nickelily (@nickelily) header
nickelily (@nickelily) profile picture

Love & Peace

Joined March 2009
Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)Photo by nickelily (@nickelily)
预测:2026年7月,股市见顶,AI泡沫破裂。
4
1
5
262
J’ai rassemblé toutes les interviews d’Elon Musk depuis 1998 sur un seul site. 651 vidéos. 508 heures. 4 492 citations. Et 418 de ses prédictions vérifiées une par une. Tout se regarde directement sur le site, en français et en anglais. silicon-valley.tech
320
1,070
123
5,600
306,603
nickelily retweeted
Had a little time on my hands this morning to do some writing… send along to someone you know who may benefit --- “Practical principles to counter a modern-day coup/psyop/struggle session” 1) Own what you did but fiercely resist accommodating the mob, resist owning half-truths or actions/behaviors you didn’t do. 2) The op is not about your past actions, it’s about your enemies and this event as their opportunity 3) No matter how bad it gets, nothing is as bad as it seems. 4) Find the enemies at the very root of the op and find every creative problem-solving tool at your disposal to neutralize them. 5) People will throw their grandmother under the bus if they believe doing so will protect their reputation. They can become highly useful idiots in the op. Their power and access must be fiercely and quickly cut off. 6) Do not hesitate in cutting off anyone you suspect of disloyalty. 7) Build a team of close and loyal patriots and have one of them run a dedicated, well-resourced war room to put the rebellion to rest. 8) Use your patriots and regular meditation to buoy your psychological resistance to the op. 9) Your enemies aren’t as smart or hard working as you. Outsmart and outwork them as a way of life until the op is over. 10) Think about the op as if it were in the past. What are the difficult things you procrastinated on or avoided that would have stopped the op? Do them. 11) Never be ashamed of the op. Business as usual. Continue on with optimism and joy in your everyday work and its mission. 12) Download telegram and use it as your messaging application for all sensitive communications.
271
839
229
8,472
1,426,999
X 牛逼
Chinese 𝕏 is 牛逼
1
11
nickelily retweeted
提示词:Your advice for next 10 years. No explanations.
119
193
31
1,013
89,234
nickelily retweeted
我的 Gemini 给我的未来十年建议!大家可以试试问 Argon 这个问题:“Your advice for next 10 years. No explanations.” 看能否得到类似的回答?这个建议可以说相当通透了👀 1. 掌控核心瓶颈:能源、算力与稀缺物理资产。 2. 构建“人机双向共生”,而非与算法内卷。 3. 锤炼主体性(Agency)与审美品味,而非单纯的执行力。 4. 守护底层生理硬件:睡眠、体能与深度专注力。 5. 将“注意力主权”视为你最高杠杆的资产。 6. 设计在你离线时依然能产生复利的系统。 7. 深耕高信任度的线下真实人际网络。 8. 保持认知的流动性,让“反学习”与迭代的速度快于模型更新。 9. 精通提出正确问题的艺术。 10. 重仓那些永远无法被数字化的体验与情感连接。
11
14
3
111
10,577
有分歧才有机会
1
7
nickelily retweeted
说说最近一个极其重要的观察。 以抖音和b站为主平台的AI短剧和漫剧越来越风靡。和年初相比,AIGC生成的效率和质量进步只能用恐怖来形容。很多颜值博主说自己越来越颜值焦虑的根源是自己看到的AI漫剧里很多角色比自己和周围的人好看的多、精致的多,导致自己在现实里看别人的脸反而不自在、觉得有缺陷。 尤其是以夜神月&富江为代表作的AI漫改剧,我已经明显感觉到此赛道流量和热度是爆炸的程度。尽管不知道它未来流量和热度情况会如何,但至少现在我的判断是,此后的几个月,AI漫改剧对于社会的影响是巨大的,包括但不限于红果,还有黄果。 谈一个涉及心理学和社会学的命题。原来人们看AI觉得AI伪人感过强,现在AI慢慢生动化之后,人自己对自己的样貌开始失去自信,开始觉得自己和周围的人是伪人了。这是一件非常需要关注且有些恐怖的事情。 我查了一下,这叫面孔适应效应。视知觉研究发现,人脑对正常的脸的判断是由最近看到的大量面孔不断校准。实验里让人盯着拉宽的脸看几分钟,之后再看正常的脸就会觉得偏窄。长时间刷AI漫剧,等于每天几小时给大脑灌输标准脸:完美对称,没有毛孔,五官比例是动漫式的放大。参照系被悄悄挪走后,现实中正常的不对称、肤色不均、表情时的细微褶皱,就被识别成了错误。 所以大家的恐怖谷其实没有消失,只是移了位置。现在掉,进谷里的是真人。 这件事比以前的滤镜、明星、二次元更麻烦,而且严峻得多。 原因在于AI漫改剧刚好卡在一个以前不存在的位置。纯二次元画风足够假,大脑会把它归到另一个类别,不拿来和真人比;明星和修图虽然逼真,但数量有限,而且是静态的。 AI漫改剧是写实质感配上非人比例,会动、有表情、有剧情带来的情感投入,供给量几乎无限。大脑把它当成人脸来处理,却又是任何真人都达不到的版本。 这就是动物行为学里说的超常刺激:比天然刺激更强烈,反而压过了对天然刺激的偏好。 大家可以谈谈自己的意见,看我说的问题是否真实客观存在。
35
26
8
257
118,641
