春日ゆう:
ケーキタイム〜🍰🍰🍰
#超かぐや姫
#ヤチヨ誕生祭2026
#かぐや誕生祭2026
#pixiv pixiv.net/artworks/14712…
https://twitter.com/KaKi3572607908/status/2076273569458475335
ケーキタイム〜🍰🍰🍰
#超かぐや姫
#ヤチヨ誕生祭2026
#かぐや誕生祭2026
#pixiv pixiv.net/artworks/14712…
https://twitter.com/KaKi3572607908/status/2076273569458475335
🔥9
&'a ::rynco::UntitledChannel
什么叫做 S39 试图用 COPV 鱼雷击沉机器船浮标啊 https://fixupx.com/NASASpaceflight/status/2075625452857696454
30X_Squid:
think fast!
https://twitter.com/stainless_squid/status/2076320682653229081
think fast!
https://twitter.com/stainless_squid/status/2076320682653229081
NSF - NASASpaceflight.com:
@SpaceX Epic content!
Love how Ship 39's final act was to torpedo the buoycam with a COPV 😅
https://twitter.com/NASASpaceflight/status/2075625452857696454
❤3
Forwarded from Hacker News
Networking and the Internet, from First Principles (🔥 Score: 154+ in 2 hours)
Link: https://readhacker.news/s/6Yi2Q
Comments: https://readhacker.news/c/6Yi2Q
Link: https://readhacker.news/s/6Yi2Q
Comments: https://readhacker.news/c/6Yi2Q
Faza
Networking and the Internet
How electricity, light, and radio move data across independently operated networks.
Morgz! 🫘🖌️:
Tom as an anime girl Page 2
SFWイラスト
https://twitter.com/morgzkun/status/2076639877987549424
Tom as an anime girl Page 2
SFWイラスト
https://twitter.com/morgzkun/status/2076639877987549424
🥰4🐳1
John D. Cook:
"Armed with generative AI, a B student can produce A work while turning into a C student." -- Nicholas Carr
https://twitter.com/JohnDCook/status/2076363738467184781
"Armed with generative AI, a B student can produce A work while turning into a C student." -- Nicholas Carr
https://twitter.com/JohnDCook/status/2076363738467184781
😁4
Forwarded from Hacker News
Show HN: Super Dario (🔥 Score: 150+ in 1 hour)
Link: https://readhacker.news/s/6YrVY
Comments: https://readhacker.news/c/6YrVY
Link: https://readhacker.news/s/6YrVY
Comments: https://readhacker.news/c/6YrVY
🥰1
scr.c:
不知道是不是暴论:感觉可以提出一个黄油Benchmark,因为压根不可能有厂家把黄油作为训练对象搞不好还会重点清洗,而前后端系统等都是模型数据重点,用户给它的任务极大概率和它的知识高度拟合所以很容易大力出奇迹。因此模型在黄油上表现(打满Context让它学习上下文)才是它的真实智力水平(?)。
https://twitter.com/silsrc/status/2076688559281770720
不知道是不是暴论:感觉可以提出一个黄油Benchmark,因为压根不可能有厂家把黄油作为训练对象搞不好还会重点清洗,而前后端系统等都是模型数据重点,用户给它的任务极大概率和它的知识高度拟合所以很容易大力出奇迹。因此模型在黄油上表现(打满Context让它学习上下文)才是它的真实智力水平(?)。
https://twitter.com/silsrc/status/2076688559281770720
🐳12