analysts to investigate alerts and coordinate responses. AI agents introduce the possibility of performing portions of that work at machine speed, provided organizations can establish appropriate controls around what those agents are permitted to do.
The Conversation Moves To LinkedIn
The engagement around Wiener’s post illustrates how quickly these questions are becoming part of the public cybersecurity conversation.
Sharon Rosenman commented, “Slowing down innovation is not a good idea and will practically not happen. Great insight.” Noam Bar-Lev wrote, “100%. Slowing things down gives attackers the edge.” Chris Hoard described the discussion as “A good read. Thanks.”
Not every reaction needs to endorse the entire argument for the underlying question to gain traction. The comments show professionals engaging with the tension between AI safety, technological acceleration and defensive readiness.
For Wiener, the practical implication is that cybersecurity organizations should not wait for AI-enabled attacks to become more sophisticated before adapting their defenses.
The Mate Security CEO’s argument is ultimately about closing the gap between offensive and defensive capabilities. If AI is going to transform both sides of cybersecurity, the question may no longer be whether organizations should adopt AI for defense, but whether their defensive systems can develop quickly enough to keep pace with the threats they are designed to counter.
