Brain Dumps

发信人: OneFlyingPig (time), 信区: Military 标  题: Cao 在特斯拉任职期间向iCloud账户上传源代码 发信站: BBS 未名空间站 (Thu Mar 21 13:20:46 2019, 美东) 小鹏硅谷的自动驾驶团队又出事了，特斯拉起诉前 Autopilot 计算机视觉科学家、小 鹏自动驾驶感知负责人 Guangzhi Cao 窃取 Autopilot 源代码。 根据特斯拉的说法，Cao 在特斯拉任职期间是 AP 团队具备访问 AP 源代码权限的 40 人之一，从去年开始，他向自己的 iCloud 账户上传 AP 相关源代码的完整副本。 他转移了超过 30 万份与 AP 相关的文件和目录，在去年年底受聘小鹏后，他从电脑上 删除了 12 万份文件并断开了他的 iCloud 账户。 在特斯拉任职的最后一天，他清空了自己的浏览器记录。 2019 年 1 月 3 日，他突然向特斯拉递交辞呈，随后特斯拉发现他跳槽到小鹏汽车出 任感知负责人。 特斯拉认为，Cao 和小鹏没有合法权利使用特斯拉 Autopilot 的相关技术，这是特斯 拉五年多的研发成果和超过数亿美元投资的结晶。

To make a long story short, I'm currently enthusiastically experimenting with   Prometheus . As part of this I'm trying out Prometheus's support for 'black box' status and health checks, where you test services and so on from the outside (instead of the 'white box' approach of extracting health metrics from them directly). The Prometheus people don't seem to be too enthusiastic about black box metrics, so it's perhaps not surprising that   the official Prometheus blackbox exporter   is somewhat underdocumented and hard to understand. The three important components in setting up blackbox metrics are   targets ,   modules , and   probers . A   prober   is the low level mechanism for making a check, such as making a HTTP request or a TCP connection; the very limited set of probers is built into the code of the blackbox exporter (and Prometheus is probably unenthused about adding more). A   module   specifies a collection of parameters f...

上周五新西兰基督城清真寺枪击案的凶嫌之一、 28 岁的澳大利亚人 Brenton Tarrant ，在出发去行凶之前曾经在网上放出一份他自己写的 “ 宣言 ” ，标题叫 “The Great Replacement” 。 文件洋洋洒洒多达74页，据他自己在文中透露之前还写过一个240页的版本，后来一时冲动自责全部删除，这次发动袭击前的两个星期又重新动笔写了个简明的版本。 这个小小的细节，说明了他的内心曾经有过的挣扎和冲突。 宣言书的开头，引用了英国诗人迪伦·托马斯(Dylan Thomas)的名作《不要温和地走进那个良夜》(Do not go gentle into that good night)——也许你还有印象，这首诗曾经在《星际穿越》里出现过好几次。 以往类似的枪杀案，媒体往往要费尽九牛二虎之力，寻找各种蛛丝马迹，找到很多人采访，才能拼凑出凶手大致的情况。 如果凶手被当场击毙，那他的确切动机和行事逻辑，可能就永远都成为一个难解的谜。 但这次不一样，Brenton Tarrant 留下的这份宣言书用一问一答的形式，列举出80多个问题一一作答。 通过这些回答，他坦诚地交代了自己的背景、经历、思想演变的过程，还有之所以开枪屠杀的原因和目的。 除了问答，宣言书的另一半篇幅是他阐述自己对这个世界的偏激看法： 当前的欧洲，白人的生育率过低，同时又有大批移民定居； 在这两个因素的共同作用下，欧洲很快就会“变色”，非白人取代白人成为欧洲的人口主体，欧洲各国到了生死存亡的紧急关头，面临亡国灭种的危机。 澳大利亚和新西兰都属于他所认为的“欧洲”的范畴，因为这些国家的人都是欧洲白人的后代。 这个混杂着部分事实和受害妄想的“理论”，成为他一切行为的出发点。 所以在明知杀戮平民的行径有如纳粹的情况下，他仍然一意孤行。 他的目的是，希望这样做能吓退那些想移民到欧洲的人，更希望自己制造的冷血杀戮能“...

