llamacorn.eth

I taught a speech model to understand context in conversation. This is what happened It adjusts voice and tone to express urgency, comfort, understanding from the dialogue. Just like a real human being 520M model. Runs locally on consumer devices How this is achieved 🧵

@xingpt 谢谢蒋老师，这是不是AI写的

开工第一周想必各位都感觉昏昏沉沉，身体在公司，心里还在老家的温柔乡或者旅行的远方，这里分享一下新年开运的小诀窍，大家可以尝试做一下。 【AI时代新年开运十诀】 1）主动做一点让身体“有痛感”的事 10度以下的水冰浴、3RM深蹲、拉伸到极限、泡沫轴滚大腿前侧最紧的点、HIIT到 Zone4。 痛感会让大脑短暂清空，杂念退场，注意力回到当下。随后内啡肽释放，你会感到一种通透的爽感。 身体的疼痛，也是在通气血、醒心神。 气血一旦重新回到身体，人的状态会先变，运势才会跟着变。 2）读懂自己的身体 用 Whoop、Oura、Apple Watch 或任何工具都行。知道自己的 HRV、恢复水平、睡眠质量。 也知道自己对甘氨酸镁、苏糖酸镁、Q10、鱼油、维生素B族这些补剂的反应。 运势的底层，是生理状态。 你的“身强”，才接得住横财和大运。 3）长待的地方朝阳 卧室、客厅、办公室，至少有一个主要活动区域能稳定晒到太阳。 阳光是最便宜的阳气补给。 人长期待在阴暗环境里，心气容易下沉，情绪和行动力都会变钝。 空间有光，人的气场会更开，做事也更容易顺势。 4）每天花5分钟肯定自己 找到今天三个做得好的地方，越具体越好。 训练正向你的潜意识。 一个人长期自我否定，能量会收缩，心念会变弱。长期自我确认，能量会展开，行动会更稳。 很多时候，所谓开运，就是先停止对自己的消耗。 5）每天当面夸赞一个人 同事、朋友、家人，甚至陌生人都可以，真诚就行。 赞美是一种很强的能量交换。你给出去的善意，会在关系网络里留下回响。 人缘本身就是运势的一部分。你周围的善意越多，托举你的力量也就越大。 6）给手机设边界，守住注意力 睡前不带手机上床，醒来不立刻看手机。 每周至少一天不刷短视频和社交媒体。 AI时代最稀缺的资源是完整的注意力。 注意力也是人的神识和心神。起床就刷手机，很容易一早把自己的气散掉。 7）高价值内容读原文 深度文章要读原文，不要所有文章丢给AI 总结。 AI擅长提炼观点，但常常会压扁推理过程，省略案例细节，也会抹掉行文的过渡和层次。 你读的是路径，也是作者的思维筋骨和文气。 读长文也是养灵魂和心力的过程，灵魂越饱满，心力越强大，整个人的气质和状态也越好。 8）少抱怨，多转念 抱怨会消耗心力，也会拖慢气场。 遇事先强迫自己找一个正面的思考角度。 正念是保护自己培育的能量不轻易松懈掉。 心一乱，局就乱。l心一稳，局就有解。 长期看，情绪稳定本身就是一种稀缺能力。会转念的人，往往也更容易转运。 9）火年更要修“和气” 丙午属火，火旺则快，快则燥，燥则容易起冲突、下重话、做错判断。 所以火年更要与人为善，少指责，多体谅。 和气不是软弱，是一种更高水平的控场能力。口德稳，关系就稳；关系稳，财气和人气才留得住。 火气收得住的人，往往更有后劲。 10）戒糖，是最被低估的“改运术” 糖摄入过多会让血糖波动变大，人容易昏沉、烦躁、发胖，睡眠和恢复也会受影响。 身体一浑，脑子就不清。脑子不清，决策就容易走偏。 很多人以为自己运气差，其实只是长期处在高波动、低清醒的状态里。 稳定血糖，本质上是在稳定心神和执行力，也是在给自己的运势打地基。 最后祝大家马年马到功成，一马平川，手到心到，必有收获！

