Ed Yong is back writing again!
Most Americans simply aren’t thinking about COVID with the same acuity they once did; the White House long ago zeroed in on hospitalizations and deaths as the measures to worry most about. And what was once outright denial of long COVID’s existence has morphed into something subtler: a creeping conviction, seeded by academics and journalists and now common on social media, that long COVID is less common and severe than it has been portrayed—a tragedy for a small group of very sick people, but not a cause for societal concern. This line of thinking points to the absence of disability claims, the inconsistency of biochemical signatures, and the relatively small proportion of severe cases as evidence that long COVID has been overblown. “There’s a shift from ‘Is it real?’ to ‘It is real, but …,’” Lekshmi Santhosh, the medical director of a long-COVID clinic at UC San Francisco, told me. Yet long COVID is a substantial and ongoing crisis—one that affects millions of people. However inconvenient that fact might be to the current “mission accomplished” rhetoric, the accumulated evidence, alongside the experience of long haulers, makes it clear that the coronavirus is still exacting a heavy societal toll.
The company could have saved itself a giant headache by building in robust data record-keeping from the start, she says. Instead, it is common in the AI industry to build data sets for AI models by scraping the web indiscriminately and then outsourcing the work of removing duplicates or irrelevant data points, filtering unwanted things, and fixing typos. These methods, and the sheer size of the data set, mean tech companies tend to have a very limited understanding of what has gone into training their models.
she really gets it. Lots of interesting thoughts