@webdevcody Yes. To me models follow the goals you give it, their not really creative or know your full vision unless you spell it out completely for them ahead of time if the problem is complex then this makes everything harder to verify overtime when your not writing the code yourself
convex-typegen v0.3 by @codingwithjamal
Rust type generation for Convex just got a major update - unique args struct names, fallible args conversion, Convex-correct optional fields, and a lighter codegen-only path with default-features = false.
For Rust developers who want better DX with Convex.
https://t.co/aoZVXpVlMl
@_sarant@AnthropicAI Their research focus is on alignment and language models — image generation is a different model architecture (diffusion models, etc.) There’s no obvious mission-driven reason for Anthropic to compete with Midjourney, DALL-E, or Sora — it’s not core to their goals
Building an AI agent for @ClarityTextCom and looking for papers/resources on RAG and context retrieval — specifically hybrid search (FTS + vector) and how to rank/filter retrieved context effectively. How do you decide what’s relevant enough to include in the context window?
@Abhishekcur i really want to like c++, theres a beauty in the madness but the toolchain, legacy DX, and the years of using better keep me away from it for any big projects :(