Osmo, a New Jersey-based AI startup, is working to give computers a sense of smell. At its lab, researchers catalog odors using descriptive terms such as “nutty,” “peachy,” and “marine,” then cross-reference those notes with chemical analyses to train what the company calls the world’s first foundation model for smell. Founder Alex Wiltschko says the goal is to give machines olfactory capabilities, and he claims the model has already reached “human-level olfactory intelligence.” Osmo says the technology is already helping fragrance brands develop new scents, and it has broader ambitions, including detecting counterfeit goods and diagnosing diseases. Skeptics remain unconvinced that the system can deliver on those promises, leaving open the question of how far AI-powered smell could really go.
Elsewhere, hydropower is under strain. Data from the US Energy Information Administration shows the US Southwest generated 28% less electricity from hydropower between April and August than the ten-year average. Most American hydropower relies on dams that hold back reservoirs, and an extreme, prolonged drought has pushed the region’s two largest reservoirs to dangerously low levels. Climate change is compounding the problem, putting pressure on a key source of renewable power.
The grid faces its own challenges. Transformers, though nearly invisible, are essential for moving electricity efficiently over long distances, yet they are complex, in short supply, and largely unchanged in design for about a century. As demand grows from electric vehicles, data centers, and other sources, companies are developing solid-state transformers that could ease supply shortages and reduce power losses. The technology is still in development, but it may become an important part of the future electric grid.
Source: The Download: AI learns to smell and hydropower struggles via MIT Technology Review
