Recent developments have reinforced the view that in the coming years multiple, overlapping risks could trigger cascading societal and economic disruptions. The author draws on discussions with CEOs, government insiders and major investors and judges it plausible—not certain—that between now and 2028 we could see system‑level shocks such as AI‑linked cyberattacks and sudden economic turbulence.
How this differs from past crises
Past major shocks—like the 2008 financial crisis or the 2020 pandemic—were largely single, extreme events. The key concern now is a chain reaction: one shock provoking others faster than institutions and markets can absorb them. The article terms this dynamic “chain‑reaction chaos,” occurring in an era where the velocity of change outpaces the capacity of individuals, firms, markets or governments to fully process and respond.
Which risks are stacking up?
The piece lists several relatively high‑probability risks that, individually, would test already volatile markets and politics. Together they raise the odds of linked disruptions:
-
Cyber disruption:
- In recent days AI‑identified flaws in PHP—the software powering millions of websites—took government sites offline globally. The U.S. National Security Agency (NSA) urged organizations to patch immediately or risk outages.
- Major banks discovered unauthorized AI agents inside customer accounts, some attempting actions customers had not requested. Jason Clinton, deputy chief information security officer at Anthropic, warned that recent incidents will likely become “a lot worse before they get better.” Salesforce CEO Marc Benioff labeled the situation “Code red” on X.
- According to reporting cited in the article, AI companies and banks are already preparing “day after” plans, and some people at top AI firms estimate the chance of a larger cyber hit in coming months is well above 50%.
-
A partial AI investment bubble bursting:
- This situation differs from 1999 because large corporations with big balance sheets are underwriting the AI boom. Yet those investments are increasingly debt‑financed: Goldman Sachs expects roughly one third of Big Tech’s AI capital expenditures this year to be funded with debt.
- A dramatic collapse is not necessary to produce significant economic and political pain. Big companies are deeply interconnected as investors, customers and suppliers. The article points to Meta’s recent weak earnings and its comments about difficulties raising capital for AI as an example of how investor sentiment can be spooked. The author judges a sharp correction in the near term to be a greater than 50% possibility.
-
Rapid AI self‑improvement (RSI):
- Recursive self‑improvement could accelerate AI capabilities faster than most anticipate. People inside Anthropic, OpenAI and Google see this dynamic emerging. If managed well, this could yield long‑term benefits; but it could also produce acute disruption in 2027, threatening industries, companies and jobs and prompting political demands—potentially joining figures like Senator Bernie Sanders (I‑Vt.)—for a slowdown or pause, which would unsettle markets.
-
War escalation and an oil price shock:
- The article reports that advisers around President Donald Trump believe he is likely to resume conflict with Iran after an election; Iran has said it would expand targets in retaliation. Meanwhile, the U.S. has reduced staffing at multiple Middle East embassies amid increased threats.
- Escalation becomes especially consequential when oil supplies are tight and gasoline and diesel prices are already high. The author notes that sustained $120‑per‑barrel oil would cause substantial economic pain.
-
Risks tied to an older, less‑constrained president:
- The piece stresses this is not a partisan medical claim but an actuarial observation: Donald Trump, age 80, would be the oldest person to serve as president over the next two years, and the probability of cognitive decline and incapacitating health episodes rises substantially in one’s eighties.
- The author adds that similar concerns would apply to Joe Biden; both presidents, the article argues, lack surrounding figures who can reliably challenge them. The risk spectrum ranges from sudden medical episodes to consequential judgments about war and the exercise of presidential power.
Practical steps for individuals
The author frames these recommendations as pragmatic measures rather than alarmism. Three concrete actions are advised:
-
Be informed:
- Learn about AI—its benefits, risks and failure modes—and rethink what you expect from your employer and elected representatives in managing the technology.
-
Be vigilant about security:
- Lock down passwords: use a secure password manager and unique passwords per account, employ passkeys where available, and protect email and phone accounts with an authenticator app. Whoever controls those recovery channels can reset everything else. Do not give passwords or login codes to anyone—human or AI agent—who contacts you unsolicited.
-
Be financially prepared:
- AI is already a central market theme. If you have a 401(k) or other investments, understand how concentrated you are in particular sectors and pursue smart diversification strategies. Follow a few reliable news sources that can interpret rapid developments.
Bottom line
Any single one of the enumerated risks might not materialize. But when many plausible, high‑impact risks accumulate simultaneously, the overall probability of a crisis beyond direct control rises day by day.
