The Sketch.dev team writes about building an AI coding agent, covering agent loops, containers, merge queues, and an LLM-caused outage.
HN commenters mention Sketch as a similar, well-organized coding assistant that needs little babysitting, and the agent loop post is cited by its author.
> The software world is very close to building a super intelligent senior software developer. Companies like this will ask all the best things a software engineer does automatically. Now claude will add it into the coding agents itself.
Except Claude is more expensive than an actual senior software developer. Otherwise, why are many companies terrified of the usage bill that gets printed on the invoice?
The nonsense in "tokenmaxxing" was a complete marketing scam and illusion of cheap tokens which in reality were heavily subsidized.
The entire point is detecting bad code before it reaches production. [0] AI generated or not.
[0] https://sketch.dev/blog/our-first-outage-from-llm-written-code
The thing that's working really well for me is parallel research tasks.
I can pay full attention to the change I'm making right now, while having a couple of coding agents churning in the background answering questions like:
"How can I resolve all of the warnings in this test run?"
Or
"Which files do I need to change when working on issue #325?"
I also really like the "Send out a scout" pattern described in https://sketch.dev/blog/seven-prompting-habits - send an agent to implement a complex feature with no intention of actually using their code - but instead aiming to learn from which files and tests they updated, since that forms a useful early map for the actual work.
I use https://sketch.dev, which seems to be approximately the same kind of thing. I find it is very organized, it does not need nearly as much babysitting as this guy is talking about. Perhaps a difference is the "todo read" and "todo write" tools, which it uses to break the main goal into achievable tasks, and then judge itself on if it has finished correctly or not. Also, Sketch is easy-going with corrections: if you tell it that it went the wrong direction, it quickly adjusts and follows your direction.
Anyway, two big thumbs up for agentic dev workflows from me, but only so far with Sketch.
> In practice, most of what people generate is boring and easy to validate: you know within a couple minutes whether it's the right shape, whether anything sticks out the wrong way, the way a veteran chess player can quickly pattern-match a whole chessboard. When it isn't boring, you read carefully (and expensively), or you just say "no, try again, give me something more boring", or you break your LLM generation into smaller steps that are easier to pattern match and recurse.
nope!
https://sketch.dev/blog/our-first-outage-from-llm-written-code
Hi, author here. I mean the piece of code that calls the model and executes the tool calls. My colleague Philip calls it “9 lines of code”: https://sketch.dev/blog/agent-loop
We have built two of them now, and clearly the state of the art here can be improved. But it is hard to push too much on this while the models keep improving.
Here's a blog post about it: https://sketch.dev/blog/merde
And some tech details: https://sketch.dev/blog/jagged-git-repo
(not mine, I just think it's cool)