以下部分引用 AI 总结,现阶段 AI 仍然有幻觉。请以内容的原文为准。
数据新鲜度群聊 07-03 ✓卖方 07-01 ⚠Wrap 06-30 ⚠X 07-03 ✓页面生成 07-05
scaling
3 atoms · 跨 3 天 · 首见 2026-06-09 · 最近 2026-06-16
三色: 🟦 fact 0 · 🟥 take 3 · stance ▲3/▼0/◆0
来源: X 3
时态: fresh:3
标签: 好思考:3
叙事 (narrative) (3)
- 2026-06-09 · 持续有效
but scaling just works
tags: 好思考 · → daily
value: direction=positive
- 2026-06-16 · 是GPT-3成功的关键洞见
THAT is the key insight that led to GPT-3. THAT is what Ilya saw. step #1: take a simple, elegant algorithm. step #2: shove compute at its face.
tags: 好思考 · → daily
- 2026-06-15 · bet against scaling is a failure mode
Always worth thinking about whether you’re implicitly making a bet against scaling, and how to course correct (unless that’s the bet you’d like to make!). Certainly a failure mode I’ve dealt with.
tags: 好思考 · → daily
时间轴 (近 20)
- 2026-06-16 ·
narrative · 是GPT-3成功的关键洞见 · → daily
- 2026-06-15 ·
narrative · bet against scaling is a failure mode · → daily
- 2026-06-09 ·
narrative · 持续有效 · → daily
← 实体目录 · 系统日志