From Storage to Data Virtualization

Do you remember Primary Data? Well, I loved the idea and the team but it didn’t go very well for them. It’s likely there are several reasons why it didn’t. In my opinion, it boiled down to the fact that very few people like storage virtualization. In fact, I expressed my fondness for Primary Data’s technology several times in the past, but when it comes to changing the way to operate complex, siloed, storage environments you come across huge resistance, at every level!
The good news is that Primary Data’s core team is back, with what looks like a smarter version of the original idea that can easily overcome the skepticism surrounding storage virtualization. In fact, they’ve moved beyond it and presented what looks like a multi-cloud controller with data virtualization features. Ok, they call it “Data as a Service,” but I prefer Data Virtualization…and being back with the product is a bold move.
Data Virtualization (What and Why)
I’ve begun this story by mentioning Primary Data first, because David Flynn (CEO of HammerSpace and former CTO of Primary Data) did not start this new Hammerspace venture from scratch. He bought the code which belonged to Primary Data and used it to build the foundation of his new product. That allowed him and his team to get on the market quickly with the first version of HammerSpace in a matter of months instead of years.
HammerSpace is brilliant just for one reason. It somehow solves or, better, hides the problem of data gravity and allows their Data-as-a-Service platform to virtualize data sets by presenting virtualized views of them available in a multi-cloud environment through standard protocols like NFS or S3.
Yes, at first glance it sounds like hot air and a bunch of buzzwords mixed together, but this is far from being the case here… watch the demo in the following video if you don’t trust me.
The solution is highly scalable and aimed at Big Data analytics and other performance workloads for which you need data close to the compute resource quickly, without thinking too much about how to move, sync, and keep it updated with changing business needs.
HammerSpace solutions have several benefits but the top two on my list are:

  • The minimization of egress costs: This is a common problem for those working in multi-cloud environments today. With HammerSpace, only necessary data is moved where it is really needed.
  • Reduced latency: It’s crazy to have an application running on a cloud that is far from where you have your data. Just to make an example, the other day I was writing about Oracle cloud, and how  good they are at creating high-speed bare-metal instances at a reasonable cost. This benefit can be easily lost if your data is created and stored in another cloud.

The Magic of Data Virtualization
I won’t go through architectural and technical details, since there are videos and documentation on HammerSpace’s website that address them (here and here).  Instead, I want to mention one of the features that I like the most: the ability to query the metadata of your data volumes. These volumes can be anywhere, including your premises, and you can get a result in the form of a new volume that is then kept in sync with the original data. Everything you do on data and metadata is quickly reflected on child volumes. Isn’t it magic?
What I liked the least, even though I understand the technical difficulties in implementing it, is that this process is one-way when a local NAS is involved… meaning that it is only a source of data and can’t be synced back from the cloud. There is a workaround, however, and it might be solved in future releases of the product.
Closing the Circle
HammerSpace exited stealth mode only a few days ago. I’m sure that by digging deeper into the product, flaws and limitations will be found.t is also true that the more advanced features are still only sketched on paper. But I can easily get excited by innovative technologies like this one and I’m confident that these issues will be fixed over time. I’ve been keeping an eye on multi-cloud storage solutions for a while, and now I’ve added Hammerspace to my list.
Multi-cloud data controllers and data virtualization are the focus of an upcoming report I’m writing for GigaOm Research. If you are interested in finding out more about how data storage is evolving in the cloud era, subscribe to GigaOm Research for Future-Forward Advice on Data-Driven Technologies, Operations, and Business Strategies.

Voices in AI – Episode 62: A Conversation with Atif Kureishy


About this Episode

Episode 62 of Voices in AI features host Byron Reese and Atif Kureishy discussing AI, deep learning, and the practical examples and implications in the business market and beyond. Atif Kureishy is the Global VP of Emerging Practices at Think Big, a Teradata company. He also has a B.S. in physics and math from the University of Maryland as well as an MS in distributive computing from Johns Hopkins University.
Visit to listen to this one-hour podcast or read the full transcript.

