wkyleg

The main issue I experience with LLMs, and the one that seems to most inhibit my further adoption is lack of ability of agents to remember relevant context.

A few years ago everyone was using RAG, embeddings, databases on top of models. Now models with access to local markdown and memory files (like OpenClaw) seem to be readily outperforming these databases with grep and simple UNIX tools.

Is this an inherent issue in scaling LLMS? Does Obsidian work that much better for most people? It anyone finding anything that actually outperforms markdown?

At this point the main bottleneck in my adoption seems to be memory and persistent long term context, not quality or reliability of the models.

I'm curious if there are any technical or scaling metrics we could use to forecast where this will end up going.

Right now, JavaScript scales well with a single-threaded event loop. Certainly not as fast as something like Go for async tasks, but enough to power much of the web and be easy to write.

Why hasn't anyone abstracted the event loop model to scale across multiple machines or utilize modern processors? Perhaps with something more like an Actor model or Erlang's BEAM?

It seems like just getting the JavaScript concurrency model as an abstraction over multicore or multi-machine concurrency would be one of the easiest ways to achieve this. I realize that this is still technically difficult, but programming tends towards "just porting things to JavaScript." I would love to have something like Phoenix framework, just built with JavaScript/TypeScript, and I can scale a back end by bumping size of a machine or scaling horizontally.

What are DIY projects that can use technical ability and/or time to obtain something that typically costs much more as a paid product?

Building a custom PC comes to mind (although this has changed a bit) as does buying on furniture and refinishing.