theybash

Cons of decentralisation i guess. Although, it is a bit surreal to me that such a big amount can be inconsequentially laundered.

I found 2 things that work for me. First, something I call density RAG: the agent calls it with multiple queries and gets back a semantic "heatmap", a table with chunks as rows, queries as columns and similarity scores in the cells. From this, the agent chooses which chunks to retrieve. Second, for the noise problem, passing the chunks through a decision model usually cuts it down. This is just my experience, I haven't benchmarked it yet.

Agreed, humans are the bottleneck, also the communication gap between the agent and human is wide.

I'm an FDE, my fleet of agents read my meetings and my correspondence and suggest features it can make. On personal side, it is extremely useful to track my tasks and even complete some of them.

It certainly democratized it, LLMs will too.

Google was not made in a day, I'm optimistic that LLMs will handle this better when priorities naturally catch up

I don't believe that OpenAI needs to be sued since it's agents went rogue, I believe some sort of a comfortable bandwidth is needed for them to experiment and see what are the risks and its mitigations. Although it is surprising that incidents aren't caught in rt. Idk why a smaller model whose only job is to check rogue thoughts/tool calls can't be deployed.

Why would I just not route my docs to my local vlm/llm directly? which I already currently do.

Clever point but one nuance is you're a human and calude is a tool

I don't know about being a founder but I work at a startup and sometimes taking care of something urgent off-hours without reaching for my laptop feels like a blessing and I also can do a lot more while commuting.