Raschka to discuss DeepSeek-V4, GLM-5.2, Kimi K3 and open-weight coding agents
hugobowne · x · 2026-07-28
A livestream preview between Hugging Face’s Huggo Bowne and Sebastian Raschka will cover open source, the latest LLMs, and agent frameworks.
The discussion topics include:
- What DeepSeek-V4, GLM-5.2, Kimi K3, and the upcoming Qwen3.8 are actually doing differently
- What stronger open-weight models mean for local coding agents
- Which parts of current agent systems will survive the next model release
- What Raschka learned from starting with a small Qwen3 model and iterating with evaluation and inference work
More from coding & agent
- LangChain argues companies will rent models but own workflows, evals and memory — hwchase17 · 2026-07-28
- GrokTerm puts Grok voice and MCP into a real multi-tab Rust terminal — Daniel_Farinax · 2026-07-28
- Kimi K3 report says AgentENV uses Firecracker microVMs for agent sandboxes — TheNickWalsh · 2026-07-28
- Dev Perspective: Auth and Infra Remain Biggest Friction in AI Agent Development — vboykis · 2026-07-28
- Every reports almost all of OpenAI now uses Codex internally — every · 2026-07-28
- One user moved a valuable domain to a new registrar entirely with Codex — corbtt · 2026-07-28