🔋𝓣𝓮𝓼𝓵𝓪4680🔋

3.3K posts

🔋𝓣𝓮𝓼𝓵𝓪4680🔋 banner
🔋𝓣𝓮𝓼𝓵𝓪4680🔋

🔋𝓣𝓮𝓼𝓵𝓪4680🔋

@Roystable

Tesla owner since 2021, Husband, Father, Stand with Elon.

Katılım Eylül 2015
790 Takip Edilen266 Takipçiler
Sabitlenmiş Tweet
🔋𝓣𝓮𝓼𝓵𝓪4680🔋
I promise: I'll buy a tesla when the stock hits $600 (PT of @Garyblack00 ),M3 or MY would be OK,before that I should deal with my Toyota.
English
3
0
13
0
MayHe
MayHe@An_maymayhe·
英伟达老板,黄仁勋年轻的时候也是好帅👍,有没有觉得看起来和刘德华年轻时候很像。都是大帅哥!
MayHe tweet media
中文
165
34
460
87.2K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
SemiAnalysis
SemiAnalysis@SemiAnalysis_·
SpaceX just filed their S1. SemiAnalysis research is cited! (1/5) 🧵
SemiAnalysis tweet media
English
9
24
414
59.1K
✧ 𝕀𝔸𝕄𝔸𝕀 ✧
谷歌新模型发布了,不出意外,依然强大得近乎残酷。过去一年,Google 的 token 月使用量同比去年增长了 7 倍。站在人类文明的尺度看,我从不怀疑 AI 会极大提高生产率,甚至在未来某个阶段将全球 GDP 推高数倍。 但也正因为如此,我越来越被另一个问题困扰。 我真正担心的,并不是“人类会不会被 AI 替代”这种表层问题。相比之下,更让我不安的是: 如果智能一旦被点燃,就能以如此恐怖的速度自我增强、扩散、改造经济和能源系统,那为什么宇宙依然如此空旷?为什么我们看不到任何其他文明留下的清晰痕迹? AI 的迅速进化,反而让费米悖论变得更加刺眼。 过去我们可以说,也许生命罕见,也许智慧罕见,也许技术文明罕见。但今天,当我们亲眼看到智能可以被工程化、规模化、token 化,并且以指数速度渗透进人类社会的每一个角落时,一个更冷峻的问题浮现出来: 如果一个文明真的跨过了 AI 这道门槛,它理论上不应该很快变成一种宇宙级现象吗? 它应该更快地发现新物理,更快地掌握能源,更快地自动化生产,更快地制造探测器、殖民轨道空间,甚至向恒星尺度扩张。哪怕不变成银河帝国,至少也应该在宇宙中留下某种可观测的工程痕迹。 可我们看到的,是沉默。 这让我开始怀疑,所谓“大过滤器”也许并不在智能诞生之前,而是在智能爆发之后。真正困难的不是产生智慧,而是一个文明能否在获得递归增强的技术能力之后,依然保持足够的协调、克制和长期主义。 也许宇宙并不缺少火种。 它缺少的是那些在掌握恒星级火焰之后,仍然没有把自己烧毁的文明。
中文
19
7
122
23.2K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
Sawyer Merritt
Sawyer Merritt@SawyerMerritt·
NYT in new article: "Early reviews of the @Tesla Semi suggest that it could be a much-needed hit for the company. And it could shake up the staid business of truck manufacturing in the same way that Tesla’s cars upended the auto industry. Demand for the Semi appears strong. California trucking firms have asked the state government for subsidies to help them buy more than 1,200 Tesla trucks. That’s more than all the applications for other electric trucks since the state’s incentive program began in 2019." Ivan Torres, a driver for Nevoya, a San Francisco-based trucking company, is a big fan of the Semi. He said the Semi was quieter than diesel rigs. And he can run the air-conditioning while waiting to drop off a load. That is sometimes not possible with diesel trucks because California limits how long those vehicles can idle in residential areas or near schools and hospitals to minimize pollution. It will also be cheeper to operate. Diesel prices have risen around 50% in the last couple months, making battery power more attractive. Electricity is much cheaper per mile, even in California, where utility rates are relatively high.
Sawyer Merritt tweet media
English
49
133
1.2K
208.3K
DJ
DJ@congressdj·
@SawyerMerritt @Starlink So what’s stopping the user from going to any of these sites using Starlink, directly?
