Xiaolei Lian

[Xiaolei Lian @xiaolei_1993]

@tspy @thsottiaux 这有啥的，我就是这样用公有云训练打比赛的。把公有云的开关机、数据传输写成 skill 不就得了

[Xiaolei Lian @xiaolei_1993]

@Lgk007 ？你做饭这么慢吗？

[Xiaolei Lian @xiaolei_1993]

@NFT_Chen Sam 又要瘫痪在椅子上了

[Xiaolei Lian @xiaolei_1993]

@CoderDaMing 好家伙这是调 codex 还是要高考写作文

[Xiaolei Lian @xiaolei_1993]

I built this with Grok — try it out: https://t.co/DL5HlAZh6m

[Xiaolei Lian @xiaolei_1993]

I built this with Grok: https://t.co/YTr9ySsCcS

[Xiaolei Lian @xiaolei_1993]

@derrickcchoi 例如打 kaggle

[Xiaolei Lian @xiaolei_1993]

@derrickcchoi 在数据科学建模上 codex 的 workflow 没有 claude-code 效果好。
claude-code 的思路更清晰且步骤有条理，codex 在思路上有些混乱。

[Xiaolei Lian @xiaolei_1993]

Just shipped Jiuzhou Xiangqi 3D — a playable 3D version of Chinese chess, built with GPT‑5.6 Sol.
Play it in your browser:
https://t.co/pEv2Pu8VTs

[Xiaolei Lian @xiaolei_1993]

@thsottiaux I used GPT‑5.6 Sol to create a playable 3D version of Chinese chess, and I’d love to keep improving it:

https://t.co/4rMAyuP7rr

I’ve used up my current quota during the build, so I’d really appreciate a reset. Happy to share some detailed feedback on my experience with Sol!

[Xiaolei Lian @xiaolei_1993]

@R1s666 @claudeai 醒醒吧，我两个号封一年了还是 review 中，A/ 根本不看的。

[Xiaolei Lian @xiaolei_1993]

@BrandonKHill ？

[Xiaolei Lian @xiaolei_1993]

@Khazix0918 我发现打 kaggle k3 能和 fable 比划比划 于是下个月我准备退订 Claude max

[Xiaolei Lian @xiaolei_1993]

@daimajia 微蹬了一下 确实不错

[Xiaolei Lian @xiaolei_1993]

@thsottiaux @Da7_Tech obviously, reset rate limit is not enouth

[Xiaolei Lian @xiaolei_1993]

Moving up to rank 12. rising like my electricity bill. #kaggle - https://t.co/rw9K4JKjEv

[Xiaolei Lian @xiaolei_1993]

@thsottiaux Stop tweeting and start chasing Claude code please

[Alias @Alias_labs]

🚀 Alias Mini App is LIVE! 🚀

📲 Dive into the future of #AI and #Crypto. We're on a mission to transform your world with #BlockchainTech.

Join the revolution at https://t.co/qn7J7JsDp8

#AliasMiniApp

[Xiaolei Lian @xiaolei_1993]

@elonmuskADO Twitter forever

[Xiaolei Lian @xiaolei_1993]

@Jousefm2 When training video generation tasks, will adding the Laplacian message passing mechanism generate videos that are more in line with physical intuition?