Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Multi-Agentic Software Development Is a Distributed Systems Problem

kirancodes.me

Post reply on HN
Print view View original

61–70 of 75 posts

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#61

Post by socketcluster »

rdevilla wrote: post #30
socketcluster wrote: post #29

I tried my hand at coding with multiple agents at the same time recently. I had to add related logic to 4 different repos. Basically an action would traverse all of them, one by one, carrying some data. I decided to implement the change in all of them at the same time with 4 Claude Code instances and it worked the first time. It's crazy how good coding agents have become. Sometimes I barely even need to read the code…

Please stop. I can't take HN seriously any more with comments like this here.

Hmmm. Have you used Claude Code for coding? I'm not saying it's always accurate but for a lot of coding tasks, it's insanely accurate. It's like mind reading.

Like for complex bugs in messy projects, it can get stuck and waste thousands of tokens but if your code is clean and you're just building out features. It's basically bug free, first shot. The bugs are more like missing edge cases but it can fix those quickly.

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#62

Post by rdevilla »

socketcluster wrote: post #61
rdevilla wrote: post #30

Earlier quoted context omitted.

Please stop. I can't take HN seriously any more with comments like this here.

Hmmm. Have you used Claude Code for coding? I'm not saying it's always accurate but for a lot of coding tasks, it's insanely accurate. It's like mind reading. Like for complex bugs in messy projects, it can get stuck and waste thousands of tokens but if your code is clean and you're just building out features. It's basically bug free, first shot. The bugs are more like missing edge cases but it can fix those quickly.

This site is just inundated with LLM slop now. Moltbook everywhere, all day, every day.

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#63

Post by bmitc »

mrothroc wrote: post #31

I've been running a multi-agent software development pipeline for a while now and I've reached the same conclusion: it's a distributed systems problem. My approach has been more pragmatic than theoretical: I break work into sequential stages (plan, design, code) with verification gates. Each gate has deterministic checks (compile, lint, etc) and an agentic reviewer for qualitative assessment. Collectively, this looks…

How did you launch, orchestrate, and run the agents? Did you build your own framework, or are you just using the various CLI tooling?

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#64

Post by SpicyLemonZest »

jbergqvist wrote: post #3

Doesn't this whole argument fall apart if we consider iteration over time? Sure, the initial implementation might be uncoordinated, but once the subagents have implemented it, what stops the main agent from reviewing the code and sorting out any inconsistencies, ultimately arriving at a solution faster than it could if it wrote it by itself?

Subagents working on shared state are primarily a context window hack. They're powerful to the extent that they enable solving problems an agent with global state couldn't solve due to context pollution. I'm sure there are caveats, but to first approximation, a main agent that can comprehend the entire code in enough detail to sort out those inconsistencies could have just written the code itself.

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#65

Post by EagnaIonat »

kang wrote: post #57
EagnaIonat wrote: post #51

Earlier quoted context omitted.

It scales fine if done correctly. Even with the weights the extra context allows it to move to the correct space. Much the same as humans there are terms that are meaningless without knowing the context.

Would it be possible to make GPT3 from GPT2 just by prompting? It doesn't work/scale

Bit of a straw-man there.

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#66

Post by don_esteban »

airstrike wrote: post #42
21asdffdsa12 wrote: post #32

Earlier quoted context omitted.

So one hub architect agent for overview- which generates tokens for the spoke agent and receives architectural problem reports from spoke agents?

That seems to be the most common "fix" people deploy today, but it's still structurally flawed due to inherent limitations in LLMs.

A hybrid approach is needed, precise/algorithmic manager, with LLM workers.

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#67

Post by 21asdffdsa12 »

airstrike wrote: post #42
21asdffdsa12 wrote: post #32

Earlier quoted context omitted.

So one hub architect agent for overview- which generates tokens for the spoke agent and receives architectural problem reports from spoke agents?

That seems to be the most common "fix" people deploy today, but it's still structurally flawed due to inherent limitations in LLMs.

Everything is a compromise- everything is "structurally" flawed. Who is to say- that the brain itself is not something similar, a labyrinthine city, with domain specific shortlived LLMs "excitements" that wake up other short live LLMs? Where not resolving an excitement builds up a frustration counter until a new agent is born whos sole purpose is to solve this problem?

Re: Multi-Agentic Software Development Is a Distributed Systems Problem

#70

Post by mrothroc »

bmitc wrote: post #63
mrothroc wrote: post #31

I've been running a multi-agent software development pipeline for a while now and I've reached the same conclusion: it's a distributed systems problem. My approach has been more pragmatic than theoretical: I break work into sequential stages (plan, design, code) with verification gates. Each gate has deterministic checks (compile, lint, etc) and an agentic reviewer for qualitative assessment. Collectively, this looks…

How did you launch, orchestrate, and run the agents? Did you build your own framework, or are you just using the various CLI tooling?

I created my own framework. Long ago it started as shell scripts that I used in conjunction with aider. It was a very manual process.

It's grown over time to be a full MCP and CLI with stages and gates defined in YAML. I was thinking about open sourcing it but since the code grew organically I would need to do extensive cleanup to make it presentable.

But I do walk through the process on page 9: https://michael.roth.rocks/research/trust-topology/#9

Post reply on HN