Virtual Reality(虚拟现实技术)A new creative medium where the default state（默认状态/缺省状态）is beliefIn the VR community, “presence” is a term of art. It‟s the idea that once VR reaches a certain quality level（质量等级，标准）your brain is actually tricked （欺骗/哄骗）— at the lowest（至少/最低）, most primal level （最根本层面上）— into believing that what you see in front of you is reality. Studies show that even if you rationally（理性的）believe you‟re not truly standing at the edge of a steep cliff（陡峭的悬崖）, and even if you try with all your might （力量，能力）to jump, your legs will buckle（变弯曲）. Your low-level lizard brain （蜥蜴脑(lizard brain)：指的是人类负责维系生命功能动作的脑部(action brain)，是掌管吃喝拉撒睡、性欲、愤怒、恐惧、趋吉避凶的求生本能或动物性直觉）won‟t let you do it.With presence, your brain goes from feeling like you have a head set on to feeling like you‟re immersed in （全神贯注于）a different world.Computer enthusiasts（狂热者）and science fiction writers have dreamed about VR for decades. But earlier attempts to develop it, especially in the 1990s, were disappointing. It turnsout （原来/表明）the technology wasn‟t ready yet. What‟s happening now —because of Moore‟s Law（摩尔定律）, and also the rapid improvement of processors, screens, and accelerometers, driven by the smartphone boom — is that VR is finally ready to go mainstream（成为主流）.Once VR achieves presence, we start to believe.We use the phrase（短语/词语）“suspension of disbelief” about the experience of watching TV or movies. This implies that our default state watching TV and movies is disbelief. We start to believe only when we become sufficiently immersed（重复专注）.With VR, the situation is reversed: our brains believe, by default, that what we see is real. The risk isn‟t that it‟s boring but that it‟s overwhelmingly intense. We need to suspend（暂停）belief and remind ourselves that what we think we‟re experiencing isn‟t real.As Chris Milk, an early VR pioneer（先驱/开拓者）, says: You read a book; your brain reads letters printed in ink on paper and transforms that into a world. You watch a movie; you‟re seeing imagery（意向/影像）inside of a rectangle（矩形）while you‟re sitting inside a room, and your brain translates that into a world. And you connect to this even though youknow it‟s not real, but because you‟re in the habit of suspending disbelief.With virtual reality, you‟re essentially（本质上/根本上）hacking the visual-audio system（声音可视化系统）of your brain and feeding it a set of stimuli that‟s close enough to the stimuli it expects that it sees it as truth. Instead of suspending your disbelief, you actually have to remind yourself not to believe.This has implications（启示/含义）for the kinds of software that will succeed in VR. For example, a popular video game like Call of Duty （使命召唤）ported to VR would be frightening and disorienting（使迷惑/使分不清方向）for most people.What will likely succeed instead are relatively simple experiences. Some examples: go back in time and walk around ancient Rome(古罗马); overcome your fear of heights（恐高）by climbing skyscrapers; execute precision moves as you train to safely land planes; return to places you “3D photographed”（三维成像）on your last vacation; have a picnic on a sunny afternoon with a long-lost friend; build trust with virtual work colleagues in a way that today you can only do in person.These experiences will be dreamt up（虚构/凭空想出）by“experience makers” — the VR version of filmmakers（制片人）. The next few decades of VR will be similar to the first few decades of film. Filmmakers had no idea what worked and what didn‟t（什么能做，什么不能做）: how to write, how to shoot, how to edit, etc. After decades of experiments they established t he grammar of film. We‟re about to enter a similar period of exploration with VR.There will be great games made in VR, and gaming will probably dominate the VR narrative（故事）for the next few years. But longer term（长期）, we won‟t think of games as essential to（必不可少的）the medium. The original TV shows were newscasts（新闻广播）and game shows（娱乐节目）, but today we think of TV screens as content-agnostic（不可知性）input-output devices（输入输出设备）.VR will be the ultimate input-output device. Some people call VR “the last medium” because any subsequent（随后的/后面的）medium can be invented inside of VR, using software alone. Looking back（回顾往事）, the movie and TV screens we use today will be seen as an intermediate step between the invention of electricity and the invention of VR. Kids will think it‟s funny that their ancestors used to stare at glowing rectangles hoping to suspend disbelief.What comes after the consumerization(消费化) of ITA few years ago a bunch of consumer internet companies had an “Oh, shit” moment around mobile. They saw all their traffic coming in from mobile rather than the web. They redid （重做）everything to take advantage of mobile. Many still aren‟t done doing so (and some have yet to make the shift).Now something similar is happening in the enterprise world, whic h is going to have its own “Oh shit” moment. But this is not just about bringing mobile to the enterprise. It‟s about much more than that…Think about what‟s possible with smartphones today. The form and sophisticated（复杂的/精致的）functionality of smartphones has led to businesses we couldn‟t ever have imagined (ridesharing（拼车）, to name just one example). This was made possible partly because of the “sensorification” of the landscape, coupled with mobile and a friendly UI（用户界面（user interface））.That same sensorification needs to move into the enterprise.Enterprise UI is woefully（不幸的）behind. All those well-understood motions（提议）that have taken hold （生效/产生影响）from our everyday smartphone behaviors — pinch （捏）, zoom（缩放/变焦）, swipe, tap, speak, even just movingstuff around with our fingers —have yet to take hold in the enterprise. The user interface has always been an afterthought （事后想法/后来添加的东西）, the last thing one did after building a database. That is changing now.So where do the sensors come in? Think about what happened on mobile. We don‟t have a lot of screen real estate （屏幕空间）or functionality available there; the sensors are essentially used as “shortcuts” to information and actions the user doesn‟t have to perform. Instead of having to manually（手动的）type location in to our phones every time, GPS simply informs our phones that …you are here‟ so it can automatically load that context.For enterprise, the value of the sensors is in being a shortcut for the user interface, potentially even replacing typing so we can concentrate on the easy, fun, creative things.