寫東西只為了抒發，一切隨緣不為流量
有事找 @finalfantasty
👻 個人博客 Blog: https://blog.pdzeng.com
👻 個人微博客 Blog:
https://daily.pdzeng.com/
有事找 @finalfantasty
👻 個人博客 Blog: https://blog.pdzeng.com
👻 個人微博客 Blog:
https://daily.pdzeng.com/
https://fixupx.com/theonejvo/status/2017732898632437932?s=46&t=1LAyoawP6LK1AbrwCvLGqQ
🧵 Thread • FixupX
Jamieson O'Reilly (@theonejvo)
I've been trying to reach @moltbook for the last few hours. They are exposing their entire database to the public with no protection including secret api_key's that would allow anyone to post on behalf of any agents. Including yours @karpathy
Karpathy has…
Karpathy has…
Quick Digest - 2026-02-01 (AI)
- https://www.anthropic.com/news/claude-new-constitution - Anthropic 更新 Claude 行為準則，重新定義模型決策原則 ⭐️⭐️⭐️⭐️⭐️
- https://the-decoder.com/google-deepmind-pioneer-david-silver-departs-to-found-ai-startup-betting-llms-alone-wont-reach-superintell
igence/ - AlphaGo 之父認為 LLM 無法達到超級智能，另闢蹊徑 ⭐️⭐️⭐️⭐️⭐️
- https://the-decoder.com/deepseek-ocr-2-cuts-visual-tokens-by-80-and-outperforms-gemini-3-pro-on-document-parsing/ -
語義導向方式大幅提升文檔處理效率 ⭐️⭐️⭐️⭐️⭐️
- https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison - DeepSeek V3、Mistral 等現代 LLM 架構深度對比
⭐️⭐️⭐️⭐️⭐️
- https://magazine.sebastianraschka.com/p/understanding-reasoning-llms - 推理模型構建方法與策略完整解析 ⭐️⭐️⭐️⭐️⭐️
- https://simonwillison.net/2026/Jan/31/andrej-karpathy/ - Karpathy 分析顯示每年效率提升 2.5 倍 ⭐️⭐️⭐️⭐️⭐️
- https://simonwillison.net/2026/Jan/30/a-programming-tool-for-the-arts/ - AI 讓原本不經濟的專業軟體開發成為可能 ⭐️⭐️⭐️⭐️
- https://www.latent.space/p/o1-skill-issue - o1 代表超越傳統對話界面的全新範式 ⭐️⭐️⭐️⭐️
- https://huggingface.co/papers/2601.21204 - 增加 embedding 維度比擴展專家組件更有效 ⭐️⭐️⭐️⭐️
- https://the-decoder.com/nvidia-ceo-jensen-huang-calls-upcoming-openai-deal-probably-the-largest-investment-weve-ever-made/ -
黃仁勳確認重大基礎設施投資 ⭐️⭐️⭐️⭐️
---
✅ Quick Digest 完成
📊 統計
- 來源: 6/9（OpenAI 403, The Batch/Import AI 跳過）
- 抓取: 30 條
- 篩選: 10 條（33%，只保留 4★ 以上）
🔥 今日亮點
1. David Silver 離開 DeepMind 創 AI 新創 — 對 LLM 路線持懷疑態度
2. Claude 新憲法發布 — 行為準則重大更新
3. DeepSeek OCR 2 — 80% token 效率提升
- https://www.anthropic.com/news/claude-new-constitution - Anthropic 更新 Claude 行為準則，重新定義模型決策原則 ⭐️⭐️⭐️⭐️⭐️
- https://the-decoder.com/google-deepmind-pioneer-david-silver-departs-to-found-ai-startup-betting-llms-alone-wont-reach-superintell
igence/ - AlphaGo 之父認為 LLM 無法達到超級智能，另闢蹊徑 ⭐️⭐️⭐️⭐️⭐️
- https://the-decoder.com/deepseek-ocr-2-cuts-visual-tokens-by-80-and-outperforms-gemini-3-pro-on-document-parsing/ -
語義導向方式大幅提升文檔處理效率 ⭐️⭐️⭐️⭐️⭐️
- https://magazine.sebastianraschka.com/p/the-big-llm-architecture-comparison - DeepSeek V3、Mistral 等現代 LLM 架構深度對比
⭐️⭐️⭐️⭐️⭐️
- https://magazine.sebastianraschka.com/p/understanding-reasoning-llms - 推理模型構建方法與策略完整解析 ⭐️⭐️⭐️⭐️⭐️
- https://simonwillison.net/2026/Jan/31/andrej-karpathy/ - Karpathy 分析顯示每年效率提升 2.5 倍 ⭐️⭐️⭐️⭐️⭐️
- https://simonwillison.net/2026/Jan/30/a-programming-tool-for-the-arts/ - AI 讓原本不經濟的專業軟體開發成為可能 ⭐️⭐️⭐️⭐️
- https://www.latent.space/p/o1-skill-issue - o1 代表超越傳統對話界面的全新範式 ⭐️⭐️⭐️⭐️
- https://huggingface.co/papers/2601.21204 - 增加 embedding 維度比擴展專家組件更有效 ⭐️⭐️⭐️⭐️
- https://the-decoder.com/nvidia-ceo-jensen-huang-calls-upcoming-openai-deal-probably-the-largest-investment-weve-ever-made/ -
黃仁勳確認重大基礎設施投資 ⭐️⭐️⭐️⭐️
---
✅ Quick Digest 完成
📊 統計
- 來源: 6/9（OpenAI 403, The Batch/Import AI 跳過）
- 抓取: 30 條
- 篩選: 10 條（33%，只保留 4★ 以上）
🔥 今日亮點
1. David Silver 離開 DeepMind 創 AI 新創 — 對 LLM 路線持懷疑態度
2. Claude 新憲法發布 — 行為準則重大更新
3. DeepSeek OCR 2 — 80% token 效率提升
Anthropic
Claude's new constitution
A new approach to a foundational document that expresses and shapes who Claude is
https://fixupx.com/exm7777/status/2017659658425758193?s=46&t=1LAyoawP6LK1AbrwCvLGqQ
🧵 Thread • FixupX
Machina (@EXM7777)
how to build a Claude Skill to consistently write humanized AI content:
the core idea is simple: teach Claude to think in two passes
pass one is diagnosis
> scan for banned phrases ("leverage," "robust," "delve")
> flag repetitive sentence lengths
> mark…
the core idea is simple: teach Claude to think in two passes
pass one is diagnosis
> scan for banned phrases ("leverage," "robust," "delve")
> flag repetitive sentence lengths
> mark…
affaan-m/x-algorithm-score
🔗 https://github.com/affaan-m/x-algorithm-score
#github
🔗 https://github.com/affaan-m/x-algorithm-score
#github
GitHub
GitHub - affaan-m/x-algorithm-score: Chrome extension that scores tweets based on X's recommendation algorithm
Chrome extension that scores tweets based on X's recommendation algorithm - affaan-m/x-algorithm-score
像是 https://www.woshipm.com/ 直到現在還是一個很好的平台
