Your daily source for science discovery, wonder and awe ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
September 16, 2025—A RAND report on whether AI could wipe out humanity, "time crystals" and why we love specific smells but find others off-putting. —Robin Lloyd, Contributing Editor | | A 3D rendering of the ozone hole evolution in 2025. CAMS | | Why We Love Specific Smells | The sense of smell is highly individual, according to a recent survey, but some trends emerged. For example, people tend to prefer the odors of more complex molecules than those of simpler ones. The team also collected replies to whether the smell of individual molecules matched any of 16 descriptive terms, such as "spices" or "burnt." Those results were mapped onto participants' collective pleasantness scores for the molecules, revealing that the most pleasant smells are described as "sweet," "flower" or "fruit." Unpleasant smells are associated with scents described as "decayed," "ammonia/urinous" or "sweat" reported Scientific American senior editor Clara Moskowitz, data journalist Miriam Quick and Scientific American senior graphics editor Jen Christiansen.
Why this matters: Artificial noses that can identify scents as well or even better than humans are the long-term goal of such research. "Digital olfaction" could be used to detect disease or household aromas, such as nearly burnt toast or rotting food in a refrigerator. How it works: In the graphic below, adapted by Christiansen for this newsletter, each black dot represents the median pleasantness and familiarity ratings of a specific molecular compound's smell. A correlation is clear: participants generally perceived more familiar smells as more pleasant (top right quadrant) and less familiar smells as unpleasant (bottom left quadrant). Participants also rated odors in six other dimensions: intense, warm, cold, edible, disgusting and irritating. Some secondary patterns are evident. For example, odors that smell edible are positively correlated with pleasantness and skew toward the right (blue). | | Miriam Quick and Jen Christiansen; Source: "A Dataset of Laymen Olfactory Perception for 74 Mono-Molecular Odors," by Antonie Louise Bierling et al., in Scientific Data, Vol. 12; February 2025 (data) | | A complete wipeout of our species by rogue artificial intelligence (AI) is plausible but very unlikely, according to a recent RAND analysis, writes Michael J.D. Vermeer, a senior physical scientist at the institute. He and his colleagues focused on three existential threats: pandemics, climate change and nuclear war. It's possible for two of these scenarios to work, but humanity should keep AI around for its benefits, the team reported. How it works: Nuclear war, while cataclysmic, would fail to wipe out all of us, the team concluded. AI couldn't kill us all with "garden-variety anthropogenic climate change," Vermeer and colleagues write, but it could produce and release more potent greenhouse gases at a large scale that could destroy all habitats fit for humans. Similarly, a single-pathogen plague would be insufficient, but a combination of pathogens could do the job provided it could find isolated stragglers. What the experts say: An AI effort to destroy all humans would have to overcome four major constraints, Vermeer writes. First, it would have to set an objective to achieve our extinction. Second, to carry out such a plan, it would need to gain control over key systems such as the means to launch nuclear weapons or manufacture chemicals. Third, it would require some help along the way from humans to hide its actions. Ultimately, it would require autonomy, to ensure total extinction. | | | | |
Can you unscramble this image? Reassemble the cover of our April 27, 1889 supplement, captioned "new treatment of locomotor ataxia by suspension." To celebrate Scientific American's 180th anniversary, we're publishing a jigsaw every weekday to show off some of our most fascinating magazine covers over the years. Take a tour here through all the cover-art jigsaws. | | We always like to hear from you. Please send thoughts, queries and other feedback to us at: newsletters@sciam.com. —Robin Lloyd, Contributing Editor
| | | | |
Subscribe to this and all of our newsletters here. | | | | |