Hello,
October is National Cybersecurity Awareness Month and CISA’s theme this year is “Securing the Next 250.” That’s less a slogan for posters in a break room and more an assignment. Our nation has just marked two and a half centuries, and our work now is to make sure the digital foundations of commerce, government, critical infrastructure, and daily life are ready for the next quarter-millennium. That work will not be accomplished through cyber-awareness theater. It will be achieved through ever-increasing velocity; by discovering, early and without sentiment, that which does not work, discarding what does not work, and concentrating talent and resources on what does.
For too long the field has treated cybersecurity technology design and analysis as a closed priesthood. Ideas and candidates move slowly. Weaknesses surface late. Ego and sunk cost keep mediocre approaches on life support. That tempo is a luxury we no longer have. Adversaries will not wait for our committee calendars. Neither will the systems we already operate.
Recent work with advanced artificial intelligence language models shows what a different OPTEMPO looks like. Teams working with Anthropic’s Claude examined a post-quantum digital signature candidate called HAWK that had advanced, for approximately two years, through NIST’s standardization process. The model did not invent a new branch of mathematics. Instead, it read the literature, combined known ideas, and located a symmetry in the scheme’s lattice structure that earlier analysis had not turned into a practical attack. The result was a faster key-recovery method against the proposed parameters. Restoring the intended security margin would have required larger keys and slower operations, erasing the size and efficiency that had made HAWK attractive. HAWK was withdrawn. That isn’t a scandal. It’s the process working as it should, only faster. No production system used HAWK. A flawed candidate left the field before it could consume years of implementation effort.
The same research refined an attack on a laboratory version of AES, the workhorse cipher that already protects bulk data across the internet. Full AES remains far beyond practical reach. Machines amplified and accelerated the efforts of human cryptographers, increasing the speed and thoroughness of literature review, reasoning, experiment, and runnable code so that the human experts could spend their time where it counts: choosing which questions matter, checking every claim, deciding what the result means for real systems, and choosing the best way forward. Velocity amplified the human contribution by concentrating it.
Leaders do not need to become technologists to use this lesson. They already oversee inventories, vendor roadmaps, upgrades, staging, and rollback. The same discipline applies here. Demand analysis tools and processes that raise the velocity of honest testing. Reward the early kill of a weak approach even when the approach has a household name for a champion, a budget line, or a conference paper attached to it. Protect the promising designs by giving them more scrutiny, not less. Refuse the consultant habit of inflating complexity until only the consultant can interpret it. Knowledge acquisition is a duty of leadership. Synergy between skilled people and capable models is not a threat to that duty. It is how the duty gets discharged at the speed the problem requires.
National Cybersecurity Awareness Month will produce four habits: strong passwords, multifactor authentication, phishing resistance, timely updates. Those habits matter; they make life harder for ordinary attackers. However, they are not sufficient for securing the next 250 years of American digital infrastructure. That task requires developing a different skill: the willingness to let evidence, generated faster than before, decide which technologies we keep and which we bury. Whose ox is gored is irrelevant. What survives contact with serious analysis is what we should field.
This isn’t new news. We can treat October as a communications campaign and return in November to the same slow cycle of proposal, attachment, and late surprise. Or we can treat it as a mandate to put better instruments in the hands of the people who design, buy, and operate the locks, and act on what those instruments find. One path wastes the years we still have. The other builds the foundations the next 250 will actually need.
Build it right, America.
Adam Firestone
Editor-in-Chief
Leave a Comment