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

Apache Burr: Build reliable AI agents and applications

burr.apache.org

Post reply on HN
Print view View original

81–90 of 125 posts

Re: Apache Burr: Build reliable AI agents and applications

#81

Post by fnordpiglet »

thedougd wrote: post #27
hmokiguess wrote: post #5

How does this compare to https://strandsagents.com/ ? I'm interested in tools in this space, right now I'm not attached to one, but Bedrock + Serverless on Agent Core feels like the "easy guided path" though I don't like the platform lock-in

Curious about other experiences. I’ve been playing with this stack and left wondering if Strands provides any secret sauce with Agent Core. So far it doesn’t feel that way and sometimes they even feel at odds with each other.

I think strands and agentcore don’t have a ton of overlap but I have found agentcore makes running strands frameworks pretty easy and relatively inexpensive. I like strands in that it’s the most mature (IMO) that isn’t specific to a model provider - although I found using things like vertex to require a lot of custom work to get caching and other things to work properly, the maturity is all on bedrock.

I’ll need to dig into burr - I’m not finding strands has an insurmountable maturity to it but it’s not carrying a lot of weird opinion (like some of the other OSS frameworks seem to be highly infected with) and is pretty practical in what it exposed and does, and is most like the agentic frameworks I’ve worked with inside FAANG. If burr can meet that and keep growing I’d probably look at moving to it as I also get a sense strands has a bit of the Amazon “highly probable to be abandoned once the managers and PMs get promoted” feeling.

Re: Apache Burr: Build reliable AI agents and applications

#82

Post by andai »

sdesol wrote: post #79
andai wrote: post #75

Earlier quoted context omitted.

(I'm not the guy but) That's funny, I had the same idea the other day. Keeping summaries of files. Haven't tested that yet. Another thing I've been thinking is how, most parts of a file are not relevant to the whole system. Like there are parts where they intersect, and those seem to be the most important ones for capturing the big picture. You wanna be able to see the entire "skeleton". So I thought the summary mayb…

If you include the following: https://github.com/gitsense/chat/blob/main/base-state/analyz... In your chat with AI, include the above file and let it know what your requirements are and I can create the analyzer and include it. You can also think of my tool as data prepping tool. So if you have a clear prompt the AI can review the file during analysis and remove all unnecessary code so the extracted metadata will the…

> If a developer wanted to change X, would these keywords help them find this file?

I think the best way to generate these is with a sub-agent. Tell it to try and solve a problem that involves editing this file, and see what it starts grepping for.

This ties in with this idea that the tools and designs should be what comes naturally to the LLM, i.e. what it's already been trained on. And the most straightforward way to do that is to let it reach for it.

Like when you reach in the darkness for an object. Where your hand lands is exactly where it should be.

Re: Apache Burr: Build reliable AI agents and applications

#83

Post by andai »

dd8601fn wrote: post #77
brotchie wrote: post #6

I'm still on the fence about agent frameworks, they have their place, and it depends on the nature of the agent: e.g. "Low latency, return a good enough response in 3 seconds, vs. working for 3 hours on a problem." BUT, if you boil it down, an agent really is context building, making an LLM call, executing requested tool calls, parsing the final model output, returning it to some frontend. There's extensions like mem…

I’ve said something similar about dozens of frontend frameworks. It’s massive abstraction and convolution for some future payoff that’s obviously never going to happen. But sometimes people just need something to do, or something fun to play with, and “the next guy” rarely matters that much… so who cares that you’ve saddled them with the result of your paid playtime?

So before AI I had the experience, more often than not, that it would take me longer to figure out how to use someone else's thing (or get it to do some particular thing, which often turned out to be impossible), than to just make my own.

And that was before I could just ask the computer to make it for me!

But most people seem to be the other way around. They'd rather deal with abstractions and boilerplate instead of writing the actual code.

Re: Apache Burr: Build reliable AI agents and applications

#84

Post by bananamogul »

elijahbenizzy wrote: post #39

One of the co-creators/maintainers here! Will try to answer Qs over the day.

Do you ever abbreviate the project name as A. Burr and how many Hamilton jokes[1] result from this?

[1] https://en.wikipedia.org/wiki/Burr%E2%80%93Hamilton_duel

Re: Apache Burr: Build reliable AI agents and applications

#85

Post by nico »

sgc wrote: post #76
nico wrote: post #43

On a tangent, can anyone recommend good coding agent orchestration tools or platform? Something to launch, manage and monitor codex or claude agents in multiple machines Ideally self-hostable/open source I know claude code has a lot of that internally built in already, but it’s claude-only

I was looking at their docs and Burr has agent cookbooks to get started with this, and it can handle multi-machine workflows. Is this not what you were looking for? I am not sure how it integrates and uses skills etc, but it seems like it should work to me. https://burr.apache.org/docs/examples/agents/

Thank you. I want something like that, but that uses codex/claude code as agents (through their corresponding subscriptions), instead of having to create ad-hoc agents + api keys

Re: Apache Burr: Build reliable AI agents and applications

#87

Post by pjmlp »

kristjansson wrote: post #25
brotchie wrote: post #6

I'm still on the fence about agent frameworks, they have their place, and it depends on the nature of the agent: e.g. "Low latency, return a good enough response in 3 seconds, vs. working for 3 hours on a problem." BUT, if you boil it down, an agent really is context building, making an LLM call, executing requested tool calls, parsing the final model output, returning it to some frontend. There's extensions like mem…

Obscuring core logic is the most egregious part of most agent frameworks. One needs a clear view of what, exactly, is being sent to the underlying language model, and what's coming back. Everything in an 'agentic' application is realized as a sequence of tokens or a call to a provider eventually. It should be clear and obvious from ~all layers of the app what that's going to look like.

Unfortunately agent orchestration frameworks feels like the second coming of BEPL, and the incentives are all wrong.

Re: Apache Burr: Build reliable AI agents and applications

#89

Post by ivanmontillam »

redlewel wrote: post #88

Didn't bother reading more after seeing the vibeslopped landing page that took Also "700+ Discord Members" is not any type of endorsement of a technology or service.

Not to kick dirt on their faces, but I also think Discord is just not the proper chat community open source projects should have.

Might be my IRC mind talking, but a proprietary service like Discord, or Slack, or even Telegram at times, is just not suitable for the target community, as there's often a data privacy concern.

I never had a Discord account, and I'm hoping I can keep that streak.

Re: Apache Burr: Build reliable AI agents and applications

#90

Post by msradam »

I have enjoyed using this framework in my personal and work projects, having a reliable stateful workflow for AI models while getting free observability. I stitched a tool that allows mounting a Burr state machine as an MCP, giving agents a rail to follow, and no matter how complex the state machine gets the MCP tools are constrained to state machine navigation: https://github.com/msradam/theodosia

I am currently working on skills-to-state-machine conversions, since a lot of popular skills out there are already written as phases for an AI model to follow, so it would be great to leverage the explicit functionality of Burr to make that more reliable. Thank you for this amazing project.

Post reply on HN