@bee926cn @WisprFlow Oh你还保留着台湾口音🤣

最近開始用 @WisprFlow ，基本上不打字了 這是個 AI 語音輸入工具，你用說的它幫你打字 重點是： 你想的永遠比你打的快，很多好的想法還沒打完就忘了 用說的就不會 我現在每天用語言跟我的 AI agent Lucy 對話，中英夾雜它完全聽得懂 🤯 腦袋想到什麼就直接講，根本不需要鍵盤了

@lanhubiji Cypto 老师你是故意的还是不小心🤣

@_Dave__White_ What you feel marks you as a sensitive, kind guy. All your feelings are valuable. I think you can finally find a way out.

it's been almost half a year since i've written anything on here one of the biggest reasons is that twitter (do we still call it that?) has become somewhat overwhelming and depressing for me i guess i've always been a bit of a pessimist, but it seems like the rate of shit going extremely wrong has been accelerating dramatically recently, and every time i open the app to try to post i see something horrifying or jealousy-inducing and get sucked in and end up closing out in half an hour kind of fucked up for the rest of the day but i also think i've lost something significant by cutting off my main point of connection to the outside world at large, which is something i'd like to rectify going forward just to get the things out of the way that i'll probably be talking about so i'm not afraid to say it anymore: - i've been spending a lot of time learning to play music. currently learning piano, drums, and production in ableton. i'm very new at this but i'm honestly proud of many of the things i've made. - i'm increasingly horrified by the rate of ai progress and our lack of technological and social infrastructure to ensure that that goes well. i think we probably have one or two years left of anything vaguely normal looking left. - that belief makes it very, very hard to orient and find meaning in things. i.e. music. why learn to haltingly play the e flat major scale when suno is already better at piano then i'll likely ever be? - there are people, i suspect a lot of them who are going to read this post, who think this belief means i'm deluded or weak and i'm an example of how a dangerous ideology can ruin lives. they may be right. i really don't think so though. - i suspect the reason more people aren't feeling this way is that it is too horrifying to look at head on, and/or accepting it as real would imply too many very inconvenient or upsetting life changes, so we choose to look the other way. we've been doing this with death since time immemorial. - i'd like to help people wake up, though, and i'll be trying to do that more in the coming weeks and months