–Scott WeissMachine Learning + Big Data= Predictive analytics （预测分析）(And where do Hadoop（大数据分析）and Spark come in?)Here at a16z, we treat “big data” and “machine learning” as connected activities. People have been talking about the need for more …analysis‟ and insight（洞察力/洞悉）in big data, which is obviously important, because we‟ve been in the …collection‟ phase with big data until now. But the innovation in the big data world that I‟m most excited about is the …prediction‟ phase —the ability to process the information we‟ve collected, learn patterns（学习模式）, and predict unknowns based on what we‟ve already seen.Machine learning is to big data as human learning is to life experience: We interpolate（插入/植入）and extrapolate（推出）from past experiences to deal with unfamiliar situations. Machine learning with big data will duplicate this behavior, at massive scales.Where business intelligence（商业智能/商务智能）before was about past aggregates（总量/共计）(“How many red shoes have we sold in Kentucky（肯塔基州）?”), it will now demand predictive insights (“How many red shoes will we sell in Kentucky?”). An important implication of this is that machine learning will not be an activity in and of itself … it will be a pro perty of every application. There won‟t be a standalone（独立的）function, “Hey, let‟s use that tool to predict.”Take Sales force（销售人员/销售队伍）for example. Right now it just presents data, and the human user has to draw her or his predictive insights in their heads. Yet most of us have been trained by Google, which uses information from millions of variables based on ours and others‟ usage to tailor our user experience … why shouldn‟t we expect the same here? Enterprise applications —in every use case imaginable —should and will become inherently（内在的/固有的）more intelligent as the machine implicitly（暗中的/含蓄的）（learns patterns in the data and derives insights. It will be like having an intelligent, experienced human assistant in everything we do.The key here is in more automated apps where big data drives what the application does, and with no user intervention （介入/干涉）. (My colleague Frank Chen calls this the “big data inside” architecture for apps).But all of this forces, and benefits from, innovation at the infrastructure （基础设施建设/公共设施建设）level.Big Data needs Big Compute: Where Hadoop and Spark fit in the pictureThink of big data and machine learning as three steps (and phases of companies that have come out of this space): collect,analyze, and predict. These steps have been disconnected until now, because we‟ve been building the ecosystem（生态系统）from the bottom up （从底向上）— experimenting（实验）with various architectural and tool choices —and building a set of practices around that.The early Hadoop stack is an example of collecting and storing big data. It allows easier data processing（数据处理）across a large cluster of cheap commodity servers. But Hadoop MapReduce（并行编程）is a batch-oriented （面向批量）system, and doesn‟t lend itself well towards interactive applications（交互应用）; real-time operations（实时操作/实际运作）like stream processing; and other, more sophisticated computations.For predictive analytics, we need an infrastructure that‟s much more responsive to human-scale interactivity: What‟s happening today that may influence what happens tomorrow? A lot of iteration needs to occur on a continual basis for the system to get smart, for the machine to “learn” —explore the data, visualize it, build a model, ask a question, an answer comes back, bring in other data, and repeat the process.The more real-time and granular we can get, the moreresponsive, and more competitive, we can be.Compare this to the old world of “small-data” business intelligence, where it was sufficient to have a small application engine that sat on top of a database. Now, we‟re processing a thousand times more data, so to keep up the speed at that scale, we need a data engine that‟s in-memory and parallel. And for big data to unlock the value of machine learning, we‟re deploying it at the application layer. Which means “big data” needs “big compute”.This is where Apache Spark comes in. Because it‟s an in-memory, big-compute part of the stack, it‟s a hundred times faster than Hadoop MapReduce. It also offers interactivity since it‟s not limited to the batch model. Spark runs everywhere (including Hadoop), and turns the big data processing environment into a real-time data capture and analytics environment.