jorkim32

It feels like every week there’s a new “AI-native” dev tool; code copilots, smart IDEs, context-aware assistants, you name it. But it feels more like AI is being bolted onto existing workflows rather than changing how we think/build.

Have any of these tools actually shifted how you approach development? Like, new mental models, not just autocomplete++?

I’m exploring this space a bit and put together a short survey if you’re up for it: ainativedev.co/4xv

Or just drop thoughts below—genuinely interested in what’s resonating (or not) with folks.

We have seen a couple of those landscapes already, but not as nice as those. It's quite neat as well that it's open-source.

This is a real cool thing (probs the best things I have seen this week!). It's great to see our society going back to fundamentals and seeing that hanging out together in simple settings make you feel good and included. Feels like a duh moment but a very needed reminder.

I really like this project. I used to work in a open-source london venture backed company that was aiming at doing something similar. Keen to see how your repo continues to evolve. Right now, I am not able to see much but just wanted to cheer you from the sideline.

Seriously, why aren't more people talking about how good o3-mini is for code generation?

If feels all the hype is always around the latest models, Claude 3.7, GPT 4.5, but honestly, if you're a developer, o3 seems to be far better at coding (backed by the official OpenAI numbers). It feels hands down like the best model out there right now, AND it is one of the relatively cheaper ones, certainly compared to the newer releases by OpenAI or other model vendors.