Transcript Excerpt

Byron Reese: This is Voices in AI, brought to you by GigaOm, I’m Byron Reese. Today my guest is Atif Kureishy. He is the Global VP of Emerging Practices, which is AI and deep learning at Think Big, a Teradata company. He holds a BS in Physics and Math from the University of Maryland, Baltimore County, and an MS in distributive computing from the Johns Hopkins University. Welcome to the show Atif.
Atif Kureishy: Welcome, thank you, appreciate it.
So I always like to start off by just asking you to define artificial intelligence.
Yeah, definitely an important definition, one that unfortunately is overused and stretched in many different ways. Here at Think Big we actually have a very specific definition within the enterprise. But before I give that, for me in particular, when I think of intelligence, that conjures up the ability to understand, the ability to reason, the ability to learn, and we usually equate that to biological systems, or living entities, and now with the rise of probably more appropriate machine intelligence, we’re applying the term ‘artificial’ to it, and the rationale is probably because machines aren’t living and they’re not biological systems.
So with that, the way we’ve defined AI in particular is: leveraging machine and deep learning to drive towards a specific business outcome. And it’s about giving leverage for human workers, to enable higher degrees of assistance and higher degrees of automation. And when we define AI in that way, we actually give it three characteristics. Those three characteristics are: the ability to sense and learn, and so that’s being able to understand massive amounts to data and demonstrate continuous learning, and detecting patterns and signals within the noise, if you will. And the second is being able to reason and infer, and that is driving intuition and inference with increasing accuracy again to maximize a business outcome or a business decision. And then ultimately it’s about deciding and acting, so actioning or automating a decision based on everything that’s understood, to drive towards more informed activities that are based on corporate intelligence. So that’s kind of how we view AI in particular.
Well I applaud you for having given it so much thought, and there’s a lot there to unpack. You talked about intelligence being about understanding and reasoning and learning, and that was even in your three areas. Do you believe machines can reason?
You know, over time, we’re going to start to apply algorithms and specific models to the concept of reasoning, and so the ability to understand, the ability to learn, are things that we’re going to express in mathematical terms no doubt. Does it give it human lifelike characteristics? That’s still something to be determined.
Well I don’t mean to be difficult with the definition because, as you point out, most people aren’t particularly rigorous when it comes to it. But if it’s to drive an outcome, take a cat food dish that refills itself when it’s low, it can sense, it can reason that it should put more food in, and then it can act and release a mechanism that refills the food dish, is that AI, in your understanding, and if not why isn’t that AI?
Yeah, I mean I think in some sense it checks a lot of the boxes, but the reality is, being able to adapt and understand what’s occurring, for instance if that cat is coming out during certain times of the day ensuring that meals are prepared in the right way and that they don’t sit out and become stale or become spoiled in any way, and that is signs of a more intelligent type of capability that is learning the behaviors and anticipating how best to respond given a specific outcome it’s driving towards.
Got you. So now, to take that definition, your company is Think Big. What do you think big about? What is Think Big and what do you do?
So looking back in history a little bit, Think Big was actually an acquisition that Teradata had done several years ago, in the big data space, and particularly around open source and consulting. And over time, Teradata had made several acquisitions and now we’ve unified all of those various acquisitions into a unified group, called Think Big Analytics. And so what we’re particularly focused on is how do we drive business outcomes using advanced analytics and data science. And we do that through a blend of approaches and techniques and technology frankly.
Listen to this one-hour episode or read the full transcript at
Byron explores issues around artificial intelligence and conscious computers in his new book The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity.