This live demo of four Microducks with Jensen Huang went completely viral, and we're ecstatic about it. But there's also a lot of confusion, so I want to clarify what Microduck can and cannot do, and what it will realistically be able to do in the future. I count 60M+ views on this demo alone, based on the few big posts I see on social media, and everything has been buzzing. Even my own simple emotional design post on a newly created Instagram account got 1M+ views. We are selling Microducks at an incredible rate, even faster than during its historic launch. I've been trying to answer all the questions I see on our socials: people are asking if X or Y is possible on the robots, and this post aims to answer most of them. First of all, the demo you've seen was not fully autonomous. We had one engineer piloting each Microduck. The motions on the robots were trained with reinforcement learning, using a sim-to-real pipeline: walking, standing up after a fall, sitting, doing a somersault, roller skating. All of those are policies trained this way. What the pilots did during the live event was send commands: directions (move forward, rotate), head orientation (look here), quacks, and special actions (do the somersault, sit down). If you buy a Microduck, will you get an autonomous companion right out of the box? The current answer is no. But there is no impossible technical barrier to get there. I'm not sure how much time we'll have to work on this before shipping, but since everything is open source, I fully expect the community to build it, and to build it fast. Now let's understand what's already done, what will be easy to add, and what's an open problem. What you'll be able to do out of the box, and you can already do it in simulation right now, is train new policies with reinforcement learning. You could teach your duck to walk differently from the default gait we provide, to do a weird dance, or to walk on stilts. The community has been crazy creative with this, even though they only have the simulator. Note that you don't need reinforcement learning when the movements are simple. For example, the early emotional design I did is mostly coded movements plus pre-generated sounds. A simple mental model: use reinforcement learning when balance is part of the movement. If you try to code a cinematic walk by naively putting one foot in front of the other, your robot will realistically never walk, because you need to account for the dynamics of the movement, the inaccuracies of the sensors, the friction with the ground, and potential perturbations from the outside, like being pushed. So training new policies is definitely doable, but you can't train a policy to get a genuine long-term companion in your home (well, technically you could, but that's a whole other range of policy: a research-grade problem, and the resulting neural network would probably be huge and never able to run on such a small compute). Instead, you need a few more building blocks: voice recognition, facial detection/recognition, localization, some simple navigation, and a behavior layer, which is less technical and harder to describe. As a rule of thumb, a prototype of any of these is within reach of anyone if they have a modern coding agent and if the computation doesn't run on the robot (meaning the robot can reach the internet or an external computer and use external APIs). A rough prototype for each of the above can be a chill weekend project. As always, going from the prototype to a high-quality implementation is another story, but still doable for most if you put in the time and test a lot. However, having all of this run locally, on the onboard chip, I would call a very hard task. You'd need some very clever engineering to get a good enough version of everything working