The next big breakthrough in AI could be one that channels the spirit of famed scientist Albert Einstein. “The problem is that large language models [LLMs] make mistakes. And they’re always going to make mistakes, but they’re going to make fewer and fewer. At some point, they’ll make so few mistakes you can use it in medicine, you can use it in law,” Groq founder and CEO Jonathan Ross said on Yahoo Finance’s Opening Bid podcast. Ross said the ability of AI models to invent is among the next wave coming, and it’s likely to happen within his lifetime. Currently, LLMs tend to pick the most probable answer to questions, which precludes them from discovering something new. “And if it’s the most obvious, it’s not going to be good writing. It’s not going to be good science. It’s not going to invent something for you. It’s not going to invent a drug that isn’t known. That’s what we’re working on next. Invention is next,” Ross explained. While at Google, Ross designed the custom chips that the tech giant would go on to use for training its AI models. Today, Ross has a ground-floor view into how LLMs are likely to continue to evolve in the years ahead.

Full podcast interview : Groq CEO, Jonathan Ross says that large language modes will make fewer mistakes going forward.