aray07

6 ms·

61.
The four levels of AI-driven engineering(klaviyo.tech)
1 points by aray07 6mo ago|0 comments
62.
Some Simple Economics of AGI(arxiv.org)
1 points by aray07 7mo ago|0 comments
63.
Don't Use Claude's 1M Context Window for Long Coding Sessions(claudecodecamp.com)
2 points by aray07 7mo ago|1 comments
64.
by aray07 7mo ago
The dropping requirements problem is real. What's helped us is breaking the spec into numbered ACs and having the verification run per-criterion. If AC-3 fails you know exactly what got dropped.
65.
by aray07 7mo ago
I do it per feature, not per step. Write the AC for the whole feature upfront, then the agent builds against it. I haven't added a spec-validation step before coding but that's a good idea. Catching ambiguity in the spec before th
66.
by aray07 7mo ago
Agreed. The spec file is context. Writing acceptance criteria before you prompt provides the context the agent needs to not go off in the wrong direction. Human leverage just moved up and the plan/spec is the most important step. Paral
67.
by aray07 7mo ago
This is great. The tests in this case are the spec. When you give the agent something concrete to fail against, it knows what done looks like. The problem is if you skip that step and ask Claude to write the tests after.
68.
by aray07 7mo ago
i think the friction has moved upstream - now it's working on the right thing and specifying what correct looks like. i don't think we are going back to a world where we will write code by hand again.
69.
by aray07 7mo ago
yup, agree - i spend most of my time reviewing the spec. The highest leverage time is now deciding what to work on and then working on the spec. I ended up building the verify skill ( https://github.com/opslane/verify )
70.
by aray07 7mo ago
Test theatre is exactly the right framing. The tests are syntactically correct, they run, they pass but do they actually prove anything?
71.
by aray07 7mo ago
yeah honestly thats what i am struggling with too and I dont have a a good solution. However, I do think we are going to see more of this - so it will be interesting to see how we are going to handle this. i think we will need some kind of
72.
by aray07 7mo ago
thats a great idea - i have been using codex to do my code reviews since i have it to give better critique on code written by claude but havent tried it with testing yet!
73.
by aray07 7mo ago
yeah i have started using codex to do my code reviews and it helps to have “a different llm” - i think one of my challenges has been that unit tests are good but not always comprehensive. you still need functional tests to verify the spec i
74.
by aray07 7mo ago
Yeah just wanted to see what the bottlenecks would be as I started pushing the limits. Eventually made this into a verification skill(github.com/opslane/verify)
75.
by aray07 7mo ago
yeah i have been hearing a lot more about this concept of “digital twins” - where you have high fidelity versions of external services to run tests against. You can ask the API docs of these external services and give it to Claude. Wonder i
76.
by aray07 7mo ago
yup agree - i think have specs and then do verifications against the spec. I have heard that this is how a lot of consulting firms work - you have acceptance criterias and thats how work is validated.
77.
by aray07 7mo ago
yeah i agree - this is somewhat the approach I have been using more of. Write the tests first based on specs and then write code to make the tests pass. This works well for cases where unit tests are sufficient.
78.
Agents that run while I sleep(claudecodecamp.com)
429 points by aray07 7mo ago|489 comments
79.
How prompt caching works in Claude Code: experiments and architectural lessons(claudecodecamp.com)
1 points by aray07 7mo ago|0 comments
80.
Why Claude Code is just a while loop (with 20 tools)(claudecodecamp.com)
1 points by aray07 7mo ago|1 comments
81.
Show HN: Claude.md templates based on Anthropic's advice(github.com)
1 points by aray07 7mo ago|0 comments
82.
How Claude Code agent teams work under the hood(claudecodecamp.com)
1 points by aray07 7mo ago|0 comments
83.
Show HN: Claude.md templates based on Boris Cherny's advice(github.com)
10 points by aray07 8mo ago|0 comments
84.
I don't read code anymore - creator of Superpowers(claudecodecamp.com)
1 points by aray07 8mo ago|0 comments
85.
Show HN: I built an open-source CSV importer that I wish existed(github.com)
2 points by aray07 1y ago|0 comments
86.
CSVs Won't Die: The Cockroach of Data Formats(importcsv.com)
4 points by aray07 1y ago|0 comments
87.
What I learned vibe coding a WASM CSV Parser(importcsv.com)
5 points by aray07 1y ago|0 comments
88.
by aray07 1y ago
That's good feedback - working on it!
89.
Show HN: I built an open-source CSV importer that I wish existed(github.com)
5 points by aray07 1y ago|3 comments
90.
by aray07 1y ago
At my previous startup, CSV import was make-or-break for customer onboarding. We built the first version in three days. Then reality hit: Windows-1252 encoding, European date formats, embedded newlines, phone numbers in five different forma

More ›