asdev

7 ms·

91.
by asdev 11mo ago
I wonder what the accuracy is for Claude to always follow a Skill accurately. I've had trouble getting LLMs to follow specific workflows 100% consistently without skipping or missing steps.
92.
by asdev 11mo ago
You can't rely on them directly for business value, but you can rely on them indirectly is the summary of this post
93.
by asdev 11mo ago
Open source control pane for running coding agents in parallel: https://github.com/built-by-as/FleetCode
94.
by asdev 11mo ago
Have you ran your pipeline against an open benchmark like https://github.com/opendatalab/OmniDocBench ?
95.
by asdev 11mo ago
I tried your project in my exploration and ran into an issue but I can't quite remember what it was. Anyways, I would be interested in connecting and learning more about your project, I dropped my X handle in the first issue on my Gith
96.
by asdev 11mo ago
You should check out their MCP server: https://support.atlassian.com/atlassian-rovo-mcp-server/docs...
97.
by asdev 11mo ago
thanks for clarifying. I'll take a look into supporting this
98.
by asdev 11mo ago
sw engineering will be at an even higher premium if you've seen the code AI creates. AI will raise the bar to entry for sure though
99.
by asdev 11mo ago
can you elaborate why you don't want to run multiple copies of your dev stack? is it because you want to run your app from one place but be able to test multiple changes, vs having to install deps and start it from multiple folders?
100.
by asdev 11mo ago
I'll take a look. I think containers are overkill for my type of workflow. I'm working on a project where dependencies don't change often so I'm not worried about a CLI agent installing something on my machine and breaki
101.
by asdev 11mo ago
Not based on a single provider, I need to add more providers but basically supporting a provider is as simple as adding the relevant shell command for the CLI agent(claude for CC, codex for OpenAI etc). I'll try making a video at some
102.
by asdev 11mo ago
oh, and also closed source
103.
by asdev 11mo ago
I tried that, I couldn't get it to authenticate properly since I need to run some commands before running claude code. I also don't like that they rolled an entire UI when the terminal UI is good enough, it's just the paralle
104.
by asdev 11mo ago
yeah I was doing the same and it was working okay, but was hard to work in a truly parallel fashion due to agents making conflicting changes
105.
by asdev 11mo ago
Looks cool! I don't personally have the need to code from anywhere besides my laptop though
106.
by asdev 11mo ago
thanks! all feedback appreciated!
107.
Show HN: FleetCode – Open-source UI for running multiple coding agents(github.com)
103 points by asdev 11mo ago|52 comments
108.
by asdev 11mo ago
so just a way to be able to pull in more context/info about MCP servers in the form of .MD files?
109.
by asdev 11mo ago
The problem is you need to tell prompt Claude to "Store" or "Remember", if you don't it will never call the MCP server. Ideally, Claude would have some mechanism to store memories without any explicit prompting but
110.
by asdev 1y ago
I can prompt an LLM to make a mermaid diagram and iterate on it. Why do I need "ChartDB"?
111.
by asdev 1y ago
I speak for everyone when I say we don't need these videos at all and would be better off without them
112.
by asdev 1y ago
how do claude/openai get around rate limiting/captcha with their computer use functionality?
113.
by asdev 1y ago
this is a great copypasta
114.
by asdev 1y ago
Looking for more self study statistics resources for someone with a CS degree, any other recs?
115.
by asdev 1y ago
I don't think intelligence is increasing. Arbitrary benchmarks don't reflect real world usage. Even with all the context it could possibly have, these models still miss/hallucinate things. Doesn't make them useless, but
116.
by asdev 1y ago
This is what billions in capex and 100M offers gets you apparently
117.
Meta launches 'Vibes,' a short-form video feed of AI slop(techcrunch.com)
13 points by asdev 1y ago|2 comments
118.
by asdev 1y ago
Why they're working on all the application layer stuff is beyond me, they should just be heads down on making the best models
119.
Ask HN: Do you find AI code review tools useful?
1 points by asdev 1y ago|1 comments
120.
by asdev 1y ago
The problem is the research phase will fail because you can't glean tribal product knowledge from just looking at the code

More ›