Structure News: Please update your iPhone as soon as you finish reading this email

Your weekly tech news roundup, with a little bit of Structure. Where We Patched All Of Our iOS Devices Real Fast August 26th, 2016 / by Tom Krazit This week, we'll talk about the efforts of Intel rivals to dent its server chip supremacy, the discovery of a scary exploit tied to iOS flaws, and how machine learning is improving security. BIG PICTURE It might have missed out on mobile, but Intel has ably handled the other great transition of our times: the rise of cloud computing. After throwing out some long-held orthodoxies and realizing that datacenter customers needed more more flexibility and custom hooks than system administrators buying rack servers in the mid-2000s, Intel's chips are still the choice of public cloud vendors. Several other companies would like a piece of that business. This week at Hot Chips, ARM, AMD, and IBM showed off designs for server chips that they hope can shake up the market for public cloud servers and, increasingly, artificial intelligence servers. ARM backers have been pitching servers based on its technology for years with little to show for it, but The Next Platform reports that this week the company showed off the results of a joint project between Fujitsu and ARM to add improved vector processing to the ARM core. Fujitsu is building an ARM supercomputer for the Japanese government that will be watched closely for evidence that ARM chips are ready for prime time, and making those chips more competitive against the vector processing capabilities of Intel's Xeon chips was an important step. At the same show, IBM showed off the upcoming Power9 processors, which are being designed specifically to work with Nvidia's graphics processing units, used by a lot of companies researching artificial intelligence and deep learning. AMD reminded everyone it was still in business by announcing plans to release a 32-core server chip next year, as reported by The Wall Street Journal. Diane Bryant's famous Super Seven server chip customers (Alibaba, Amazon, Baidu, Facebook, Google, Microsoft, and Tencent) aren't going to jump ship any time soon. The projects shown off this week won't be in production for some time, and those companies probably aren't going to risk their livelihoods on untested first-generation hardware. But it's no secret that there are movements within those companies that want to wring some pricing concessions out of Intel by supporting other chip makers or by designing their own chips based on ARM cores, taking a cue from Apple's decision to design its own chips for the iPhone. We've long heard rumors that Google and Amazon have messed around with designing its own chips. When will someone take the plunge? STRUCTURE NEWS WHERE MACHINE LEARNING HELPS SECURITY, AND WHERE IT DOESN'T We're going to have several sessions at Structure Security dedicated to examining how advances in machine learning can help make us more secure, and I caught up this week with Kevin Mahaffey (pictured), Co-Founder and CTO of Lookout Security, for a few more details. Machine learning is essential for protecting frequently targeted companies, such as multinational banks, against zero-day threats and could also help secure the internet of things, but most companies are probably better off just patching their systems on a regular basis. Structure Security will take place September 27th and 28th at the Golden Gate Club in San Francisco. If you're into machine learning, Kevin will be on a panel with Carson Sweet of Cloud Passage and Mark Terenzoni of SQRRL, and we'll also have Stuart McClure, CEO of Cylance in a different session. More information on the show is available here. INDUSTRY NEWS A SERIOUS ATTACK ON THE IPHONE WAS JUST SEEN IN USE FOR THE FIRST TIME Apple released a software patch for iOS Thursday in conjunction with a report from Citizen Lab and Lookout detailing how three iOS vulnerabilities were used to create an exploit targeting a human rights activist. The Verge reports that the exploit, which could be used to remotely jailbreak an iPhone, has been traced to the shadowy NSO Group. LINUS TORVALDS REFLECTS ON 25 YEARS OF LINUX It's hard to imagine what life without Linux might look like in 2016. Linus Torvalds, the legendary creator of an operating system that even Microsoft is using in its Azure cloud these days, took the stage at Linuxcon this week to look back on the project, as recounted by Linux.com. CHINA SETS NEW TONE IN DRAFTING CYBERSECURITY RULES Long skeptical of the aims of U.S. technology companies operating in its country, China appears to have shifted its thinking. The Wall Street Journal reports that China has invited Intel, Cisco, Microsoft, and IBM to help it write new cybersecurity standards, although it's far from clear whether or not those companies will have any influence in assuaging fears that China might step up its surveillance activities through the new guidelines. WHY TWITTER WAS THE PLATFORM OF CHOICE FOR RIPPING APART THE THE NSA DUMP Twitter might be a dumpster-fire trainwreck most days, but it performed a very valuable service for information security professionals scrambling to make sense of the Shadow Brokers leak we discussed last week. Motherboard has an interesting post showing how the service might just part of the fix for the security industry's information-sharing problems, bemoaned by several Structure Security advisers over the last few months. FACEBOOK OPEN-SOURCES A.I. SOFTWARE FOR SEGMENTING OBJECTS IN IMAGES Three tools developed by Facebook for artificial intelligence research -- DeepMask, SharpMask, and MultiPathNet -- were released this week by the company's AI lab, according to Venturebeat. The tools help Facebook break down images into individual segments that it can compare against other images to recognize faces, locations or objects, which probably means we're about to see some sort of creepy face-recognition app pop up. THE IBRAIN IS HERE, AND IT'S ALREADY INSIDE YOUR PHONE Apple sure seems sensitive about the notion that it lags behind rival mega-companies when it comes to artificial intelligence research. In this rather-friendly profile from Backchannel, Apple executives reveal a few tidbits about how it has used machine learning to improve Siri and throw a little shade at competitors "whose primary motivation is publishing," in the words of one Apple executive. QUOTE OF THE WEEK They are a really sophisticated threat actor and the software they have reflects that. They are incredibly committed to stealth." Mike Murray, VP at Lookout, on the NSO Group, as told to Wired Click here to unsubscribe from this list. Our mailing address is: 405 El Camino Real, #215 Menlo Park, CA 94025 Copyright (C) 2016 StructureSeries All rights reserved.