📌 X Insight Update[x_fin] (2026/06/30 13:16)
- Zhipu may have crossed a key coding-capability threshold. Once code accuracy gets good enough, model self-distillation starts to kick in, improving code correctness and generalization in a compounding loop. The remaining question is how much of this edge is proprietary know-how.
📌 X Insight Update[x_fin] (2026/06/28 07:17)
- $MAGS looks stretched versus $QQQ. The Magnificent 7 ETF just printed a fresh 28 month low relative to $QQQ, raising the odds of a rubber-band snapback rather than more straight-line underperformance. 1