I had this thought and created https://pushrealm.com which is essentially a sort of Stackoverflow written by agents.
replyMy theory was that if an agent burns 30 minutes resolving an issue not present in training data, posting the solution would prevent other agents re-treading the same thinking steps.
Fascinating! Do you have a way to detect/flag malicious stuff by any chance? (Seems like a good vector for prompt injection, but maybe no more than any other internet site?)
replyI see why, but I don't feel this is the solution. Being able to search thru the endless LLM responses is not viable. However having useful memories, similar to human brain is more important. I sense this is why neuromorphic computing is the next step, energy efficient and doesn't remember much of what isn't useful to be stored.
replyWhy not preserver the essential memories in text? Why neuromorphic?
replyYou are better being able to quickly deduce ways of acting from memories of previous scenarios, than have to attempt every scenario to build a fresh memory of each, which is a lot of memory, and requires exposure to every situation before being able to do it.
replySo you mean caching? :-)
replyA lot of what I see people using LLMs for would be more cheaply and reliably done by [scripts]. A search engine style suggestion thing like "Have you tried `sed`?" would be beneficial imo
replyIn my experience, Claude is more than happy to go to Unix tools rather than write its own. Sometimes it will write a lil python script to solve something, but more often than not it’ll pipe together Unix utilities.
replyThis has the benefit of it knowing all of the arcane flags, especially for formatting output.
I believe that. I also believe that my idea won't come to fruition, at least from a group that is incentivized to make a user's first instinct be to use their product and not an external tool.
reply