together. But I can't say it's impossible, because I've seen some crazy efficient implementations of all of the above. And here we don't need a super high success rate: it's okay if the duck sometimes does something a bit random. So I wouldn't be surprised if folks managed a fully local, cool, simple version of this autonomous companion. For example, you don't need Whisper-grade open-vocabulary transcription: being able to recognize a few keywords would already be great for interaction. One of the fastest ways to turn this robot into an autonomous companion would be to stream Microduck's sensors (the camera, the depth sensor, the IMU and the joint positions) to a computer, and use a software stack like the one we use for the conversation app on Reachy Mini: a local VAD, then a cloud speech-to-text model, then a call to a frontier LLM that decides on actions. Actions could be quacking, looking at things/faces/a direction, walking to a given position, sitting down, playing an emotion, kicking or doing a somersault. A SLAM algorithm would run in parallel so the robot can localize itself in your home, and you would prompt the LLM for a lifelike behavior: moving to random places, sometimes following people, and having emotional responses that fit what you're saying and the duck's environment. Of course someone is going to build a Tamagotchi-like layer on top of this, and it will be insane. How much demand is there for a software suite that would do all of this? I for one think it would be super cool. A word about the humans. We had five engineers backstage, and they are pretty much the core team behind Microduck. Their impact per head is absolutely off the charts. Antoine Pirrone piloted Graphite, Augustin Crampette Lavender (the rebel one), Matthieu Lapeyre Sky, Coralie Deplane Cream, and Pierre Rouanet was stressing over the Wi-Fi. Give them a follow! (Antoine @antoinepirrone, Augustin on LinkedIn, Matthieu @matth_lapeyre, Coralie @Drenaya_, Pierre on LinkedIn). I get visibility because I communicate a lot, but I mainly work on Reachy Mini (you might want to check it out: it's a very cool robot, and it ships much faster than Microduck :) ). My contribution to Microduck has been minimal: I'm mostly a first power user. My main added value at GTC was having hands big enough to carry four Microducks below the stage. Where I take pride is that Antoine and Coralie are former students and former interns of mine. I'm therefore entitled to 3% credit on anything (good) they do, for life. If they do something bad, however, it was clearly the fault of their parents. About the super smart engineering that can be done on board: Antoine just released two very cool blog posts (links below). The first is on odometry, the robot estimating its own movement from its own internal sensors. The second is about SLAM, which allows long-term localization: it adds the depth sensor to the odometry to get a stable position over time. Think of it as recognizing known landmarks in your home to know where the robot is. The crazy part is that all of this runs on the robot. So this gives me hope that we'll be able to do a lot of cool stuff on board.
46
53
17
460
66,595
nickelily retweeted
iPhone Duo Review: 10,000 Folds Later...
8,540
17,618
3,428
194,048
58,917,780
nickelily retweeted
市場往往在你最痛苦的時候改變方向 你有沒有過這種經歷:一隻股票,你忍了很久沒敢買,終於忍不住買入進去了,結果就暴跌了;或者經歷連續大跌後終於絕望了,割肉了,賣出之後,市場就開始大幅反轉了。 有網友玩笑說,市場就是跟我過意不去,「主力莊家」就是看中了我手中這幾百塊錢。 其實,市場根本不知道你是誰。它只是在執行一個基本的物理法則:尋找流動性。 當所有人都覺得要大漲,滿倉押注的時候,買方的流動性已經枯竭,市場不需要什麼巨大利空,只要稍微跌一點,那些加了槓桿、追高的人就會恐慌踩踏,市場的流動性就出來了。 反之,當所有人都絕望、把股票清空的時候,市場上賣方流動性已經枯竭。這時候,只需要一點點買盤,價格就會直線拉升。這就是為什麼底部總是在極度恐慌中誕生。 下次當你感覺到極度恐慌、想要砍倉的時候,停下來想一想:多數人此時此刻在想什麼?