Self-Service Master Data Management

Once data is under management in its best-fit leveragable platform in an organization, it is as prepared as it can be to serve its many callings. It is in position to be used for purposes operationally and analytically and across the spectrum of need. Ideas emerge from business areas no longer encumbered with the burden of managing data, which can be 60% – 70% of the effort to bring the idea to reality. Walls of distrust in data come down and the organization can truly excel with an important barrier to success removed.
An important goal of the information management function in an organization is to get all data under management by this definition, and to keep it under management as systems come and go over time.
Master Data Management (MDM) is one of these key leveragable platforms. It is the elegant place for data with widespread use in the organization. It becomes the system of record for customer, product, store, material, reference and all other non-transactional data. MDM data can be accessed directly from the hub or, more commonly, mapped and distributed widely throughout the organization. This use of MDM data does not even account for the significant MDM benefit of efficiently creating and curating master data to begin with.
MDM benefits are many, including hierarchy management, data quality, data governance/workflow, data curation, and data distribution. One overlooked benefit is just having a database where trusted data can be accessed. Like any data for access, the visualization aspect of this is important. With MDM data having a strong associative quality to it, the graph representation works quite well.
Graph traversals are a natural way for analyzing network patterns. Graphs can handle high degrees of separation with ease and facilitate visualization and exploration of networks and hierarchies. Graph databases themselves are no substitute for MDM as they provide only one of the many necessary functions that an MDM tool does. However, when graph technology is embedded within MDM, such as what IBM is doing in InfoSphere MDM – it is very powerful.
Graph technology is one of the many ways to facilitate self-service to MDM. Long a goal of business intelligence, self-service has significant applicability to MDM as well. Self-service is opportunity oriented. Users may want to validate a hypothesis, experiment, innovate, etc. Long development cycles or laborious process between a user and the data can be frustrating.
Historically, the burden for all MDM functions has fallen squarely on a centralized, development function. It’s overloaded and, as with the self-service business intelligence movement, needs disintermediation. IBM is fundamentally changing this dynamic with the next release of Infosphere MDM. Its self-service data import, matching, and lightweight analytics allows the business user to find, share and get insight from both MDM and other data.
Then there’s Big Match. Big Match can analyze structured and unstructured customer data together to gain deeper customer insights. It can enable fast, efficient linking of data from multiple sources to grow and curate customer information. The majority of the information in your organization that is not under management is unstructured data. Unstructured data has always been a valuable asset to organizations, but it can be difficult to manage. Emails, documents, medical records, contracts, design specifications, legal agreements, advertisements, delivery instructions, and other text-based sources of information do not fit neatly into tabular relational databases. Most BI tools on MDM data offer the ability to drill down and roll up data in reports and dashboards, which is good. But what about the ability to “walk sideways” across data sources to discover how different parts of the business interrelate?
Using unstructured data for customer profiling allows organizations to unify diverse data from inside and outside the enterprise—even the “ugly” stuff; that is, dirty data that is incompatible with highly structured, fact-dimension data that would have been too costly to combine using traditional integration and ETL methods.
Finally, unstructured data management enables text analytics, so that organizations can gain insight into customer sentiment, competitive trends, current news trends, and other critical business information. In text analytics, everything is fair game for consideration, including customer complaints, product reviews from the web, call center transcripts, medical records, and comment/note fields in an operational system. Combining unstructured data with artificial intelligence and natural language processing can extract new attributes and facts for entities such as people, location, and sentiment from text, which can then be used to enrich the analytic experience.
All of these uses and capabilities are enhanced if they can be provided using a self-service interface that users can easily leverage to enrich data from within their apps and sources. This opens up a whole new world for discovery.
With graph technology, distribution of the publishing function and the integration of all data including unstructured data, MDM can truly have important data under management, empower the business user, be the cornerstone to digital transformation and truly be self-service.

