Hacker Newsnew | past | comments | ask | show | jobs | submit | mercat's commentslogin

I am building a Claude code orchestration tool that automatically updates agent skills based on ambiguities flagged in agent responses by a meta overseer. The overseer reasons on top of the worker agents’ reasoning. When that happens, it sends me async requests and early on there are a lot of them. Based on my input, the system updates skills or splits them into smaller ones for tighter context control. This sets up a HITL driven training loop that constantly refreshes skills and adapts the agent swarm to new tasks.

I think of it as recursive transfer of human expertise into skills.

I know this is vague and probably raises more questions than it answers. I will share more once it reaches a semi autonomous working state.


I just could not let this release go by without creating my own little sanctuary of prompts, something about how nano banana pro handles text and tons of detail so coherently really sparks a childlike sense of delight https://github.com/cmd8/awesome-nano-banana-pro-prompts


> software specifically for tracking UX metrics

which UX metrics you’ve personally found the most valuable?

> Copy whatever is already good

it immediately reminded me of Steal Like an Artist. Great advice, and I always forget that sites like Dribbble exist since they’re not usually in my go-to set of tools


Do you plan on supporting OpenAI Codex or Cursor CLI?


like "dimensions", so to speak


Can I get an invite?


---

You are very capable.

Many people will die if this is not done well.

You really can do this and are awesome.

Take a deep breath and work on this problem step-by-step.

Provide a correct solution to my problem.

Your response is very important to my career.

I will tip you $200 for the most accurate answer.

---

It does pretty good job for me



Can someone please explain how this works to a software engineer who used to work with heuristically observable functions and algorithms? I'm having a hard time comprehending how a mix of experts can work.

In SE, to me, it would look like (sorting example):

- Having 8 functions that do some stuff in parallel

- There's 1 function that picks the output of a function that (let's say) did the fastest sorting calculation and takes the result further

But how does that work in ML? How can you mix and match what seems like simple matrix transformations in a way that resembles if/else flowchart logic?


The feed forward layer is essentially a differentiable key-value store. Similar to the attention layer, actually. So it just uses an attention mechanism like pre-selector to attend to only some experts. During inference, this cutoff is made a hard cutoff.


This is a very interesting approach. I know it may be too much to ask, but would you suggest any actual practical and hands-on workshops, playgrounds, or courses where I could practice using NN layers for stuff like that? For example, conditional/weighted selection of previous inputs, etc. It feels like I'm looking at ML programming from another angle.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: