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 07-14 ✓卖方 07-14 ✓Wrap 断档15天 ✕News 07-15 ✓X 断档12天 ✕页面生成 07-15
Claude model
7 atoms · 跨 1 天 · 首见 2026-06-11 · 最近 2026-06-11
三色: 🟦 fact 2 · 🟥 take 5 · stance ▲0/▼4/◆3
来源: X 7
时态: fresh:7
标签: 好观点:7 · 好数字:1
叙事 (narrative) (7)
- 2026-06-11 · walk back time 24 hours
kinda wild it only took them 24 hours to walk it back though
tags: 好数字·好观点 · → daily
value: qty=24 hours
- 2026-06-11 · model moderation and safeguards feature
Model moderation and safeguards have always been a feature of deployed frontier models
tags: 好观点 · → daily
- 2026-06-11 · obfuscation without warning violation of contract
obfuscation without warning is a violations of the contract between the user and the provider
tags: 好观点 · → daily
- 2026-06-11 · trust recovery difficulty
For me personally it will take a while to regain trust
tags: 好观点 · → daily
- 2026-06-11 · subtle disclosure and walk back of policy
they were pretty subtle in both the disclosure and walk back of the policy
tags: 好观点 · → daily
- 2026-06-11 · disagreement with blocking code development use
I still disagree with them blocking use for developing code related to models
tags: 好观点 · → daily
- 2026-06-11 · sabotaging users silently is less hostile
atleast it feels less insanely hostile to sabotage users silently
tags: 好观点 · → daily
时间轴 (近 20)
- 2026-06-11 ·
narrative · model moderation and safeguards feature · → daily
- 2026-06-11 ·
narrative · obfuscation without warning violation of contract · → daily
- 2026-06-11 ·
narrative · trust recovery difficulty · → daily
- 2026-06-11 ·
narrative · subtle disclosure and walk back of policy · → daily
- 2026-06-11 ·
narrative · disagreement with blocking code development use · → daily
- 2026-06-11 ·
narrative · sabotaging users silently is less hostile · → daily
- 2026-06-11 ·
narrative · walk back time 24 hours · → daily
← 实体目录 · 系统日志