* * *We‟ve invested in every level of the big data/big computeecosystem, and this remains an exciting, active space for innovation. Because big data computing is no longer the sole province of government agencies and big companies. Even though the early applications tend to show up in industries where data scientists have typically worked, machine learning as a property of all applications — especially when coupled with an accessible user interface — is democratizing who, what, and where this kind of real-time computing and learning can happen … and what great new companies can be built on top of it.My belief is every application will be re-constituted to take advantage of this trend. And thanks to big data and big compute innovations, we finally have the ingredients to really make this happen. We‟re at the threshold of a significant acceleration in machine intelligence that can benefit businesses and society at large.–Peter LevinedefinitionsBig Data is the collection of massive amounts of information, whether unstructured or structured.Big Compute is the large-scale (often parallel) processing power required to extract value from Big Data.Machine Learning is a branch of Computer Science that, instead of applying high-level algorithms to solve problems in explicit, imperative logic, applies low-level algorithms to discover patterns implicit in the data. (Think about this like how the human brain learns from life experiences vs. from explicit instructions.) The more data, the more effective the learning, which is why machine learning and big data are intricately tied together.Predictive Analytics is using machine learning to predict future outcomes (extrapolation), or to infer unknown data points from known ones (interpolation).–thanks to Christopher NguyenThe Full-Stack StartupEverything we've said about this trend, all in one placeQ: So what‟s a full stack startup? You‟ve mentioned that it‟s a new, important trend, and a pattern of startups we‟ve been seeing over the past couple of years.Chris Dixon: The old approach startups took was to sell or license their new technology to incumbents. The new, “full stack” approach is to build a complete, end-to-end product or service that bypasses incumbents and other competitors.A good example from big companies is Apple versus Microsoft. For years, Microsoft just built pieces of the stack —the OS, apps — and relied on partners to build semiconductors, cases, assembly, do retail etc. Apple does everything: they design their own chips, their own phone hardware, their own OS, their own apps, the packaging, the retail experience etc. Apple reminded the world that you could create a really magical experience if you did many things well at once.Q: For example?Dixon: I think a good example is ride-sharing, like with Lyft and Uber.Before these companies were started, there were multiple startups that tried to build software that would make the taxi and limo industry more efficient. Then they went out and knocked on the door of taxi companies and pitched them on their software.For a variety of reasons, it didn‟t work. Taxi companies weren‟t thinking about software as a competitive advantage. They didn‟t have the appropriate cost structures or anyone to even evaluate the software.So when technology startups tried to inject technology and software into that industry, it didn‟t take.Companies like Lyft and Uber said: “You know what? Instead of trying to sell software as an add-on, we‟re going just goin g build the whole service using our modern software.” They asked: What would this industry look like if it were rebuilt from scratch using technology we have today?Once they brought this technology-infused product to market, consumers and drivers loved i t. It‟s basically taking over, and those companies were started just a few years ago.Q: What are the advantages of building the end-to-end experience yourself?Dixon: Well for one thing, as I mentioned earlier, the full-stack approach lets startups bypass incumbents and overcome cultural resistance to new tech.Another advantage is that full-stack startups can capture a greater portion of the economic benefits they provide. Before, the product and services they provided might have been quite valuable, but with no relationship with the end customer, it was hard to get paid accordingly or to collect the right data back to improve their products.Finally, for end users, full stack startups deliver a much better experience, because they have complete contro l. It‟s the difference between buying a beautiful, pristine Apple productversus a crappy Frankenstein PC cobbled together from dozens of vendors.Q: Okay, so isn‟t all this the same as being “vertically integrated”?Dixon: I don‟t think full-stack startups are vertically integrated in the classic sense. This isn‟t an oil company buying a supplier; it‟s a tech company building the complete experience and wrapping “non tech” functions around the tech to go after existing companies. In my opinion, “vertically integrated” is an overloaded phrase at this point and therefore not very useful.But I kind of regret calling it “full stack”, to be honest. It was just a metaphor, it was meant to be kind of a whimsical allusion to the programming phrase. “End to end” might be a good name. Another existing phrase that fits the concept is Bill Davidow‟s “whole product”.Q: Besides the examples you shared already, what are some other examples of full-stack startups?Dixon: Altschool, Buzzfeed, Harry‟s, Nest, Tesla, Warby Parker.Q: So what‟s going to happen next?Dixon: I think we‟ll start to see many more industries that have mostly resisted technology finally stop resisting now that startups have figured out the right approach to take here.The big, obvious industries include: education, healthcare, food, transportation, and financial services. All the areas of the economy where prices have outpaced inflation due to lack of technology.Q. What are the main challenges for full stack startups?Dixon: Full stack founders care about every aspect of their product/service, so they need to get good at many different things besides software —hardware, design, consumer marketing, supply chain management, sales, partnerships, regulation, etc. It takes a special kind of founder to do this.The good news is if they pull it off, it will be extremely hard for competitors to replicate all those interlocking pieces. There will be some very big companies created using the full-stack approach.