Listen: https://serve.buildtall.systems/371c674b0c3bcb9dc4965779f09c6e2ce16fbde806356c77f94ac7f8454b9bda.mp3
Marginal Revolution Read Out presents the post 'Me on the cyberattacks.' The post opens with a piece by Tyler Cowen titled 'Don't Be Fooled By Randomness,' and the central message is that the recent problems are serious, but the sky is not falling. The argument makes three points. First, cyber risk is considerably higher than it used to be, because AI-driven offense has moved ahead of cyber defense, and plenty of experts say advanced AI agents can hack a wide variety of systems. Second, not every reported incident is actually a cyberattack. Some may be ordinary service outages that the world is paying more attention to than usual, and some may be institutions patching or upgrading their systems in ways that go wrong and cause an outage. Third, the risk of a larger and costlier AI-driven cyberattack remains real. The United States has redundancy and protective systems built into national security, financial institutions, and basic infrastructure, but those protections are imperfect. The author concludes that the country needs an emergency national emphasis on cyber defense, but people should not get carried away by anecdotes or social media. Cybersecurity has been a big problem for a long time.
The first supporting link is titled 'has outraced.' That phrase appears in the author's point that cyber offense has outraced cyber defense. The linked article is 'Tyler Cowen: How to Outrun the AI Hackers' in The Free Press. It reports that an AI system detected a critical flaw in the programming language PHP, and the news leaked before a fix was available. AI commentator Zvi Mowshowitz called it probably the worst day in cybersecurity history so far. The piece also raises deeper worries about cryptography. It cites Vitalik Buterin's warning that AI could break lattice cryptography within two years and notes that Anthropic's Claude AI solved a twenty-five-year NSA cryptography mystery in thirty minutes. The author's take is that this supports the claim that AI has given the offense an edge.
The next link is titled 'patching or upgrading their systems.' It points to a post by OSINTtechnical reporting that the NSA system outage was actually just a software update, according to The Wall Street Journal. The author's point here is that not every incident is an attack. Some apparent problems may be normal service outages that the world is simply paying more attention to, and some may come from institutions patching or upgrading their systems to limit future cyberattacks, only to have something in the upgrade go wrong and cause an outage.
The author then gives three examples. The first is the 'November 12 outage,' covered in a BleepingComputer report titled 'Massive Microsoft 365 outage disrupts Outlook, Teams, Copilot worldwide.' The article details a widespread Microsoft 365 outage that affected Outlook, Teams, Copilot, Azure, OneDrive, and other services. Microsoft initially attributed the disruption to elevated infrastructure load, but later said it was a technical issue, not related to cyber or AI agents. The company stated that the incident was introduced by an issue during an earlier network capacity upgrade and surfaced after a recent network change. The author's take is that this episode seems to have been something other than a direct cyberattack.
The second example is the 'November 13 ground stop.' This refers to the Federal Aviation Administration's nationwide ground stop on November thirteenth. The linked FAA question-and-answer document could not be retrieved, so its own details are not available here. What the author says is that the FAA ground stop, like the Microsoft outage, appears to have been something other than a direct cyberattack.
The third example is the 'November 16 outage,' referring to JPMorgan Chase's November sixteenth outage. The linked AP News report could not be retrieved, so its full contents are not available here. The author's commentary is that this, too, seems to have been something other than a direct cyberattack. The title suggests simply an outage on November sixteenth, and in the author's account it is one more data point in a week that looked alarming but may mostly reflect ordinary failures and increased attention.
To caution against overreading these events, the author invokes the title of Nassim Taleb's book 'Fooled by Randomness.' The linked bookstore page could not be retrieved, but the reference is clear: it is a warning not to be too ready to look for patterns in the data. The author's point is that a cluster of outages can feel like a coordinated attack even when it is not.
Finally, the post notes that it is taken from 'a broader Free Press symposium.' The symposium article, 'Is America's Infrastructure Under Attack?' by The Free Press editors, gathers H.R. McMaster, Tyler Cowen, Michael Shermer, and other experts to weigh in on the week's outages and the possibility of foreign interference. The introduction runs through the Microsoft outage, the FAA ground stop, the bank outages, reports that Iran is developing plans for a possible dirty bomb in Washington, and the National Security Agency's internal network going down. Officials said there was no evidence the events are connected. One contributor, Mark Montgomery, argues that what we are seeing is a series of coincidences. He says the outages were attributed to improperly executed upgrades or antiquated equipment and software, and he notes that more than a hundred network or software outages occur in the United States in any given week.
Taken together, the links in this post make a consistent argument: cybersecurity is a real and growing problem, especially with AI-assisted offense, but the recent cascade of outages should not be immediately read as a coordinated cyberwar. The author wants emergency attention on defense, paired with cooler judgment about individual incidents. That is the post 'Me on the cyberattacks.'
