0xDQ (@DQbeta) profile picture
Sydney
Joined January 2010
Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)Photo by 0xDQ (@DQbeta)
为什么你花 2 万块买的旗舰智能电视,按遥控器还是卡得像十年前的安卓机? 昨天安全研究员 @log4j 把自家 3000 刀的三星电视 Root 了,扒出来的底层代码堪称年度草台班子大赏: 1. 电视后台每 500 毫秒截一次屏(连 HDMI 投屏看电脑桌面都截),打包发给近 400 家合作商。如果电视没麦克风,后台进程甚至会顺着蓝牙和回音壁去「借」一个麦克风继续听。 2. 最搞笑的来了:为什么菜单这么卡? 另一位研究员 @mike64_t 拆包发现,三星在电视那颗羸弱的 CPU 上,塞了一个叫 alex.tflite 的模型——246 万参数,2012 年古董级的 AlexNet。 这玩意每隔 2 秒抓一张画面跑推理,猜你在玩什么游戏。结果因为颜色通道写反了,它 99% 的时间什么都认不出来,纯烧 CPU。 3. 那如果你在设置里关掉「隐私追踪」呢? 代码逻辑是:先死命跑完 CPU 推理,准备上传时再看开关——发现你关了,就把结果扔进垃圾桶,2 秒后继续跑。 4. 结局更是重量级: 当号主用 OpenAI 逆向分析「自己花 3000 刀买回家的电视固件」时,OpenAI 反手一个 Cyber Abuse 把他号封了,顺带没收了剩下的 500 刀余额。 三星负责偷窥客厅,2012 年的 AI 负责烧干电视 CPU,OpenAI 负责保护三星的屎山代码不被你发现。
231
1,131
97
9,106
515,296
Replying to @engineers_feed
All mediocre writing. That includes nearly all school essays and emails, but not all books and movie scripts.
50
18
2
472
I tried this for few weeks. really helpful
Get your sleep & daytime energy and mood right: Get sunlight in your eyes as early in your day as possible. Especially on cloudy days. Extra time outside if you wake up late. Sunglasses off if you safely can. Eye glasses and contacts fine. Lights bright by day, sun-off at night.
人民不需要自由,这是最好的年代youtu.be/ZNFyceQQVNM via @YouTube
2
1
怎么用 AR 技术观看击剑比赛? 日本 Rhizomatiks 开发的“Points of a sword visualization system” 剑尖可视化系统,在 2019 世界击剑大赛上得到应用,实时合成剑尖视觉化图像,并生成运动轨迹 这套 AR 系统彻底改变了击剑、日本剑术的观看体验👏👍 #犀利而无用的知识
30
319
31
918
| ̄ ̄ ̄ ̄ ̄ ̄ ̄ ̄ ̄ | | WHY DO PYTHON | | VERSION ISSUES | | AFFECT MY | | ENTIRE LIFE | | ___________| (\__/) || (•ㅅ•) || /   づ
9
33
2
307
Data(+ML) on Kubernetes is a great workflow. Mesos was great first step, but Kubernetes makes *productionizing* ML/data workflows so much easier. As I have been saying since last 1 year+, ML/Data on Kubernetes is going to break the divide between data science/products.
One of the largest and fastest growing Internet companies (@Uber) is planning to migrate all their workloads (ML/batch/stateless/stateful/daemon) from a custom @ApacheMesos-based scheduler to @KubernetesIO. AWESOME read: eng.uber.com/peloton/
1
8
12
😂
This Post is from an account that no longer exists. Learn more
Pivotal Container Service has arrived! 🙌 Get on the fast path to enterprise-grade @Kubernetesio. Learn about the benefits of #PKS:bit.ly/2Bp5nTK
1
45
3
55
Top 7 Things You Need to Read About Project Harbor. xitter.gq/vmwopensource/status/9…
#ProjectHarbor had a landmark 2017. Here are the top 7 things you need to read about the #opensource project: bit.ly/2maCgO7
1
2
.@yfauser gives a deep dive on how #Kubernetes and @VMwareNSX work together in this #vBrownBag presentation: bit.ly/2yN1asP
7
13
东京铁塔夜景
1