0xblacklight

6 ms·

1.
by 0xblacklight 2mo ago
reviewing code as it's being written & re-steering it > reviewing it once 20k lines have been written
2.
by 0xblacklight 2mo ago
this is exactly right
3.
by 0xblacklight 2mo ago
why?
4.
Context Forking to Save Time, Tokens and Trouble(humanlayer.com)
3 points by 0xblacklight 4mo ago|0 comments
5.
Harness Engineering for Coding Agents(humanlayer.dev)
2 points by 0xblacklight 6mo ago|0 comments
6.
by 0xblacklight 8mo ago
https://blacklight.sh
7.
by 0xblacklight 8mo ago
leaving the codebase cleaner than you found it IS creating maximum value to the company because large changes are almost never walking on the knife's edge between "making the codebase better" and "making the codebase wor
8.
by 0xblacklight 8mo ago
every PR should leave the codebase cleaner than it found it
9.
Learn Cloud Architecture Through a Tower Defense Game(github.com)
2 points by 0xblacklight 10mo ago|1 comments
10.
by 0xblacklight 10mo ago
This is an excellent point - LLMs are autoregressive next-token predictors, and output token quality is a function of input token quality Consider that if the only code you get out of the autoregressive token prediction machine is slop, tha
11.
by 0xblacklight 10mo ago
Same!
12.
by 0xblacklight 10mo ago
If you find it works for you, then that’s great! This post is mostly from our learnings from getting it to solve hard problems in complex brownfield codebases where auto generation is almost never sufficient.
13.
by 0xblacklight 10mo ago
I looked when I wrote the post but the paper hasn’t been revisited with newer models :/
14.
by 0xblacklight 10mo ago
It might support AGENTS.md, you could check the site and see if it’s there
15.
by 0xblacklight 10mo ago
Yep it is opinionated for how to get coding agents to solve hard problems in complex brownfield codebases which is what we are focused on at humanlayer :)
16.
by 0xblacklight 10mo ago
I think you’re missing that CLAUDE.md is deterministically injected into the model’s context window This means that instead of behaving like a file the LLM reads, it effectively lets you customize the model’s prompt I also didn’t write that
17.
by 0xblacklight 10mo ago
I imagine it’s highly-correlated to parameter count, but the research is a few months old and frontier model architecture is pretty opaque so hard to draw too too many conclusions about newer models that aren’t in the study besides what I w
18.
by 0xblacklight 10mo ago
Hi, post author here :) Yes README.md should still be written for humans and isn’t going away anytime soon. CLAUDE.md is a convention used by claude code, and AGENTS.md is used by other coding agents. Both are intended to be supplemental to
19.
by 0xblacklight 10mo ago
Hi, post author here We used cloudflare’s AI gateway which is pretty simple. Set one up, get the proxy URL and set it through the env var, very plug-and-play
20.
by 0xblacklight 10mo ago
Hi, post author here :) I didn’t dive into that because in a lot of cases it’s not necessary and I wanted to keep the post short, but for large monorepos it’s a good idea
21.
by 0xblacklight 1y ago
tbh I think the thing that's making this new approach so hard to adopt for many people is the word "vibecoding" Like yes vibecoding in the lovable-esque "give me an app that does XYZ" manner is obviously ridiculous
22.
by 0xblacklight 1y ago
In many cases (though not all) it's the same thing that makes for great engineering managers: smart generalists with a lot of depth in maybe a couple of things (so they have an appreciation for depth and complexity) but a lot of breadt
23.
by 0xblacklight 1y ago
Among other things; coding agents that can get feedback by running a compile step on top of the linter will tend to produce better output. Also, strongly-typed languages tend to catch more issues through the language server which the agent
24.
by 0xblacklight 1y ago
if nuclear power is so much better than coal, why do we need to learn how to safely operate a reactor just to make it work? Coal is so much easier
25.
by 0xblacklight 1y ago
how does this solve it?
26.
by 0xblacklight 1y ago
> Context has never been the bottleneck for me. AI just stops working when I reach certain things that AI doesn't know how to do. It's context all the way down. That just means you need to find and give it the context to enable
27.
by 0xblacklight 1y ago
Thanks! This isn't the first time we've gotten this question. We are about to release under a source-available license with a paid self-hosting license, and a commitment to open-source under MIT if we sunset the project or the co
28.
by 0xblacklight 1y ago
This is a project we built internally for ourselves at Naptha.ai and wanted to release to the public, it's not really part of our core business. We're about to release it under a source-available license with provisions for a purc
29.
We built a dev-friendly way to ship OAuth integrations in minutes for just $1(onedollaroauth.com)
9 points by 0xblacklight 1y ago|6 comments
30.
by 0xblacklight 1y ago
OAuth is a pain. Anyone who’s built integrations with Slack, GitHub, or Discord knows how tedious and error-prone setting up auth flows, refreshing tokens, and securely storing credentials can be. We built a platform that handles all of thi

More ›