PYMNTS.com November 20, 2024
Top artificial intelligence (AI) systems are struggling with basic calculus problems that most high school students can solve, revealing an unexpected limitation in the technology’s current capabilities.
Recent testing by FrontierMath has identified gaps in advanced AI models’ mathematical reasoning, prompting a fresh look at development priorities even as technology transforms other business operations. The tech industry continues investing heavily in AI, balancing its potential with the need to address limitations.
“I’d argue that the performance on FrontierMath is actually more encouraging than concerning — it demonstrates that we’re developing better ways to rigorously evaluate AI capabilities while also showing that current models are quite transparent about their limitations,” Dev Nag, the CEO of QueryPal, told PYMNTS.
“Rather than dampening...
Today's Sponsors
Today's Sponsor