Voices in AI – Episode 19: A Conversation with Manoj Saxena

In this episode, Byron and Manoj discuss cognitive computing, consciousness, data, DARPA, explainability, and superconvergence.
[podcast_player name=”Episode 19: A Conversation with Manoj Saxena” artist=”Byron Reese” album=”Voices in AI” url=”″ cover_art_url=””]
Byron Reese: This is Voices in AI, brought to you by Gigaom. Today my guest is Manoj Saxena. He is the Executive Chairman of CognitiveScale. Before that, he was the General Manager of IBM Watson, the first General Manager, in fact. He’s also a successful entrepreneur who founded and sold two venture-backed companies within five years. He’s the Founding Managing Director of the Entrepreneur’s Fund IV, a 100-million-dollar seed fund focused exclusively on cognitive computing. He holds an MBA from Michigan State University and a Master’s in Management Sciences from the Birla Institute of Technology and Science in Pilani, India. Welcome to the show, Manoj.
Manoj Saxena: Thank you.
You’re well-known for eschewing the term “artificial intelligence” in favor of “cognitive computing”; even your bio says cognitive computing. Why is that?
AI, to me, is the science of making intelligent systems and intelligent machines. Cognitive computing, and most of AI, is around replacing the human mind and creating systems that do the jobs of human beings. I think the biggest opportunity and it has been proven out in multiple research reports, is augmenting human beings. So, AI for me is not artificial intelligence; AI for me is augmented intelligence. It’s how you could use machines to augment and extend the capabilities of human beings. And cognitive computing uses artificial intelligence technologies and others, to pair man and machine in a way that augments human decision-making and augments human experience.
I look at cognitive computing as the application of artificial intelligence and other technologies to create—I call it the Iron Man J.A.R.V.I.S. suit, that makes every human being a superhuman being. That’s what cognitive computing is, and that was frankly the category that we started off when I was running IBM Watson as, what we believed, was the next big thing to happen in IT and in enterprise.
When AI was first conceived, and they met at Dartmouth and all that, they thought they could kind of knock it out in the summer. And I think the thesis was, Minsky later said, it was just like physics had just a few laws, and electricity had just a few laws, they thought there was just a couple of laws. And then AIs had a few false starts, expert systems and so forth, but, right now, there’s an enormous amount of optimism about it, of what we’re going to be able to do. What’s changed in the last, say, decade?
I think a couple of dimensions in that, one is, when AI initially got going the whole intention was, “AI to model the world.” Then it shifted to, “AI to model the human mind.” And now, where I believe the most potential is, is, “AI to model human and business experiences.” Because each of those are gigantic. The first ones, “AI to model the world” and “AI to model the mind,” are massive exercises. In many cases, we don’t even know how the mind works, so how do you model something that you don’t understand? The world is too complex and too dynamic to be able to model something that large.
I believe the more pragmatic way is to use AI to model micro-experiences, whether it’s an Uber app, or a Waze. Or it is to model a business process, whether it’s a claim settlement, or underwriting, or management of diabetes. I think that’s where the third age of AI will be more focused around, not modeling the world or modeling the mind, but to model the human experience and a business process.
So is that saying we’ve lowered our expectations of it?
I think we have specialized in it. If you look at the human mind, again, you don’t go from being a child to a genius overnight. Let alone a genius that understands all sciences and all languages and all countries. I think we have gotten more pragmatic and more outcome driven, rather than more research and science-driven on how and where to apply AI.
I notice you’ve twice used the word “the mind,” and not “the brain.” Is that deliberate, and if so, where do you think “the mind” comes from?
I think there is a lot of hype, and there is a lot of misperception about AI right now. I like saying that, “AI today is both: AI equals ‘artificially inflated,’ and AI equals ‘amazing innovations.’” And I think in the realm of “AI equals artificially inflated,” there are five myths. One of the first myths is that AI equals replacement of the human mind. And I separate the human brain from the human mind, and from human consciousness. So, at best, what we’re trying to do is emulate functions of a human brain in certain parts of AI, let alone human mind or human consciousness.
We talked about this last time, we don’t even know what consciousness is, other than a doctor saying whether the patient is dead or alive. There is no consciousness detector. And a human mind, there is a saying that you probably need a quantum machine to really figure out how a human mind works—it’s not a Boolean machine or von Neumann machine; it’s a different kind of a processor. But a human brain, I think, can be broken down and can be augmented through AI to create exceptional outcomes. And we’ve seen that happen in radiology, at Wall Street, the quants, and other areas. I think that’s much more exciting, to apply AI pragmatically into these niches.