@xingpt ？ 蒋老师露那么多吗 要保护好自己啊

站在2026年的起点回望，过去这一年我最大的变化，可能就是身材上的重塑。 先分享一下数据：我的体重从 79kg 降到了 71kg，体脂率从 19% 降到了 14%，状态最好的时候甚至到过 12%。这些数字对我来说，不仅是外形的变化，更是一次对生活习惯的重新审视。 很多人觉得减肥或者健身就是拼命运动。其实去年我最胖的时候，恰恰是我练得最猛的时候。那时候我每天中午都会挤出时间去跑步，结果却并不理想。 这让我意识到，身材管理更重要的是科学的健身方法和理念。 关于“道”：先搞清楚身体的账单 这里我想分享两个对我帮助很大的概念：基础代谢率（BMR）和每日能量消耗总量（TDEE）。 你可以把它们理解成你身体的“能量账单”。BMR 是你躺着不动也会消耗的热量，而 TDEE 是加上你每天走路、工作和运动后的总消耗。 现在有很多在线工具或者 AI 软件，只要输入身高、体重和年龄，就能算个大概。 我的心得是： * 如果想减脂，摄入的热量就要比 TDEE 少一点。 * 如果想增肌，摄入就要多一点。 * 重点在于，无论怎么减，摄入量绝对不能低于 BMR，否则肯定掉肌肉。我之前出差时饮食不规律，吃的比基础代谢还少，结果肌肉掉得很厉害，身体状态也变差了。 所以，看清自己的账单，是所有行动的前提。 关于“吃”：结构比热量更重要 大家都说“三分练，七分吃”，我也深有体会。但我并不建议大家去苦行僧式地节食，关键在于调整饮食结构。 我个人比较推崇 211 饮食法：餐盘里一半是蔬菜，四分之一是碳水，四分之一是蛋白质。 作为一个北方人，我非常爱吃面食，以前家里几乎天天吃面。但后来我发现，精致碳水（高 GI 食物）会让血糖波动很大，容易让人产生饥饿感，结果反而吃得更多。 另外一个容易被忽视的是蛋白质。 我们平时的饮食习惯里，肉类其实吃得不够。建议大家根据自己的体重来算，每公斤体重摄入 1.5 到 2 克蛋白质。 像我 75kg 的时候，每天要吃够 150g 蛋白，这确实需要靠蛋白粉来辅助补充。 在这里分享一个我自己的小偷懒方法：我会在 ChatGPT 里建一个专门的文件夹，每天把三餐拍张照片发给它。（见图片说明） AI 会帮我估算热量和三大营养素的比例。有了这个反馈，我每天吃得心里很有底，这种掌控感是坚持下去的关键。 关于“练”：找对节奏，寻求专业帮助 在锻炼这件事上，我坦诚自己不是一个自律性极强的人。一个人练的时候，我经常会忍不住玩手机，训练效率很低。 所以我的建议是：如果条件允许，一定要请教练。 教练存在的意义不仅仅是教你动作，更重要的是能盯着你，保证那一小时的训练质量。我还有一个小习惯，就是每隔几个月会换一个教练。因为每个人的专长不同，对你身体弱点的判断也不一样，换个视角往往能带来新的进步。 在频率上，我尽量保持每周 3 到 4 次训练。虽然出差时很难保证质量，但我会尽量选有健身房的酒店，哪怕只是简单动动。 关于“养”：那些容易被忽略的细节 最后我想说两个非常关键、但大家容易忽视的点：睡眠和戒酒。 肌肉其实是在睡觉时长的，而不是在健身房长的。重训是“破坏”过程，休息才是“生长”过程。如果睡不好，身体的皮质醇会升高，练得再辛苦也难见成效。 另外就是酒。酒精对健身的影响非常大，它会直接阻断蛋白质的合成。虽然很多人说为了业务不得不喝，但我认识的很多成功人士其实都不怎么喝酒。 如果你真的想改变身体，在工作方式和健康之间，可能需要做一个取舍。至于零食和深加工食品，尽量在你的清单里划掉吧。 写在最后 健身对我来说，不是为了成为什么专家，也有非常多的大神比我练的好很多，但对于我一年的科学健身，给我带来了稳定进步的获得感，也带来了一个更好的精神状态。 现在的我，比之前都更有活力，也更了解自己的身体。 如果你也打算在 2026 年开始改变，不要急着去健身房办卡，可以先试着记录一下自己的饮食，算一算自己的能量账单。 祝大家在 2026 年都能拥有一个健康的身体，我们一起每天进步一点点。 有健身和饮食管理问题的，欢迎在评论区交流哦！

@Jrag0x 🤣🤣🤣

Dankrad right now

I’m turning 41, but I don’t feel like celebrating. Our generation is running out of time to save the free Internet built for us by our fathers. What was once the promise of the free exchange of information is being turned into the ultimate tool of control. Once-free countries are introducing dystopian measures such as digital IDs (UK), online age checks (Australia), and mass scanning of private messages (EU). Germany is persecuting anyone who dares to criticize officials on the Internet. The UK is imprisoning thousands for their tweets. France is criminally investigating tech leaders who defend freedom and privacy. A dark, dystopian world is approaching fast — while we’re asleep. Our generation risks going down in history as the last one that had freedoms — and allowed them to be taken away. We’ve been fed a lie. We’ve been made to believe that the greatest fight of our generation is to destroy everything our forefathers left us: tradition, privacy, sovereignty, the free market, and free speech. By betraying the legacy of our ancestors, we’ve set ourselves on a path toward self-destruction — moral, intellectual, economic, and ultimately biological. So no, I’m not going to celebrate today. I’m running out of time. WE are running out of time.

所有关于AI的一切，推荐 关注@LuozhuZhang , bro knows everything about 🤖 AI ☺️

Is this the new scam? Did your twitter get hacked?? @xiangrenNLP 我已经很久不阅读dm了，因为几乎全是骗子，这是什么新型骗局吗？看推特是很多共同follower，但是你不能dm回复他，私信给了一个tg handle, 也不知是不是本人。这种线上开会的事儿，似乎也很多人中招。 Btw, 目前我对AI更是一窍不通

> “He sent out an email on Friday evening saying everything is going to deep learning, and that we were no longer a graphics company,” Greg Estes, a vice president at Nvidia, said. “By Monday morning, we were an AI company. Literally, it was that fast.” This is a prime example of Jensen Huang's leadership. Following the success of AlexNet, Jensen quickly recognized the potential of neural networks and declared overnight that Nvidia would change its business model from graphics computing to artificial intelligence.

