SPONSORED BY  | | | | May 22, 2024: Paint that washes its own surface, getting inside the mind of a chatbot, and could Neosporin in the nose prevent COVID? —Andrea Gawrylewski, Chief Newsletter Editor | | | • Relentless heat led to the deaths of 645 people last year in Phoenix, Ariz. The city is scrambling to come up with funding to avert another deadly summer. | 8 min read | | | • Ninety-year-old Ed Dwight, America's first Black astronaut candidate, finally flies to space on the Blue Origin rocket. Go, Ed!| 3 min read | | | Ed Dwight was recommended by the U.S. Air Force for the NASA Astronaut Corps in 1963 but ultimately was not among those selected. Robby Klein/Contour by Getty Images | | | Researchers have developed a new type of paint containing titanium oxide nanoparticles that can prevent pollution from discoloring surfaces and maybe even remove pollution from the air itself. When exposed to the sun's UV rays, the nanoparticles in the so-called photocatalytic paint activate and remove up to 96 percent of tested pollutants added to the paint's surface. The scientists say they can even make this paint from recycled materials.
How it works: When UV light hits this special paint, it excites the electrons in the titanium oxide nanoparticles in it. The result of this process, called photocatalysis, is the production of highly reactive hydroxyl radicals. These unstable chemicals attack pollutants that come in contact with the paint, converting them into less harmful substances such as carbon dioxide and water. By adding phosphorus, nitrogen, carbon, and other elements to the nanoparticles, the researchers were able to reduce the energy needed to spark the reaction so it worked with ordinary sunlight.
What the experts say: The researchers are testing next how effectively this special paint can clean surrounding air. "We don't claim that we can remove all contamination from air," says study co-author G├╝nther Rupprechter, a chemist at the Vienna University of Technology, "but overall, it looks promising." | | | What Are the Chatbots Thinking? | The latest AI chatbots use machine learning, in which software identifies patterns in data, without rules or guidance from humans, to generate responses. These patterns can be a complete black box, even to the humans who develop the software. Computer scientists and neuroscientists are quickly developing tools that might reveal how the large language models (LLMs) that power chatbots are coming up with their answers. To figure this out, some researchers are making model decision trees to replicate the functionality of LLMs, and others are engaging in chain-of-thought conversations (much like human psychotherapy) with the chatbots to determine how they are "thinking." Other researchers are scanning LLMs' neural networks while the machines "think" to determine how a response was generated
Why this matters: These inscrutable models are now providing medical advice, writing computer code, summarizing the news, drafting grant proposals and more. Yet it is well known that such models can generate misinformation, perpetuate social stereotypes and leak private data. Not to mention their startling instances of human-like reasoning or erratic behavior.
What the experts say: Some researchers think that the companies that release LLMs (like OpenAI) should also support the research into these chatbots (rather than keeping their technology secret). "Somebody needs to be responsible for either doing the science, or enabling the science," says David Bau, a computer scientist at Northeastern University in Boston, Massachusetts, "so that it's not just a big ball of lack of responsibility." | | | SPONSORED CONTENT BY SEMAFOR | What do people in Congress, executives on Wall Street, and UN Ambassadors all have in common? | They subscribe to Semafor Flagship, a premium newsletter that delivers transparent, intelligent, and need-to-know global news right to your inbox. It's the daily global news briefing you can trust – sign up for free. | | | • Tools for combating misinformation are built around the idea of a news consumer who is a well-balanced, ethical human who can vet information calmly and reasonably. But humans are much messier than that, writes Luke Munn, a research fellow in Digital Cultures & Societies at the University of Queensland, Australia. Real people operate on hunches, loyalties and grudges, he says. "To combat misinformation, we need to start from this actually existing human, someone who is emotional, factional and frictional."| 5 min read | | | • Decades ago, 3M executives convinced a scientist that the forever chemicals she found in human blood were safe. | ProPublica | | | • Rapid DNA analysis is helping to quickly identify victims of disasters. | MIT Tech Review | | | • After listening to thousands of hours of whale song, experts think Antarctic blue whale populations are doing well. | The Guardian | | | Paints that remove air pollution, concrete that absorbs CO2, roof shingles with built-in photovoltaic cells for capturing solar power--these exciting advances are already here (or coming soon) and promise to transform the built environment in the near future. Global cities produce up to 70 percent of global carbon dioxide emissions, use up vast quantities of water, degrade water quality and produce mountains of waste. They are prime for sustainable innovation! | Email me anytime with suggestions, comments and cool tech discoveries: newsletters@sciam.com. Thanks for reading! | —Andrea Gawrylewski, Chief Newsletter Editor | P.S.: Yesterday, I erroneously said that falling space debris has killed people on Earth's surface, which isn't true. Yet. | | Subscribe to this and all of our newsletters . | | | | Scientific American
One New York Plaza, New York, NY, 10004 | | | | | | Support our mission, subscribe to Scientific American | | | | | | | | | |