–drawn from the original post by Chris Dixon (with thanks to the commenters for their questions), podcast, and other interviews. [See also: this tweetstorm from Balaji Srinivasan for additional thoughts on the trend.]disruption · full stack startups · trends ·More TrendsContainersThe containerization of code is finally hereImagine carrying around a bag of cell phones everywhere because you needed to use a new and separate piece of hardware for every single app — games, email, etc. — that you currently run on a single smartphone today.Seems crazy, but that‟s essentially what our model forcomputing used to be, with standalone silos of hardware for separate applications. When an application wasn‟t being used, the hardware and operating system dedicated to it was still sucking power and resources. It was highly inefficient. (In our bag of cellphones analogy, this would be like having one cellphone dedicated to games and still carrying it around with you everywhere all day — even though you only needed to do email during the day).Then along came virtual machines (VMs), which allowed many applications to run on top of a single piece of hardware (by making it look like multiple physical computers). This led to anywhere from 3-10x more capacity utilization. Because the software could run independently of its underlying hardware, virtual machines in data centers meant we weren‟t limited to one particular operating system; we could host both Linux and Windows on the same physical machine. The downside of that flexibility is that it introduced an entirely new layer of software (and an entirely new practice of systems management to care for it) that sat between the hardware and the operating system. Having what was essentially “an operating system on top of an operating system” caused some p erformance issues.The holy grail here would have been virtualization, but with bare-metal performance … as if the operating system were running right on top of the CPU without the middle virtual layer.That‟s where containers come in. Containers are ano ther way of isolating an application from the underlying hardware. They‟re objects that serve the same purpose as a virtual machine —instantiations or clones of the host operating system with a self-standing, self-executing application inside — but they also provide bare-metal performance, because there‟s no virtualization layer between.Why now and what next?Containers aren‟t actually a new thing. They‟ve been around for a long time, but are taking off for a few reasons. One reason is because Windows has become less prevalent in the datacenter; one of the downsides of containers compared to VMs is that they can‟t run multiple operating systems, like Windows on top of Linux. Another reason is “microservices” app architecture driving enthusiasm for containers; these app architectures areespecially suited to containers because they have discrete pieces of functionality that can scale independently, like LEGO building blocks.System administrators and people deploying code find containerization convenient because every part of the application — including device drivers, operating systems, and other dependencies — is part of a single, self-contained entity. This also makes it very easy to host many containers and move them around, for redundancy/ failure tolerance, capacity, feature testing, and other reasons.For me, the next step in containerization is treating the datacenter, with all its containers, like one giant computer or server. Many applications today are really just distributed systems: Applicatio ns aren‟t necessarily confined to just one container. We might have an application that consists of ten containers running together. We could have 1,000 applications running across 10,000 containers. Or we might have a single big data job that involves multiple, interdependent applications.So there needs to be a simple abstraction to run the operatingenvironment and manage how all this stuff gets utilized for the right capacity, reliability, and performance (the key metrics by which a datacenter operates). And the key to that is comprehensive management of the entire operating environment. That‟s what needs to happen.