jwpapi

10 ms·

211.
by jwpapi 7mo ago
quite a long time..
212.
by jwpapi 8mo ago
How do you connect verified data from the past ?
213.
by jwpapi 8mo ago
man that’s what I’m trying to build the whole time, but I keep getting json parsing errors. I’ve debugged a lot, but it seems their haiku is not consistent with the actual output. I want a hook that tells them at the end make sure you’ve bu
214.
by jwpapi 8mo ago
This goes well along with all my non-tech and even tech co-workers. Honestly the value generation leverage I have now is 10x or more then it was before compared to other people. HN is a echo chamber of a very small sub group. The majority o
215.
by jwpapi 8mo ago
Thats actually super interesting and why I really don’t like the whole .md folder structures or even any CLAUDE.md. It just seems most of the time you really just want to give it what it needs for best results. The headline is really bullsh
216.
AI Rot Management(jw.hn)
1 points by jwpapi 8mo ago|0 comments
217.
by jwpapi 8mo ago
I mean that just the way it tackles task in the core is generated differently, like inner harness, through system prompt or deeper root. F.e. Instead of answering instantly it goes through a pre-defined steps which strategy should be done,
218.
by jwpapi 8mo ago
Honestly Im not even sure how much model improvement was in the last 12 months, or it was mainly harness improvement. It feels to me like I could’ve done the same stuff with 4, if I would be able to split every task into multiple subtasks w
219.
AI-Forced ESLint Extension(jw.hn)
2 points by jwpapi 8mo ago|0 comments
220.
The Single Best One-Shot I've Ever Seen(jw.hn)
1 points by jwpapi 8mo ago|0 comments
221.
by jwpapi 8mo ago
Arguably I would think that the last year was mainly inner harness improvement instead model improvement but I could be wrong, just feels like that to me
222.
by jwpapi 8mo ago
Great article and tbh I thought it would’ve been implemented that way makes sense to hash and save mainly context I don’t expect them to care about token usage How about Kimi tho how can I play with it?
223.
by jwpapi 8mo ago
Honestly Tmux, vim, kitty, almost every terminal, shell, script is configurable. It’s what we’re used to. I wouldn’t know why you wouldn’t start allowing more config options.
224.
by jwpapi 8mo ago
Thank you I think this is also how the future of AI Implementation will be. Let’s say legal processes, One doesn’t need to finetune the outcome, but rather the verification layers. It seems related, but my experience so far is that guiding
225.
If you don't engineer backpressure, you'll get slopped(jw.hn)
2 points by jwpapi 8mo ago|2 comments
226.
by jwpapi 8mo ago
I get exactly what you are saying, but honestly man freedom is the best. I’m so happy every day as long as I have my computer an can work on taking over the world. Every day feels like school holidays to me
227.
by jwpapi 8mo ago
I’m team Geoffrey Huntley
228.
by jwpapi 8mo ago
it was just ai written thats why.. unexpectedly so from greynoise.
229.
by jwpapi 8mo ago
The way I see them acting it seems frankly to me that ruthlessness is required to achieve the goals especially with Opus. They repeatedly copy share env vars etc
230.
by jwpapi 8mo ago
Yeah I’ve talked about verification layers here https://jw.hn/dark-software-fabric . I think that is the most efficient way of approaching engineering today.
231.
by jwpapi 8mo ago
I’ve written about that a bit here https://jw.hn/dark-software-fabric AI gets a lot of big projects right if you give at all the tools to verify its own implementation if you can build a proper system to verify the solution
232.
by jwpapi 8mo ago
yeah a lot of people are just coping. If someone wants to become better or more productive, invest in engineering guardrails and verifications validation layers. There are thousand of examples where tech became obsolute and frankly it’s giv
233.
by jwpapi 8mo ago
The first rule is an antipattern. I think describing your architecture or ANY kind of documentation for your AI is an anti-pattern and blows the context window leading to worse results, and actual more deviation. The controlling systems are
234.
by jwpapi 8mo ago
I mean what you get for Claude Code Max is insane its 30x on the token price. If you don’t spend that all it’s your own fault. That must be below elecricity cost
235.
by jwpapi 8mo ago
This is my favorite article this year. Just very insightful and honest. The learnings are worth thousands for me.
236.
by jwpapi 8mo ago
I really wonder why nobody is talking about how it is more important to be able to test the code. 9/10 my ai generated code is bad before my verification layers 9/10 its good after. Claude fights through your rules. And if you cod
237.
Dark Software Engineering(julianmwagner.com)
1 points by jwpapi 8mo ago|0 comments
238.
by jwpapi 8mo ago
Honestly I’m not sure we’re not there yet, run this prompt as a ralph loop for 2 days on your codebase and see where you at...
239.
Ralph Disruption(julianmwagner.com)
1 points by jwpapi 8mo ago|0 comments
240.
by jwpapi 9mo ago
I think it seems to crystalize out a skill set of prompt engineering, debugging, system design (especially safeguard) design and potentially testing to be the most efficient for productivity. It depends obviously what kind of coding you do.

More ›