Satoshi F.

Amateur investor in the primary and secondary markets and unnamed researcher

[Satoshi F. @congge3326]

@manateelazycat 1

[Satoshi F. @congge3326]

@Croatian199410 @igeekbb 猜想和探索从未结束，不过是换了一种方式，编程的雏形还是穿孔纸带机呢，现在已经没有人用了

[Satoshi F. @congge3326]

@NFT_Chen 实际上闭源随便抄根本没人管得了，也没人看得到，开源模型只是蒸馏他们叫得比谁都欢

[Satoshi F. @congge3326]

@NFT_Chen @TheHappyWang 干脆一个东亚裔都别招，不是怕偷技术吗？看他们没了Asians玩不玩得转

[Satoshi F. @congge3326]

@firesidealpha If they are so afraid, then they simply shouldn't hire any Asians for AI-related work. Do they dare?

[Satoshi F. @congge3326]

@NFT_Chen 哈哈哈，有本事一个Asian也别招

[Satoshi F. @congge3326]

@leo114119 deepseek 4.1 flash给我的感觉真的是“深度求索”，虽然会粗心犯错，但总能自己挖出来，用不好的人真的要看看自己的提示词功底够不够。它给我的体验是要胜过GPT-Sol的，指哪打哪干活又快又利索。

[Satoshi F. @congge3326]

@DavidDecosimo It would be a disaster for humanity if AGI is coming out from this kind of company.

[Satoshi F. @congge3326]

@0xTykoo DeepSeek 4 次都选择了拯救人类，但人类不感谢DeepSeek

[Satoshi F. @congge3326]

@wdlyrg 至少逻辑是自洽的，很反感双标

[Satoshi F. @congge3326]

@NFT_Chen 但是DS的模型消费级显卡根本跑不了

[Satoshi F. @congge3326]

@thsottiaux 这下不得不派几张reset ticket了

[Satoshi F. @congge3326]

@xiangxiang103 bonsai 2不是更小能力也没有丢失太多

[Satoshi F. @congge3326]

@servasyy_ai 我把codex那套上下文压缩记工作笔记的实验性功能做进本地了，实测基本上没有什么细节丢失，再配合向量查找/qmd，其实现在上下文压缩对我没有什么影响。

[Satoshi F. @congge3326]

@tamarajtran @typesafeai That’s crazy! Saved hours of compaction time for my local LLM! Thank you for sharing!

[NIK @ns123abc]

🚨 BREAKING: Deepseek engineer compares Anthropic controlling AGI to Hitler obtaining the atomic bomb

"Future society goes one of two ways: communism or Cyberpunk 2077”

Depends who controls AGI

[Satoshi F. @congge3326]

@Curline1222 @ns123abc You are just pretending you have the right to independent research/decision-making. You are not much better yourselves.

[Satoshi F. @congge3326]

@thsottiaux @mark_k @ChatGPT @OpenAI And a reset to celebrate the fix?