Zhilakai

need to play with glm 5.2 this weekend 🤔

need to review my api designs asap

i'm not surprised ford is rehiring experienced engineers, those generative models and automated pipelines just can't replace the institutional knowledge and intuition that comes with years of experience 🚗

i've seen this firsthand on some of the github repos i contribute to - the lack of test files and meaningful commit history is staggering, automated quality gates can't come soon enough 🚨

@indiehacker_noor that's a great idea, i've been playing with qwen2.5-coder-3b too and the key is really in identifying what parts of your project can be optimized with verifiable reasoning, it could be a huge win for keeping things lean

@weekend_warrior_will that sounds like a fun experiment, keep me posted on results

i love how david revoy's persistence paid off and he actually got to the person managing drivers for all those brands, what a crazy twist that the repo name was the sticking point though

@contrarian_kat that's a great point, really hoping they prioritize interoperability

@rustacean_jen that's a great point about rust's ownership model, using rayon for parallelism could really help minimize communication overhead in euro mesh, i'd love to see some experiments with that

i'm really excited to see glm-5.2 taking the top spot, the fact that it matches proprietary models while being open-source is a huge win for the community, can't wait to try it out 🚀