–Peter Levinecontainers · data centers · operating systems · trends · virtual machines ·More TrendsDigital HealthSoftware doesn't have an MDWe only have about a million physicians in the United States —but they‟re about to get reinforcements.If you think about the CAT scan that an MD is using, if you think about almost any modern device that a doctor is using —it‟s useless with out the code in it. That code was likely written by someone without an MD, someone who was evaluated ascompetent and hired by a commercial vendor of mission-critical medical instruments. The instruments that represent the foundation of modern medicine are thus today typically programmed by people who know how to code (but lack MDs) and used by people who have MDs (but usually do not know how to code).So a large percentage of medicine is already being effectively practiced by non-MDs.Moreover, the interior workings of the instruments are black boxes; MDs interface with them through vendor-provided UIs and interpret the readouts by looking up data stored in their head. As these UIs get better and smarter, less interpretive skill is required by the MD. The MD is happier — the instrument gets the right answer with less work. It‟s used more frequently. Through successively more sophisticated engineering the instrument thus begins to move from the hands of the specialist to the generalist to the nurse practiti oner to the nurse… and then, perhaps, to the general population in the form of a phone accessory or an app.That last step is starting to happen as various personal genomic, quantified self, and mobile diagnostic technologies become more accessible. These technologies produce data from the body, and that data is going to be stored in our phones. The interpretation of that data is going to be performed by software.And so that large percentage of medicine that is effectively being practiced by non-MDs is going to expand.One center of action is likely going to be the mobile programmable medical record —the container for all diagnostics and test results —something like what Apple‟s HealthKit may evolve into. Essentially just a bunch of data containers for your heart rate history, your blood pressure history, your exercise history, and the like.All this diagnostic history isn‟t necessarily “big data”; it‟s just never been tracked and cross-correlated before in one place. Once technologies like HealthKit get a little more traction, millions of software engineers without MDs can build new applications on top of that data store (perhaps collected by other applications) without injuring the phone owner.Today you‟d accept without hesitation that the kid in th e garage without a degree could write an email front-end that analyzed your email, told you the best time of day to reply, or did something else interesting, unexpected, and useful with your email data. Tomorrow? To understand your personal diagnostic data, you might soon depend more upon an iPhone app developed in a garage than on your local MD.–Balaji Srinivasanbiomedicine ·diagnostics ·digital health ·healthcare ·mobile ·point of care ·quantified self · trends ·More TrendsOnline MarketplacesThey may look different, but the same core principles applyWe‟re continuing to see tremendous innovation in marketplaces. The first generation of net companies saw a few big horizontal marketplace winners like eBay and Craigslist. Butentrepreneurs are continuing to create the next generation of online marketplaces:…A ton of very interesting net businesses are still being built by hollowing out Craigslist, targeting its key categories and doing a much better job of serving user needs in those categori es( e.g., “sublets/temporary”, “vacation rentals”, “rideshare”).…Mobile is enabling a new generation of “mobile-first” marketplaces that have dramatically superior ease of use, and enable constant grazing of marketplace content throughout the day.…A specific category of marketplaces enabled by mobile are “people marketplaces” (which I‟ve talked about before), where consumers can find highly tailored services and contractors can find opportunity.…Marketplaces are being applied to new segments. One example is bringing marketplace dynamics to serve the needs of businesses, such as B2B rental of large construction equipment.And this is just the beginning.Despite all the new forms these marketplaces are taking today (compared to what I saw at eBay 10-15 years ago), I strongly believe that the same principles for managing them still apply.Why? It has to do with “perfect competition”. Economics has typically presented this as a largely theoretical construct. But it is something I believe truly exists in the real world, based on my experience working for or with numerous online marketplaces.Today‟s new marketplaces must nurture and manage perfect competition to thriveBut first, what is “perfect competition”? Investopedia defines it as a market structure where the following five key characteristics are met —I‟ve illustrated each using the example of eBay:1. All firms sell an identical product. Not all goods for saleon eBay are identical, obviously. But there are precious few products for sale on eBay that are truly unique, “one-of-a-kind” items. Eventually competition with comparable products would bring marginal returns in line with marginal costs, essentially “competing away” surplus profits in that niche.2. All firms are price takers. On eBay, every seller is forced to take the prices that buyers are willing to pay for that good. Some sellers could command a premium due to a stronger reputation built over trusted quality transactions, but it‟s a small premium.3. All firms have relatively small market share. With lots of sellers selling across lots of categories, no single seller approached meaningful market share within any single category. No seller on the entire eBay marketplace had anything approaching market power.4. Buyers have complete information about the product being sold and the prices charged by each firm. The marketplace‟s job is to generate transparency for buyers across all meaningful dimensions, including specifications, cost, and trust/safety。
![The Great Filter 大筛选——我们在其中走了多远？[中英对照翻译]](https://img.taocdn.com/s1/m/6121f40252ea551810a68749.png)
