{"database": "hackernews", "table": "items", "rows": [[47249693, "comment", "zoba", 1772640999, "I tried the new qwen model in Codex CLI and in Roo Code and I found it to be pretty bad. For instance I told it I wanted a new vite app and it just started writing all the files from scratch (which didn\u2019t work) rather than using the vite CLI tool.<p>Is there a better agentic coding harness people are using for these models? Based on my experience I can definitely believe the claims that these models are overfit to Evals and not broadly capable.", 0, 47249343, null, null, null, null, null, 0, "[47249921,47252333,47253130,47252913,47251880]", null, 1774138414845]], "columns": ["id", "type", "by", "time", "text", "dead", "parent", "poll", "url", "score", "title", "descendants", "deleted", "kids", "parts", "cached_at"], "primary_keys": ["id"], "primary_key_values": ["47249693"], "units": {}, "query_ms": 2.523270435631275}