发信人: luren7788 (路人七七八八), 信区: SanFrancisco 标  题: 从IPO pipeline 看弯曲房市前景 发信站: BBS 未名空间站 (Fri Mar  8 13:57:02 2019, 美东) 今年貌似将是IPO丰收年。弯曲的几个大明星是 Uber, Lyft, Airbnb, Pinterest, Slack, Palantir, Robinhood。我们来看看这些公司目前的估值： Uber：120B Lyft：15B Airbnb: 30B Slack: 10B Palantir: 30B Robinhood: 5B Pinterest: 10B 加起来 $200B 以上。当然这些公司都很久了。融资好多轮。但是一般来说普通职员， 包括马工的股份，总共应该占10%左右。所以总值有 $20B。这些公司总部和大部分关键 员工都在弯曲。我们算75%吧，那就是 $15B 的财富。 拿了这么多钱干嘛？首先当然是买房。大部分员工都是无房族，都等了这么多年了，从 年轻小伙变成了中年大叔，头发都等白了。终于能买房了，这个高兴啊。 那我们假设其中2/3拿来进入房市，也就是$10B。能买多少房子呢？算两倍的Leverage ，入市的钱就是$20B。平均2米的房子，那就可以买一万栋。所以光是这几个公司今明 两年就会增加刚需“标准房”一万栋。 打开redfin看看弯曲市场上有没有一万栋2米左右的房子。

Lots of people make the mistake of thinking there's only two vectors you can go to improve performance, high or wide. High - throw hardware at the problem, on a single machine Wide - Add more machines There's a third direction you can go, I call it "going deep". Today's programs run on software stacks so high and so abstract that we're just now getting around to redeveloping (again for like the 3rd or 4th time) software that performs about as well as software we had around in the   1990s and early 2000s . Going deep means stripping away this nonsense and getting down closer to the metal, using smart algorithms, planning and working through a problem and seeing if you can size the solution to running on one machine as-is. Modern CPUs, memory and disk (especially SSDs) are unbelievably fast compared to what we had at the turn of the millenium, yet we treat them like they're spare capacity to soak up even lazier abstractions. We keep thinking that co...

Prometheus is   architected  for   reliability   of alerting, how do you set it up? For a setup that can gracefully handle any machine failing, we'll need to run two Prometheus servers and two Alertmanagers. First we'll run the Alertmanagers on different machines, and setup a mesh between them: # On a machine named "am-1": wget https://github.com/prometheus/alertmanager/releases/download/v0.15.3/alertmanager-0.15.3.linux-amd64.tar.gz tar -xzf alertmanager-*.linux-amd64.tar.gz cd alertmanager-* ./alertmanager --cluster.peer=am-2:9094 # On a machine named "am-2": wget https://github.com/prometheus/alertmanager/releases/download/v0.15.3/alertmanager-0.15.3.linux-amd64.tar.gz tar -xzf alertmanager-*.linux-amd64.tar.gz cd alertmanager-* ./alertmanager --cluster.peer=am-1:9094 To verify that the Alertmanager mesh is working correctly, create a silence in one Alertmanager. If it shows up in the other, then all is well. Next we configure the Prometheus ...

ederal Aviation Administration managers pushed its engineers to delegate wide responsibility for assessing the safety of the 737 MAX to Boeing itself. But safety engineers familiar with the documents shared details that show the analysis included crucial flaws. Share story   By   Dominic Gates   Seattle Times aerospace reporter As Boeing hustled in 2015 to catch up to Airbus and certify its new 737 MAX, Federal Aviation Administration (FAA) managers pushed the agency’s safety engineers to delegate safety assessments to Boeing itself, and to speedily approve the resulting analysis. But the original safety analysis that Boeing delivered to the FAA for a new flight control system on the MAX — a report used to certify the plane as safe to fly — had several crucial flaws. That flight control system, called MCAS (Maneuvering Characteristics Augmentation System), is now under scrutiny after two crashes of the jet in less than five months resulted i...