Unlike pure LLMs, which rely solely on neural networks' pattern recognition abilities, Neurosymbolic LLMs first convert natural language instructions into symbolic programs and then use external symbolic interpreters, such as programming language interpreters, to execute them.
Symbolic AI has been an approach to AGI that has had continuous research since the very beginning. If there are real advances being made in this field, then it's big news, but this article is, ironically, very obvious A.I. slop itself.
Avoid the repetitive verbal diarrhoea and go straight to the papers that it links:
"Neuro-Symbolic Concepts" by Jiayuan Mao, Joshua B. Tenenbaum, Jiajun Wu