You know, it’s really interesting because there’s been a twenty-year effort called OpenWorm Project, to take the nematode worm’s brain, which is 302 neurons, and to model it. And even after twenty years, people in the project say it may not be possible. And so, if you can’t do a nematode… One thing is certain, you’re not going to do a human before you do a nematode worm.
Exactly. You know the way I see that, Byron, is that I’m more interested in “richer,” and not “smarter.” We need to get smarter but also we need to equally get richer. By “richer,” I don’t mean just making money, by “richer,” I mean: how do we use AI to improve our society, and our businesses, and our way of life? That’s where I think coming at it in the way of “outcome in,” rather than “science out,” is a more pragmatic way to apply AI.
So, you’ve mentioned five misconceptions, that was one of them. What are some of the other ones?
The first misconception was, AI equals replacing human mind. Second misconception is, AI is the same as natural language processing, which is far from the truth—NLP is just a technique within AI. It’s like saying, “My ability to understand and read a book, is the same as my brain.” That’s the second misconception.
The third is, AI is the same as big data and analytics. Big data and analytics are tools that are used to capture more input for an AI to work on. Saying that big data is the same as AI is saying, “Just because I can sense more, I can be smarter.” All big data is giving you is more input; it’s giving you more senses. It’s not making you smarter, or more intelligent. That’s the third myth.
The fourth myth is that AI is something that is better implemented horizontally versus vertically. I believe true AI, and successful AI—particularly in the business world—will have to be verticalized AI. Because it’s one thing to say, “I’ve got an AI.” It’s another thing to say, “I have an AI that understands underwriting,” versus an AI that understands diabetes, versus an AI that understands Super Bowl ads. Each of these require a domain-specific optimization of data and models and algorithms and experience. And that’s the fourth one.
The fifth one is that AI is all about technology. At best, AI is only half about technology. The other half of the equation has to do with skills, has to do with new processes, and methods, and governance on how to manage AI responsibly in the enterprise. Just like when the Internet came about, you didn’t have the methods and processes to create a web page, to build a website, to manage the website from getting hacked, to manage updates of the website. Similarly, there is a whole AI lifecycle management, and that’s what CognitiveScale focuses on: how do you create, deploy, and manage AI responsibly and at scale?
Because, unlike traditional IT systems—which do not learn; they are mostly rules-based systems, and rules-based systems don’t learn—AI-based systems are pattern-based, and they learn from patterns. So, unlike traditional IT systems that did not learn, AI systems have an ability to self-learn and geometrically improve themselves. If you can’t get visibility and control over these AI systems, you could have a massive problem of “rogue AI”—is what CognitiveScale calls it—where it’s irresponsible AI. You know that character Chucky from the horror movie, it’s like having a bunch of Chuckys running around in your enterprise, opening up your systems. What is needed is a comprehensive end-to-end view of managing AI from design, from deployment, to production, and governance of it at scale. That requires a lot more than technology; it requires skills and methods; and processes.
When we were chatting earlier you mentioned that some people were having difficulty scaling their projects, that they began in their enterprise, making them kind of enterprise-ready. Talk about that for a moment. Why is that, and what’s the solution to that?
Yes. I’ve talked to over six hundred customers just in the last five years—everything from IT level to board level and CEO level. There are three big things that are going on that they’re struggling with getting value of AI. Number one is, AI is seen as something that can be done by data scientists and analytics people. AI is far too important to be left to just data scientists. AI has to be done as a business strategy. AI has to be done top-down to drive business outcomes, not bottom-up as a way of finding data patterns. That’s the first part. I see a lot of science projects that are happening. One of the customers called it darts versus bubbles. He says, “There are lots of darts of projects that are going on, but where do I know where the big bubbles are, which really move the needle for a multibillion-dollar business that I have?” There is a lot of, I call it, bottom-up engineering experiments that are going on, that are not moving the needle. That’s one thing.
