Is AI slowing down? Is AI at the end of the road? Even if the LLMs (which is only a subset of AI but

Is artificial intelligence plateauing? You've probably heard arguments recently that artificial intelligence is hitting a plateau and it's not going to get that much better. There are lots of clickbaity headlines like this one with the CEO of Google saying that AI is slowing down and the low hanging fruit is gone. I saw this tweet from John Rush and I thought it was a really interesting perspective. He's saying that even if we're hitting a plateau and the large language models aren't going to get that much better because there's not that much more data to train them on, if we compare it to the human brain and the fact that our brains haven't got that much bigger in the last couple of hundred thousand years, and then we remember that that hasn't stopped us from progressing our technology, then we come to this really interesting argument that even if we have hit plateau of artificial intelligence, what we can do with it, what we can build on top of it, we haven't even started to tap yet. So next year and beyond, it's going to be about the agents, the tools, the software that we build using artificial intelligence. Right now, most of us are still using the core tools, the underlying technology like ChatGPT, rather than things that are built on top of these language models. Increasingly, software tools and agents are going to be integrated into our daily life. Think things like Siri, but you know, good.