> I see it as a positive that AI will push folks to think about verification (you cannot truly trust the AI, after all), observability (verifying that the thing actually works in production), architecture (we’ll generate more code, faster, and it’ll need to be organized) and constraints (when shipping faster, you hit constraints faster around resource usage, performance, reliability, etc.)
I resonate with this. Especially when working with a local LLM that isn't as sharp as frontier models, verification is essential. I try to enforce a strict project structure with clean docs that avoid keeping old "2 weeks ago we fixed X by doing Y" type comments. Creating tools and processes to ensure each step is correct before moving on, which includes tests for every major functionality. These help the agent stay on track and within the guardrails.
> I see it as a positive that AI will push folks to think about verification (you cannot truly trust the AI, after all), observability (verifying that the thing actually works in production), architecture (we’ll generate more code, faster, and it’ll need to be organized) and constraints (when shipping faster, you hit constraints faster around resource usage, performance, reliability, etc.)
I resonate with this. Especially when working with a local LLM that isn't as sharp as frontier models, verification is essential. I try to enforce a strict project structure with clean docs that avoid keeping old "2 weeks ago we fixed X by doing Y" type comments. Creating tools and processes to ensure each step is correct before moving on, which includes tests for every major functionality. These help the agent stay on track and within the guardrails.
10 month old story and getting out of date, at the time (4+1+4 points, 0+2+4 comments) https://news.ycombinator.com/item?id=46522485 https://news.ycombinator.com/item?id=46525532 https://news.ycombinator.com/item?id=46638167