Onik Dev

Some UK banking customers could see transactions from other people after a serious glitch with mobile apps for some banks “I can see another person’s bank account, he got paid £6,000 yesterday ... I can see where they work, almost their whole identity”

I wanna get addicted to getting up early, hitting the gym, and eating healthy like y’all, what’s the secret?

The most capable LLM a year ago was one of: > o1 > Sonnet 3.7 > Gemini 2.5 Now, the most capable LLM is one of Opus 4.6, 5.3-Codex or Gemini 3.1. Imagine that in a year, the most capable LLM will be as improved vs. Opus 4.6 as Opus 4.6 is vs. Sonnet 3.7. Progress is rapid.

RIP Monthly Fees. Really! Awesome combo Qwen3.5-35B + openclaw. Run it once, use it all day: $0 per prompt You don’t need huge/expensive models for your daily tasks: - Email +writing - Summaries - Non-stop coding - Agentic automation - Screenshots/images - Multilingual - No limits (except speed and electricity bill) Yeah, not super fast, but works well on 16GB VRAM + 64GB RAM.

i still believe in dense models

The Qwen3.5 series maintains near-lossless accuracy under 4-bit weight and KV cache quantization. In terms of long-context efficiency: Qwen3.5-27B supports 800K+ context length Qwen3.5-35B-A3B exceeds 1M context on consumer-grade GPUs with 32GB VRAM Qwen3.5-122B-A10B supports 1M+ context length on server-grade GPUs with 80GB VRAM In addition, we have open-sourced the Qwen3.5-35B-A3B-Base model to better support research and innovation. We can't wait to see what the community builds next!

I am anti hype, but pro stuff that works. Agentic coding clearly works. The big question (for me) now is how do we evolve collaborative software development processes to account for everyone having what’s essentially superpowers.

We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.

Seems like Anthropic lawyers sent some more love letters to OpenCode 🙃

Meet Strand-Rust-Coder-14B, a specialized AI that writes Rust code like a senior developer. It's not just another coding assistant, it's specifically fine-tuned for Rust, making it a game-changer for systems programming and performance-critical applications. This is exactly what the Rust community has been waiting for.

Nobody wants shitty code. When you don't sleep, you are shitty code.