zlacker

[parent] [thread] 4 comments
1. branaf+(OP)[view] [source] 2026-01-26 23:24:40
You're describing a multi-agent long horizon workflow that can be accomplished with any programming language we have today.
replies(1): >>999900+k6
2. 999900+k6[view] [source] 2026-01-27 00:01:39
>>branaf+(OP)
I'm always open to learning, are there any example projects doing this ?
replies(2): >>fwip+db >>branaf+cc
◧◩
3. fwip+db[view] [source] [discussion] 2026-01-27 00:36:22
>>999900+k6
yes "now what?" | llm-of-choice
◧◩
4. branaf+cc[view] [source] [discussion] 2026-01-27 00:43:56
>>999900+k6
The most accessible way to start experimenting would be the Ralph loop: https://github.com/anthropics/claude-code/tree/main/plugins/...

You could also work backwards from this paper: https://arxiv.org/abs/2512.18470

replies(1): >>999900+6e
◧◩◪
5. 999900+6e[view] [source] [discussion] 2026-01-27 00:58:08
>>branaf+cc
Ok.

I'm imagining something like.

"Hi Ralph, I've already coded a function called GetWeather in JS, it returns weather data in JSON can you build a UI around it. Adjust the UI overtime"

At runtime modify the application with improvements, say all of a sudden we're getting air quality data in the JSON tool, the Ralph loop will notice, and update the application.

The Arxiv paper is cool, but I don't think I can realistically build this solo. It's more of a project for a full team.

[go to top]