LEMONed

this is so dumb and pathetic. macrumors.com/2026/05/27/met…

@jajia 不知道哇 应该没有吧

@LEMONed 陽江是啥來著？

三個碳水臉！ E159 "碳水脸"被定性"挑动对立",官媒为何集体下场和概念斗争? | 中产内斗 | 身材焦虑 | 网信办 youtu.be/GTJHyNwsu0E?si…

in other words, Qwen is 94.6% bullshit.

@Cloudflare Verifying you are human. This may take indefinitely.

AP wins Pulitzer Prize for China surveillance reporting: buff.ly/3tStbFt

the @cloudflare security verification is extremely disgusting. even more so day by day.

@instagram is there any way I can contact help? you have dozens of help articles that are not helping at all.

In your 20s, you’ll meet someone who’s obsessed with you and wants you all the time. And this person will be your manager at work.

curse you @adobe

@bfishadow 说明该商家颇具规模，不然就直接从义乌发了。

非常离谱…… 我前阵子在某澳大利亚电子商务网站（由Shopify强力驱动）购买了一个小玩意儿，商家说 Worldwide Delivery，邮费只需要五十元。 然后第二天商家 email 跟我联系问我中文地址和电话…… 所以当我今天收到顺丰速运预告说有一票货从宁波发来的时候，我的心情是非常复杂的…… #一件小事

谁通知一下牢A？

🤔 伍雷如果没有润，他会不会替许家印辩护？

@jajia 我经常记不住「费迪南大公」

今天直播腦子不轉，把瓦爾登湖說成瓦登爾湖⋯⋯⋯網友說：這人真的讀書嗎⋯⋯ 對我真是奇恥大辱啊⋯⋯ 有人犯過類似錯誤嗎？ 四川大學一食堂有一道菜叫鍋巴雞片，每天都有學生說成雞巴鍋片。

🚨BREAKING: OpenAI published a paper proving that ChatGPT will always make things up. Not sometimes. Not until the next update. Always. They proved it with math. Even with perfect training data and unlimited computing power, AI models will still confidently tell you things that are completely false. This isn't a bug they're working on. It's baked into how these systems work at a fundamental level. And their own numbers are brutal. OpenAI's o1 reasoning model hallucinates 16% of the time. Their newer o3 model? 33%. Their newest o4-mini? 48%. Nearly half of what their most recent model tells you could be fabricated. The "smarter" models are actually getting worse at telling the truth. Here's why it can't be fixed. Language models work by predicting the next word based on probability. When they hit something uncertain, they don't pause. They don't flag it. They guess. And they guess with complete confidence, because that's exactly what they were trained to do. The researchers looked at the 10 biggest AI benchmarks used to measure how good these models are. 9 out of 10 give the same score for saying "I don't know" as for giving a completely wrong answer: zero points. The entire testing system literally punishes honesty and rewards guessing. So the AI learned the optimal strategy: always guess. Never admit uncertainty. Sound confident even when you're making it up. OpenAI's proposed fix? Have ChatGPT say "I don't know" when it's unsure. Their own math shows this would mean roughly 30% of your questions get no answer. Imagine asking ChatGPT something three times out of ten and getting "I'm not confident enough to respond." Users would leave overnight. So the fix exists, but it would kill the product. This isn't just OpenAI's problem. DeepMind and Tsinghua University independently reached the same conclusion. Three of the world's top AI labs, working separately, all agree: this is permanent. Every time ChatGPT gives you an answer, ask yourself: is this real, or is it just a confident guess?

一个反直觉的事实：包括 DeepSeek、千问在内的中国大模型，其逻辑推理能力的底座是用英文世界的高质量语料（代码、论文、技术文档）训练出来的。模型从英文语料中习得的推理模式，在输出中文时被保留了下来——低语境、重逻辑、主谓宾清晰。 这让大量中国网民第一次在日常场景中持续接触到了结构化、高信息密度的中文表达。过去这种风格只存在于翻译体媒体（比如纽约时报中文网）或少数作者的输出中。曾经我第一次看这些媒体的感受是，原来中文新闻可以这么写... 现在每天跟 DeepSeek、Kimi 对话、让它写邮件改方案，就是在被这种语言风格反复浸泡。对于过去浸泡在官话和高语境语言环境的中文使用者来说，算是一个语言环境的升级吧.

@lyricwai @cryptonerdcn 我无奈表示挺同意的

@cryptonerdcn 我昨天想了这个事情，觉得，长远看品味差也没关系，最后有个什么机制优胜劣汰就行。就跟咱们体内的垃圾基因似的。

这篇最有用的点就一句：AI把开发成本降低后，不管好的坏的点子都能被完成的人模狗样，反而增加了开发者判断的难度。 其实就是“品味才是护城河”的另一种说法。

@jajia 推荐两个电影，Elysium《极乐空间》和 In Time《时间规划局》。对「人类分化为两个阶级」的设定都还不错。

下播晚了一點，和上海朋友約見面，雖然已經誠意滿滿地打車來了，但還是遲到了🥹

@fxtrader The Chinese government floods X search results with porn whenever there is political unrest—to prevent their citizens from finding out real-time information. This has been a difficult problem to solve but we are aware & working on it.