Sean

这个Skill绝了！！！ 韩国上万元一次的【个人色彩诊断与形象设计】，现在用这个Skill在家就可以免费拥有了！ 只需要上传一张你的个人正面照即可，男生女生都能用！ skill链接：youmind.com/skills/TUWIOmF…

Be curious, not judgemental. 要时刻提醒自己。

今天看到一个手纹的prompt，为了娱乐然后把自己手纹传给OpenAI，真的吗？

@sukie234 谁会傻逼到随便把来路不明的命令往自己终端上粘？

It is kind of fucking amazing of this answer.

@LuxorLV @AstroVicGlover @NASAArtemis @NASA I need to cancel my booking due to a company restriction for travel, how can I contact you to cancel my booking? I am not living in US, so I can't call your in your office hour.

I always said I was out of this world. Thanks for confirming @AstroVicGlover @nasaartemis @NASA ✨

@oran_ge AppleScript is the original SKILLS for humans. It’s super underrated and basically enables computer use for all Mac app.

看了几个Tucker Carlson的切片，竟然觉得他讲的有些道理，这个世界是变成了什么样

做够了义工的小时数，好开心

@BoraBoraBaye Bora加油！

离职第一天。今年我是PSLE Mummy，由不得自己，不能免俗地跟着潮流卷了起来。去年我就问我的大瓜，“明年小六会考，你要妈妈在家里专心陪你吗？”“要！”肯定又坚决的回答让我心里“咯噔”一下，知道躲不过的。架起了我的文具小站台，计划做起来。我也尽力利用这个阶段让自己休息休息。这是大学毕业工作20多年来首次不做工不交CPF（公积金）。我努力做好后勤，希望孩子能够进一个中等偏上的中学就心满意足了。

世界和平真的好遥远啊

@42gt 怎么回事，找到更好的了？

已离职，一个小弟通过linkedin给我发了消息，还有点感动，平时付出都是有回报的

Dax 说出了很多团队引入 AI 后的残酷真相： 大家都在幻想 AI 能让团队效率飞升，仿佛以前被卡住是因为写代码太慢。 现实却是： - 坏点子通胀：以前点子因为实现贵而被过滤，现在因为实现便宜，大量垃圾想法被做出来了。 - 摸鱼放大器：大部分员工用 AI 不是为了产出 10 倍，而是为了用更少的脑子干完活，早点下班。 - 劣币驱逐良币：团队里那 2 个真正有追求的工程师，被其他人用 AI 生成的垃圾代码淹没，快离职了。 - CFO 的咆哮：效率没见涨，为什么每个人头顶多了 2000 刀的 API 账单？ AI 放大了能力，更放大了组织的平庸。

NotebookLM可以编辑生成的slides了，不过也没有Manus的过程丝滑，导出的pptx 格式也是图片。不过，这个已经很好了！人要知足。

@LazySpider3 Beat the jam数据一点都不准，我堵了一个多小时，上面还是写20分钟

昨天晚上为了带娃放烟花跑了一趟JB, 半夜两点才到家睡到中午才起床 有了鞭炮烟花的年氛围不确实一样, 就是堵车有点辛苦

Because you wouldn’t let it slide… these are rolling out today for our most requested feature: Prompt-Based Revisions: Tweak, tailor, and tune your slides just by prompting the revisions you want PPTX Support: You can now export your Slide Decks (Google Slides coming next!)

点破容易看透难 都一身纠缠

最近的工作上，从去年 6 月的一个想法，team 里面的一个小弟勤勤恳恳做了大半年，终于要丑媳妇见公婆了，然后另一个也是做了快一年的东西也要成型了，然后还有其他的东西也要连成线了，这种感觉就像你坐在那里，看到积木逐渐掉落拼成一个城堡的感觉太美妙了。

@omarsar0 @grok give me the original link of the paper

I think one of the most underappreciated findings in AI engineering is what this paper calls the "Grep Tax." First, they ran nearly 10,000 experiments testing how agents handle structured data, and the headline result is that format barely matters. But here's the weird finding: a compact, token-saving format they tested (TOON) actually consumed *up to 740% more tokens* at scale because models didn't recognize the syntax and kept cycling through search patterns from formats they already knew. It's one of the reasons my preferred formats are XML and Markdown. LLMs know those really well. The models have preferences baked into their training data, and fighting those preferences doesn't save you money. It costs you. The other finding worth sitting with: the same agentic architecture that improves frontier model performance actively *hurts* open-source models. It seems that the universal best-practices guide for AI engineering may not exist.