Autumn always arrives quietly, but the world rarely does. As the days shorten and the season turns, two stories in particular have been pulling at that quiet this week — one unfolding on the ground, one unfolding in the space between labs and podiums.
Yemen’s Houthi rebels have just pulled off one of their biggest territorial gains in years. Over the past week, they swept down Yemen’s western coast, capturing the port town of Mocha, the strategic Perim Island, and — this weekend — the Greater and Lesser Hanish islands in the southern Red Sea, giving them effective control over the Bab al-Mandeb Strait, one of the world’s busiest shipping chokepoints. The Houthis say they seized roughly 5,400 square kilometers of territory in the process. Saudi Arabia’s East-West oil pipeline was reportedly struck on 11th September and Yemeni military leaders say they requested Saudi air support to hold the line — but according to Axios, when Saudi Crown Prince Mohammed bin Salman asked Trump directly on 10th September for US airstrikes against the advancing Houthis, Trump declined, choosing instead to offer intelligence and targeting data rather than direct force. Two days later, Trump told reporters the Houthis “called u, and they don’t want to fight with us… they’re letting most ships go through.” Yemen analyst Mohammed Al-Basha told CNN that “all signs now point to a large-scale Saudi retaliation and that likelihood is increasing by the day” — while Saudi Arabia insists it doesn’t want all-out war. It’s a strait squeezed from two sides now: Iran leaning on Hormuz, the Houthis leaning on Bab al-Mandeb and Washington, notably, choosing to watch rather than intervene.
Meanwhile, a very different kind of existential argument has broken out inside the AI industry itself. Last week, Anthropic researcher Jacob Coxon resigned, writing that leading AI labs are “racing straight to self-improving superintelligence and gambling with our lives.” His former colleague Evan Hubinger, Anthropic’s alignment research lead, backed him publicly, saying he personally believes there is a “greater than 10%” chance AI could kill all humans within the next decade. He’s not alone — AI pioneer Geoffrey Hinton has separately put similar odds on catastrophe and Yoshua Bengio has gone further. Anthropic’s CEO followed with a public essay calling for the industry to slow down. Trump’s response, when asked Thursday whether he shared any of those concerns: “No, I don’t have any… I have concerns that if we don’t win AI, we’re going to be put in a very bad position.” Asked directly about guardrails against AI turning on humanity, he added: “It’s going to be fine. We’ll always have something to stop them. We’ll have a little gear. Boom.”
Join us for this week’s Global Risk Forum sessions, 461 and 462 — come share your take and whatever else around the world has been on your mind.
- Global Risk Forum 461
- Thursday, 17th September at 16:00 BST
- REGISTRATION >>
- Global Risk Forum 462
- Friday, 18th September at 09:00 BST
- REGISTRATION >>
