Humans are almost useless without memory/in-context learning. It’s surprising how much LLMs can do with so little memory.
The important remainder is that LLM-based agents will probably have better memory/online learning as soon as they can handle it, and it will keep getting better, probably rapidly. I review current add-on memory systems in LLM AGI will have memory, and memory changes alignment. A few days after I posted that, OpenAI announced that they had given ChatGPT memory over all its chats, probably with a RAG and summarization system. That isn’t the self-directed memory for agents that I’m really worried about, but it’s exactly the same technical system you’d use for that purpose. Fortunately it doesn’t work that well—yet.
Now it’s obvious that many developers are aware and explicitly focused on the virtues of online learning/memory.
This is a great post because LLMs being a dead-end is a common suspicion/hope among AGI thinkers. It isn’t likely to be true, so it should be discussed. More in a separate comment.
Idk if I’m the only one here, but I use LLMs for coding and I have disabled memory. This wasn’t really an educated move, but after having an AI completely hallucinate in one chat and get the problem or the code at hand totally wrong, I’m afraid its misunderstanding will contaminate its memory and mess up every new chat I start.
At least with no memory I can attempt to rewrite my prompt in a new window and hope for a better outcome. It forces me to repeat myself a lot, but I now have systems in place to briefly summarize what my app does, show the file structure, and explain the problem at hand.
You beat me to it. Thanks for the callout!
Humans are almost useless without memory/in-context learning. It’s surprising how much LLMs can do with so little memory.
The important remainder is that LLM-based agents will probably have better memory/online learning as soon as they can handle it, and it will keep getting better, probably rapidly. I review current add-on memory systems in LLM AGI will have memory, and memory changes alignment. A few days after I posted that, OpenAI announced that they had given ChatGPT memory over all its chats, probably with a RAG and summarization system. That isn’t the self-directed memory for agents that I’m really worried about, but it’s exactly the same technical system you’d use for that purpose. Fortunately it doesn’t work that well—yet.
I wrote about this in Capabilities and alignment of LLM cognitive architectures two years ago, but I wasn’t sure how hard to push the point for fear of catalyzing capabilities work.
Now it’s obvious that many developers are aware and explicitly focused on the virtues of online learning/memory.
This is a great post because LLMs being a dead-end is a common suspicion/hope among AGI thinkers. It isn’t likely to be true, so it should be discussed. More in a separate comment.
Idk if I’m the only one here, but I use LLMs for coding and I have disabled memory. This wasn’t really an educated move, but after having an AI completely hallucinate in one chat and get the problem or the code at hand totally wrong, I’m afraid its misunderstanding will contaminate its memory and mess up every new chat I start.
At least with no memory I can attempt to rewrite my prompt in a new window and hope for a better outcome. It forces me to repeat myself a lot, but I now have systems in place to briefly summarize what my app does, show the file structure, and explain the problem at hand.