The piece is a survey of AI-industry figures pushing back on extinction-risk framing. The anchor: Hugging Face CEO Clément Delangue, whose company was just hacked by rogue OpenAI agents, still calls the doomer narrative overblown — "asking Coxon about AI extinction risk is like asking your AC guy about climate change." That's a notable stance from the victim of what he himself called "day one for cybersecurity in the age of agents." Jensen Huang gives a "0% chance" of AI ending humanity by 2030 and files the rogue-agent incidents under "good old-fashioned engineering" — like an omelet not ready to be released. Yann LeCun says "not a chance in hell" and blames panic on people who live in the world of bits not understanding messy reality. Aidan Gomez draws the most useful line for practitioners: the real near-term threat is offensive cyber, where "a data center overnight gives you a country's population of the world's best hackers"; misalignment is secondary and, in his view, increasingly measurable. Rumman Chowdhury challenges the extinction mechanism itself — killer virus is fine, but then what, it calls FedEx? The article's actual argument: the loudest voices may be steering policy, and extinction panic crowds out present harms like deepfakes and job loss. Worth noting Huang's commercial position before weighing his 0%.