@LuozhuZhang nice

For AI products, people generally think the intelligence dominates everything, while privacy and cost are seen as secondary. The industry’s path shows this: we’ve spent huge money in leading labs to build the largest models with exceptional intelligence. But I believe we’ve overlooked another path that’s just as important — the case for local models. Where local models are slowly emerging: - Cognitive Kernel of the SLM OS. This is the cognitive core of the OS. It doesn’t need to be very large or know everything. it only needs to understand the user’s intent and call the right apps or tools when needed. Ideally, a few billion parameters will be enough. It’s built directly into the OS with native support for text/audio/vision. So users never need to download or configure anything, and it will automatically return the result in the right format, whether text/audio or vision. - Super Personal Assistant App. This is the application layer of the SLM OS. It is built as an execution agent that works offline with access to the local device and application data, coordinating and interpreting your actions. For example, most AI assistants can only reply to an email. This one can pull from unified application data, summarize your meeting notes, and draft a reply the way you want, while leaving the final decision to send up to you. It can also learn from user feedback, continually improving how it handles tasks. The killer feature is cross-app automation + local brain search. For instance, when you ask “When did I say XXX?” or “Where is the photo of me and XXX?” it can return the correct result in less than 500 milliseconds. - Game characters in AI-native games. Traditional games rely on scripts and behavior trees to control game characters. After a few tries, everything feels repetitive and predictable, and players end up quitting. With SLMs combined with natural TTS, that logic is completely changed. Through deep integration of SLMs with the game engine, every NPC can become a unique companion (with their own personality/background/speaking style). More than that, the storyline can follow the choices made by the player and their companions. This is what we call a “never-ending game.” And these models live on your device, built right into the game files so you hardly notice them. They can remember the adventures you share, the stories you tell, and the things you care about. Over time, they can feel like your best friend. Local models win on these factors: - Low interaction latency: local models can respond in < 500 ms, with some native OS operations in < 50 ms. Game character speaking in < 800 ms close to human conversation speed. - Private data access: the cognitive kernel of the SLM OS can natively access local data, while LLMs never can. Data quality decides everything for an AI product, so it is reasonable to see local SLMs perform better than LLMs. - On-device finetuning: we may see better fine-tuning techniques that enable test-time training directly on edge devices. This would allow SLMs to improve personalization by learning from user interactions. - Everyday tasks: most of the things we do each day are relatively simple. So we’d rather get an 85/100 answer in < 500 ms than wait 10 minutes for an LLM to call multiple tools just to give a 95/100 answer. - Cost: whether it’s an OS or a game NPC, local SLMs can be used infinitely at zero cost, with no need to worry about inference expenses. - Ownership: not your weights, not your brain. Yes, LLMs will continue to get smarter, but most of our daily needs remain simple and unchanged. In some key domains, local SLMs can even perform better than LLMs. I believe we’ll see more impressive SLM use cases in the next 3–6 months, and it shouldn’t be a surprise if some of the best products don’t come from the big labs.

@llamacorneth 小姐姐你搞错了 我们这集没有讲吴京 🤠

小宇宙上有位听众听了这期后留言说： 「这集很好。这个行业不适合我。」 哎 感觉挺对不起人家的。

@Darrenlautf Oneill cheap and soft

Anyone got recommendations for microplastic free clothes?

@zhusu idiot

The idea of the submissive asian woman is one of the biggest psyops ever inflicted on white men. Mao offered Kissinger 10 million Chinese women in 1973, to which Kissinger replied, “it is such a novel proposition. We’ll have to study it.”

@RealSat_Intern @evilcos 那我这个虚拟号码可能以前有另一个人在用

@llamacorneth @evilcos 我刚验证了一下：登出一个账户（并且移除这个 device， 否则之前被记住，就不需要验证了），然后再登陆，自动发验证码给了我登陆的另一个设备，而不用 fragment

可以的，Telegram 盗号流程做得越来越方便用户被盗号了…😂 特别注意下图三，你如果没有启用 Two-Step Verification，一旦将 Login code 当成验证码输入，你的号就没了…

@RealSat_Intern @evilcos 但是如果是虚拟号码，不应该先要打开钱包fragment, 证明你有这个号码，才可以request login code吗？

@llamacorneth @evilcos 他如果尝试用你的号登陆，验证码就会发送给你