GIF
English
5
3
83
1.3K
Sawyer Merritt
Sawyer Merritt@SawyerMerritt·
Delta Airlines CEO in new interview on why they chose Amazon Leo over @Starlink: “Amazon brings a lot more than just satellite technology. They bring great retailing capability and Amazon Prime and video gaming technologies, which Starlink does not have. I think the opportunities, in terms of the improved bandwidth with a much lower price point than what we’ve ever seen from Starlink, will make a big difference. Of course, we expect Starlink will be warning people that we’re going to go with an inferior product. But I’m not too worried about partnering with Amazon.” bloomberg.com/news/articles/…
English
577
47
1.2K
779.8K
WangNextDoor
WangNextDoor@WangNextDoor2·
5月18日,西安电子科技大学段宝岩院士带领的“逐日工程”团队迎来重大科研进展,成功突破空间太阳能电站与微波无线传能多项关键核心技术,自研完成多动目标微波无线传能地面验证系统,推动我国相关技术正式迈向工程化应用阶段。 空间太阳能电站如同太空微波充电桩,可摆脱卫星供电局限,搭建太空无线能源补给网络。团队优化推出分布式欧米伽设计方案,攻克一对多精准无线传能难题,可同时为多个太空飞行器、地面移动设备供电。 实测成果数据优异:百米级距离实现1180瓦功率输出,直流-直流传输效率20.8%,波束收集效率88.0%;无人机在时速30公里、相距30米场景下,稳定接收143瓦电能。 项目同步升级光电转换效能,实现天线集成轻量化。经权威评审,项目整体技术达国际领先水平,航天国防与民用产业化应用潜力巨大。 我表示观望态度,按理说有这技术那不是今后电动车充电也都可以不用充电枪了?成本如何,体积大小如何,工业化程度如何等等。。。国内的一些大学的测试数据还有待后续观察,产业化能力也有待后续跟进。
WangNextDoor tweet media
中文
3
1
4
2K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
海伦子Hellen
海伦子Hellen@peng_hellen·
星舰的猛禽发动机,采用的是全流量分级燃烧循环,有两个预燃室同时在工作,一个主要燃烧液氧,一个主要燃烧甲烷,两股力量产生的高温高压气体同时驱动涡轮泵,最后一起汇入到主燃烧室里。 而传统的火箭发动机,要么只有一个预燃室驱动,要么不是全流量通过,所以涡轮泵的功率自然更小。 这也是为什么单个三代猛禽的燃烧室压力超过了330bar,属于断层式领先,更高的燃烧室压力意味着同样大小的发动机喷嘴,可以喷发出更大的推力。 每一台三代猛禽的推力是280吨,助推器上33台,也就是说起飞推力变成了9240吨,而二代星舰是7600吨,当年载人登月的土星五号只有3400吨。 猛禽的厉害之处还体现在,它不是一个只能造几台原型机的矜贵产品,而是可以量产且反复使用的火箭发动机。 本期视频讲述了:有效观看星舰发射你需要的一些信息储备 youtu.be/v10NHFmg6Lc?si… via @YouTube
YouTube video
YouTube
中文
31
16
191
29.8K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
Sawyer Merritt
Sawyer Merritt@SawyerMerritt·
Tesla has just increased Model Y prices in the U.S. for the first time in two years. New prices: • Model Y Premium RWD: $45,990 ($1,000 increase) • Model Y AWD: $49,990 ($1,000 increase) • Model Y Performance: $57,990 ($500 increase) No price change for base Model Y RWD ($39,990) or Model Y Y AWD ($41,990).
Sawyer Merritt tweet mediaSawyer Merritt tweet mediaSawyer Merritt tweet mediaSawyer Merritt tweet media
English
157
318
3K
424.9K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
ImNotHarsh | 📈💸
ImNotHarsh | 📈💸@imnotharsh·
$INTC just set a CPU World Record with Core i9 processor. Chinese overclocker wytiwx smashed a new world record by pushing an Intel Core i9-14900KF past 9.2 GHz on an ASUS ROG Maximus Z790 Apex board. Extreme cooling, insane volts, and pure silicon lottery. Peak desktop overclocking. 14th-gen Intel still isn’t done yet.
Intel Technology@IntelTech

