closed labs

5 atoms · 跨 5 天 · 首见 2026-04-15 · 最近 2026-06-17

三色: 🟦 fact 0 · 🟥 take 5 · stance ▲2/▼1/◆1

来源: X 5

时态: fresh:4 · aging:1

标签: 好观点:2 · 好思考:2 · 好问题:1

叙事 (narrative) (4)

• 2026-06-17 · 优化推理栈
  they have many more engineers working on optimizing the inference stack
  tags: 好思考 · → daily
• 2026-06-15 · developed strongest language models past five years
  why is it that the closed labs have generally developed the strongest language models at any given time in the past five years?
  tags: 好问题 · → daily
• 2026-06-09 · 会对不构建独立模型和基础设施的实体进行侮辱性支配
  if you don't build out sovereign and independent models+infra closed labs will patronize you to an insulting degree
  tags: 好观点 · → daily
• 2026-04-15 [老化中] · ability to dominate open-weight models on capabilities via RL and distribution
  The RL dominated training era has increased the relevance of distribution to real-world use-cases as a key factor in continued capabilities improvements. This is the first clear technical area that closed labs can dominate open-weight models on capabilities, potentially leveraging online RL directly based on user feedback.
  tags: 好观点 · → daily
  value: direction=can dominate

预测 (forecast) (1)

• 2026-05-03 · RSI and frontier gains potential if RSI online
  If RSI is already online, frontier gains will become ridiculous within a few months.
  tags: 好思考 · → daily

时间轴 (近 20)

• 2026-06-17 · narrative · 优化推理栈 · → daily
• 2026-06-15 · narrative · developed strongest language models past five years · → daily
• 2026-06-09 · narrative · 会对不构建独立模型和基础设施的实体进行侮辱性支配 · → daily
• 2026-05-03 · forecast · RSI and frontier gains potential if RSI online · → daily
• 2026-04-15 · narrative · ability to dominate open-weight models on capabilities via R · → daily

← 实体目录 · 系统日志