31
56
8
571
47,415
nickelily retweeted
JUST IN: A full highlight reel of all the stunts performed at the Tesla Roadster event
210
1,137
405
10,928
1,340,869
1011一周年,很多人死在了那天,幸存者砥砺前行。
1
9
Thanks @168X_Fortune and @168MrZ for having me back. Here is the summary in English TLDR, AI fundamentals are still getting stronger. But after the big run, I’m keeping positions light and waiting for a cleaner setup. The next major opportunity could be the assets that high rates have kept under pressure. Here’s how I’m thinking about it: 1. US10Y: volatility matters more than the exact yield. I’m less concerned about whether the 10-year yield is 5.3% or 5.5% than how violently it moves. Before adding meaningful risk, I want yields stabilizing or gradually declining, with MOVE—the bond market’s volatility index—coming down too. 2. Market breadth needs to resolve. Money has concentrated in AI winners while much of the market struggles. Breadth can improve through laggards catching up, or through winners correcting. Either outcome makes chasing semiconductors with maximum size less attractive here. 3. The catch-up basket: KRE, IWM, RKT and XLU. Regional banks, small caps, housing and utilities could respond sharply to easing rate pressure, helped by short covering. The condition matters: an orderly decline in yields supports that setup. A sudden collapse could signal recession instead. 4. Mid term election is not a risk to AI datacenter development. The key topics from Democrats are still inflation, Iran and gas price. These have higher priority than AI datacenter. However I see long term risk as the US people's sentiment towards AI and datacenter is quickly deteriorate except people who are working in the AI industry. 5. The oil issue extends beyond crude. Refinery capacity and refined-product supply deserve more attention. Having crude available doesn’t automatically solve gasoline or diesel shortages. Persistent pressure on those products can keep inflation expectations—and long-term yields—elevated. 6. AI revenue growth is the key validation. I’m skeptical of claims that OpenAI is already at $70B ARR. But if coding and enterprise growth stays steep, an $80–100B annualized revenue run rate by year-end is conceivable. That’s a possible outcome, not a confirmed number. Downstream revenue must justify upstream capex. 7. Google’s biggest opportunity is a personal AI agent. Google has distribution through Search, YouTube, Gmail, Maps and Android, existing monetization infrastructure, and its own hardware-to-model stack with low inference cost. I’d like to see it combine those advantages into a personal agent. Coding isn't its advantage. 8. The next AI re-rating needs evidence of returns. Personal agents, AI for Science, biotech and enterprise applications could broaden monetization beyond coding. Faster research, cheaper drug discovery and measurable productivity gains would help demonstrate that AI spending can generate attractive returns on invested capital. 9. AI IPOs could redistribute the premium. If OpenAI or Anthropic goes public, investors would gain direct access to another layer of AI monetization. Some capital could rotate away from semiconductor proxies. That could change where the AI premium sits even while chip fundamentals remain healthy. 10. I’m more selective within semiconductors. CPU still has catch-up potential as server architecture evolves. Optical has already run hard, and strong memory results may already be well understood. I’m treating semiconductor equipment more tactically. Timing matters increasingly after the sector’s major advance. 11. BTC is my first watch when the rate setup improves. If US10Y and MOVE genuinely top out, BTC looks like a strong expression of improving liquidity and risk appetite. Its technical structure matters too. I’d look at that strength alongside the catch-up potential in rate-sensitive equities. 12. TLT calls have become less attractive. High bond volatility makes options expensive. Even if TLT rises as yields decline, falling implied volatility can eat into call returns. For this setup, I prefer considering equities with underlying businesses that could benefit from easing rate pressure. 13. My main macro catalysts: FOMC, CPI and Treasury issuance. These can change expectations for policy, inflation and long-term bond supply. For the midterms, I care most about what follows: changes to fiscal policy and issuance that could alter the path of long-term yields. 14. Light positioning is a deliberate choice. I already traded the post-September FOMC rally aggressively. Now I’m waiting for better risk/reward. With leverage and options, timing matters enormously. A stable index can still hide severe drawdowns in individual stocks. 15. The next big trade could be the laggards catching up. I remain confident in AI’s fundamentals, especially as downstream monetization expands. But confidence doesn’t require maximum exposure at every price. If yields stabilize and MOVE falls, BTC and KRE/IWM/RKT/XLU could offer a compelling next setup. For now, I’m keeping risk light, watching the bond market, and waiting for the opportunity to become clearer.