New CPU frequency world record! Extreme overclocker wytiwx pushed an #IntelCore i9 14900KF over 9.2GHz on an @ASUS_ROG Maximus Z790 Apex motherboard. More details on HWBot: ms.spr.ly/6011vR7DF

English
3
19
121
13.9K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
ImNotHarsh | 📈💸
ImNotHarsh | 📈💸@imnotharsh·
BREAKING: $NVDA May leverage $INTC Intel Foundry's EMIB-T advanced packaging technology for 4-chip flagship Rubin Ultra Per UBS analyst note, $NVDA Nvidia is reportedly considering two SKUs for its upcoming Rubin Ultra AI GPU: 1. a standard 2-chip version and, 2. higher-end 4-chip flagship that may leverage Intel’s EMIB-T advanced packaging technology. This follows confirmed supply-chain challenges where Nvidia scaled back an ambitious single-package 4-die design due to $TSM CoWoS-L warpage, yield, and capacity constraints. The move highlights Nvidia's efforts to diversify packaging options beyond heavy TSM reliance while maintaining high HBM4E memory targets for 2027 production. While the Intel collaboration remains unconfirmed speculation, it aligns with broader industry trends and Intel Foundry’s push into high-end AI assembly. Rubin base chips are still expected late 2026, with Ultra following in 2027.
ImNotHarsh | 📈💸 tweet media
English
8
45
307
26K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
Sawyer Merritt
Sawyer Merritt@SawyerMerritt·
Panasonic has raised its EV battery outlook due to an increase in Tesla sales. "Demand is expected to exceed full-year 2026 levels, supported by strategic customer's recent recovery in market share, despite significant revisions to their model lineups," the company said. investors.com/news/tesla-sal…
English
38
180
1.8K
223.1K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
Inside the conflict
Inside the conflict@InsidConflict·
Elon in Beijing doing full "Kekius Maximus at a state dinner" mode Tim Cook standing there like: "Bro this is why we can't have nice things... but also please buy more iPhones in China" Elon: thumbs up, yap mode 100, zero chill Only Elon can turn a fancy dinner into a live meme. Legend.
English
1
2
11
12.4K
🔋𝓣𝓮𝓼𝓵𝓪4680🔋 retweetledi
Macro_Lin | 市场观察员
聊一下Intel的EMIB-T,这个技术可能在未来两年改变先进封装的竞争格局。 先说背景。AI加速器的封装面积越来越大,一颗芯片装不下的计算和内存,需要把多颗die封装在一起协同工作。当封装面积超过单次光刻曝光的极限,就需要用”局部硅桥”来连接相邻的die。这个思路,TSMC和Intel各自走出了不同的实现路径。 TSMC的CoWoS-L,L代表LSI(Local Silicon Interconnect),用局部硅桥做相邻die之间的高密度互连,全局布线靠InFO RDL扇出层完成,整个工艺在圆形晶圆载板上加工。Intel的EMIB,同样是局部硅桥,但嵌入的是矩形有机基板,全局布线靠基板上的金属层完成。两者底层思路相近,但制造平台和全局布线的密度存在本质差异。 早期EMIB有一个结构性短板。硅桥里没有垂直通孔,电源从基板底部送不穿桥,只能绕路。偏偏最需要电的PHY电路就坐在桥的正上方(信号线越短越好,所以PHY放在die最边缘),电源反而要从die中心横向传过来。路径长,压降大,瞬态响应差。HBM3e时代还能应对,到HBM4接口宽度翻倍、pin速率推到6.4Gbps以上,这条供电路径就成了瓶颈。 EMIB-T就是解这个问题的。T代表TSV,在桥里打铜通孔,电源从基板直接垂直穿过桥送到上面的PHY。同时桥里集成MIM电容和接地铜网格,就地稳压滤噪。供电路径从横向绕行变成垂直直连。 这个改动释放了两层约束。第一层是供电能力,可以支撑HBM4/4e级别的功耗需求。第二层是封装面积上限,Intel目标是2026年做到120×120mm封装、集成8倍光罩面积的硅,2028年推进到120×180mm、超过12倍光罩。 跟CoWoS-L的竞争,核心差异在制造平台。CoWoS-L在圆形晶圆载板上完成全部工艺,面积越大,晶圆边缘的浪费越大,RDL层的均匀性和翘曲控制难度也非线性上升。EMIB-T在矩形有机基板上多放桥,每个桥的工艺独立重复,整体良率虽然也随桥数量指数衰减,但衰减速率稳定可控。封装越大,EMIB-T在成本上的优势越明显,良率的可控性也越高。反过来说,CoWoS-L的InFO RDL层提供了更高密度的全局布线和电源分配能力,加上NVIDIA等客户的设计生态已经深度适配CoWoS工艺,迁移成本极高,这是CoWoS短期内不可替代的护城河。 目前的状态是,标准EMIB量产良率90%,Intel表示已接近传统FCBGA封装水平。EMIB-T验证良率也到了90%,但这是小批量数据,量产目标98%。郭明錤的评价很准确,从90%到98%比从零到90%难得多。2026年EMIB-T进产线验证,2027年可能出现第一批产品(Jaguar Shores),Google TPU等外部客户也在评估EMIB-T方案,2028年才是真正的放量年。当然,Intel近几年的产品节奏有过多次延期,这个时间线能否如期兑现本身就是一个需要追踪的变量。 对TSMC的影响,我的判断是分流,不是替代。NVIDIA跟TSMC的全栈绑定短期内动不了,但Google、Meta这些做自研AI ASIC的客户,以及Apple等寻求封装供应链多元化的厂商,封装面积需求大、成本敏感、又有分散供应链的动力,EMIB-T给了它们一个真实可用的备选。Bernstein估算,EMIB平台封装每颗芯片成本在几百美元量级,而Rubin级别加速器上CoWoS-L封装成本约900到1000美元。即便只从边缘蚕食,打破CoWoS垄断定价权这件事本身,对整个产业链的影响也远超实际转移的份额。 SK hynix这两天也传出在跟Intel合作研发EMIB封装。HBM供应商主动拥抱EMIB生态,说明连它们自己都不想被CoWoS产能卡住出货节奏。 Intel的EMIB-T,正在把先进封装的竞争从”谁的技术更好”转向”谁能在超大面积上同时控住成本和良率”。从目前的市场叙事来看,EMIB-T的结构性优势应该还没被充分讨论。
Macro_Lin | 市场观察员 tweet media
中文
69
70
353
46.4K
KK.aWSB
KK.aWSB@KKaWSB·
当有人给你说FSD自动驾驶已经在所有方面超过人类驾驶员时,你就给他看这个视频:
中文
630
25
544
705.7K