61–70 of 88 posts
Context switching between more than 2 threads of work is untenable if you want to really review code in depth. And with AI, you need to go through everything with a fine toothed comb
Has anyone set up a local only autonomous agent, using an open source model from somewhere like huggingface?
Still a bit confused on the details of implementing the technique. Would appreciate any explanations (thanks in advance).
Earlier quoted context omitted.
Your comment is disproportionately rude. Just because your team can't leverage multiple coding agents doesn't mean no one else can. And even if OP also can't, this is a good place to discuss possible problems and solutions for parallel development using coding agents. Please refrain from gatekeeping.
“With this approach, I can manage to have 10–20 pull requests open at once, each handled by a dedicated agent.” A quote from the post. No, I think my post is calibrated quite well considering what OPs post does to our industry.
It can mean, for example, that 2 agents worked for some time in a list of 20 TODO features and produced 20 PRs to be reviewed. They could have worked overnight even.
You're seemingly judging from the least generous interpretation, which is not constructive and is also against HN guidelines fyi.
It still amuses me how literally people took Kapathy's famous tweet around vibe coding https://x.com/karpathy/status/1886192184808149383 If people were to actually read beyond the first sentence, it would become clear very quickly that this was meant to be tongue in cheek.
Earlier quoted context omitted.
“With this approach, I can manage to have 10–20 pull requests open at once, each handled by a dedicated agent.” A quote from the post. No, I think my post is calibrated quite well considering what OPs post does to our industry.
Having 20 PRs open at once doesn't necessarily mean managing 20 agents simultaneously. It can mean, for example, that 2 agents worked for some time in a list of 20 TODO features and produced 20 PRs to be reviewed. They could have worked overnight even. You're seemingly judging from the least generous interpretation, which is not constructive and is also against HN guidelines fyi.
9-10am: I comb through our issue list, figuring out which issues are well defined, which need more input or design decision. => I pick a hand-full, lets say 10 that I kick off to run in the background, and lets say another 10 for further specification.
10-2pm: I thinker with 10 issues to figure out the exact specs and to expand the requirement list.
2pm-6pm: I review the code written by the agents, one by one. I kick off for further work things that need more input, or merge things that look good.
Earlier quoted context omitted.
“With this approach, I can manage to have 10–20 pull requests open at once, each handled by a dedicated agent.” A quote from the post. No, I think my post is calibrated quite well considering what OPs post does to our industry.
Having 20 PRs open at once doesn't necessarily mean managing 20 agents simultaneously. It can mean, for example, that 2 agents worked for some time in a list of 20 TODO features and produced 20 PRs to be reviewed. They could have worked overnight even. You're seemingly judging from the least generous interpretation, which is not constructive and is also against HN guidelines fyi.
Earlier quoted context omitted.
People find it a lot more palatable when the AI requires all this information than when software engineers do though. If I ask for clear requirements I’m asked to just figure it out. But if the AI implements nonsense without clear requirements that the fault of the specs.
Well, that’s because the software engineers are irritating when they push back and say ‘no’ or ‘wtf’. When the AI does it, it’s being polite and stuff. /s, kinda.
/s but not really?
Obviously I'm an AI-tools skeptic, but this is hilarious: > 1. Prepare issues with sufficient context > Start by ensuring each GitHub issue contains enough context for agents to understand what needs to be built and how it integrates with the system. This might include details about feature behavior, file locations, database structure, or specific requirements such as displaying certain fields or handling edge cases.…
lmao, "good writting skills" =)
The sweet spot for me is 2 agents on different projects. Surprisingly the context switch is easy. It's harder when doing 2 tasks on the same project.
Earlier quoted context omitted.
People find it a lot more palatable when the AI requires all this information than when software engineers do though. If I ask for clear requirements I’m asked to just figure it out. But if the AI implements nonsense without clear requirements that the fault of the specs.
I am amazed at how suddenly people are on board with writing clear design documentation now that it means AI can generate the code rather than humans. I wonder how much better humans would be at generating code given the same abundance of clearly-written design documentation?