🤝We're thrilled to have LoftLabs on board as our official Workshop Sponsor at #CCDPune2025. Let’s build the future of cognitive tech—together 📅 12 July 2025 📍 Tip Top International 🎟️ Registrations are now live 👉 ccd.gdgcloudpune.com #GoogleCloud #gdgcloudpune #ccdpune

We are launching a $20 million open source sustainability fund to support maintainers of critical OSS. Details below;

Please don't tell me the container has to first be compiled to wasm....

I honestly never understood how we could get to a future where computers program without the need of developers. I now understand and it's actually much easier than I would have ever thought.

I just vibe coded a react application. Haha, you guys don't have a job anymore.

I'm too old for cloudflare's cloud. I don't want to learn new ways of doing things. Just do the existing stuff and not make it difficult or expensive.

MCP Needs: 1. Streaming responses 2. "Tasks/Async tools"

LLM driven developments: You pick the tools, frameworks, architectures that LLM suggests because you know it will be writing and maintaining the code.

I'm sorry y'all, but asking o3 almost always give a better answer than asking you. Luckily I'm not here looking for quality responses.

Once we can start running these SOTA LLMs local, it's gonna get weird economically. Maybe they just don't let you because how else can they sell tokens to recouple the training cost.

The Context - Special Episode ft. GitHub & Microsoft x.com/i/broadcasts/1…

Just expose the dumb struct you idiots. cs.opensource.google/go/go/+/refs/t…

TypeScript type system is just madness.

Somebody just host ai-sdk generateText as a REST API and then compete with openrouter.

I've spent the last year building chat UI. That's innovation for you.

Lets be honest, does anyone use a local model to actually accomplish anything? Like I'd like to use one for test or something but they usually can't even do a simple tool call.