That's what they'll sell you. But fundamentally current LLM implementations seem to be the wrong methodology for 'AGI' and I don't think we're remotely close to having this.