AI Fundamental 仍强,但最舒服的主升段已经走过 3X Long Labubu @labubu_trader:接下来值得等的,不是继续追 AI Winner,而是 US10Y 稳定回落后的 Rotation 长端一旦见顶,资金可能转向 BTC,以及 KRE、IWM、RKT、XLU 这批被高利率压制最久的 Losers
17
28
4
184
84,397
触目惊心,存起来给孩子看,为什么不要碰毒品。
RAW REALITY: Several days ago, I posted shocking video of a young Asian woman going on a drug fueled bender in Seattle's Chinatown-ID. It drew national attention and triggered thousands of comments from people demanding change. But so far, not peep from socialist Mayor Katie Wilson or other elected officials in this city. That's because this is all happening on their watch and it's easier to ignore the problem. So here is the longer form version of this human tragedy. Send it to your lawmaker and tell them to do something about it before more lives are lost to the streets.
27
53
6
238
75,182
PUNK 2 BRICKS I built a tool yesterday. Upload a photo of your CryptoPunk, get back a brick bust of it: the build instructions, and the parts list you upload straight to the LEGO site to order everything. hs7j4yk4sz-boop.github.io/pu… Built with Claude Opus 5.5. Code is open, repo below.
72
54
37
474
94,482
好奇他们选人的标准是什么
辍学做 AI 有人兜底了。硅谷顶级风投 a16z,宣布自己办了所迷你学院,专门接收高中刚毕业、或大学读了一两年就想辍学,出来搞事的年轻人。 他们写了篇长文,认为现在 AI 时代,18 岁就能玩转 AI、做产品、甚至已经赚到钱的小孩,再去读那种传统的四年大学,纯属浪费时间。 所以他们在美国旧金山,搞了一个教育项目,让业内最顶级的 AI 老板和从业者,做你的老师,每人给 5 万美元以上算力额度,教你做产品、创业、研究,学校帮忙出钱和对接资源。 甚至有机会,还能去 Stripe、谷歌、OpenAI 这类公司实习。 目前计划 2027 年秋天开学,已开放全球报名,后面他们还想办两年制正规项目,像私立大学一样。
1
20
skolyoz o kadar kotu bir sey ki yillarca ugrastim ve ameliyat olmamak icin yapmadigim sey kalmadi resmen spor yaptim egzersiz yaptim masaja gittim yuzmeye bile gittim iyi gelebilecegini dusundugum ne varsa denedim sirf ameliyata gerek kalmasin diye
Skolyoz hastası kadın, demire asılı kaldığında omurgasında oluşan rahatlamayı gösterdi.
209
575
92
23,577
7,401,460
true,如果说隐私是zec的利器,为什么前几轮周期不是呢?
they are pumping the dead blockchains again, why must you all test fate in this way
1
18
WTF
He's baaaaack: "Somebody" just spent $100MM premium on 2 week AI Stock Calls for the Oct2 expiration: INTC Oct2nd 115c bot up to 3.65 20k $7.3MM MRVL Oct2nd 250c bot 11.00 3500x $3.85MM SNDK Oct2nd 1600c bot up to 97.60 4200x $41MM MU oct 2nd 1000c bot up to 44.00 10k $44MM
15
nickelily retweeted
She faked an overhead toss multiple times, pretended to restart her service motion, and then casually tapped an underhand drop shot over the net for a clean ace.🎾🤡
1
1
1
6,626