Number two is, the data scientists and application developers are struggling with taking these projects into production, because they are not able to provide fundamental capabilities to AI that you need in an enterprise, such as explainability. I believe 99.9% of the AI companies today that are funded will not make it in the next three years, because they lack some fundamental capability, like explainability. It’s one thing to find pictures of cats on the internet using a deep learning network, it’s another thing to explain to a chief risk officer why a particular claim was denied, and the patient died, and now they have a hundred-million-dollar lawsuit. The AI has to be responsible, trustworthy, and explainable; able to say why was that decision made at that time. Because of lack of these kinds of capabilities—and there are five such capabilities that we call enterprise-grade AI—most of these projects are not able to move into production, because they’re not able to meet the requirements from a security and performance perspective.
And then last but not least, these skills are very sparse. There are very few skills. Someone told me there are only seven thousand people in this world who have the skills to be able to understand and run AI models and networks like deep learning and others. Imagine that, seven thousand. I know of a bank who’s got twenty-two thousand developers, one bank alone. There is a tremendous gap in the way AI is being practiced today, versus the skills that are available in trying to get this production-ready.
That’s another thing that CognitiveScale is doing, we have created this platform to democratize AI. How do you take application developers and data scientists and machine learning people, and get them to collaborate, and deploy AI in 90-day increments? We have this method called “10-10-10,” where, in 10 hours we select a use case, and in 10 days we build the reference application using their data, and in 10 weeks we take them into production. We do that by helping these groups of people collaborate on a new platform called Cortex, that lets you take AI safely and securely into production, at scale.
Backing that up a little bit, there are European efforts to require that if the AI makes a decision about you, that you have a right to understand to know why—why it denies you a loan. So, you’re saying that that is something that isn’t happening now, but it is something that’s possible.
Actually, there are some efforts that are going on right now. DARPA has got some initiatives around this notion of XAI, explainable AI. And I know other companies are exploring this, but it’s still a very low-level technology effort. It is not coming up—explainable AI—at a business process level, and at an industry level, because explainability requirements of an AI vary from process to process, and from industry to industry. The explainability requirements for a throat cancer specialist talking talk about why he recommended a treatment, are different than explainability requirements for an investment advice manager in wealth management, who says, “Here’s the portfolio I recommended to you with our systems of AI.” So, explainability exists at two levels. It exists at a horizontal level as a technology, and it exists at an industry-optimized level, and that’s why I believe AI has to be verticalized and industry-optimized for it to really take off.
You think that’s a valid request to ask of an AI system.
I think it’s a requirement.
But if you ask a Google person, “I rank number three for this search. Somebody else ranks number four. Why am I three and they’re four?” They’d be like, “I don’t know. There are six thousand different things going on.”
Exactly. Yeah.
So wouldn’t an explainability requirement impede the development of the technology?
Or, it can create a new class of leaders who know how to crack that nut. That’s the basis on which we have founded CognitiveScale. It’s one of the six requirements, that we’ve talked about, in creating enterprise-grade AI. One of the big things—and I learned this while we were doing Watson—was how do you build AI systems you can trust, as a human being? Explainability is one of them. Another one is, recommendations with reasons. When your AI gives you an insight, can it also give you evidence to support, “Why I’m suggesting this as the best course of action for you”? That builds trust in the AI, and that’s when the human being can take action. Evidence and explainability are two of those dimensions that are requirements of enterprise-grade AI and for AI to be successful at large.
There’s seven thousand people who understand that. Assuming it’s true, is that a function of how difficult it is, or how new it is?
I think it’s a function of how different a skill set it is that we’re trying to bring into the enterprise. It is also how difficult it is. It’s like the Web; I keep going back to Internet. We are like where the Internet was in 1997. There were probably, at that time, only a few thousand people who knew how to develop HTML-based applications or web pages. AI today is where the Internet was in 1996 and 1997, where people were building a web page by hand. It’s far different from building a web application, which is connecting a series of these web pages, and orchestrating them to a business process to drive an outcome. That’s far different from optimizing that process to an industry, and managing it at the requirement of explainability, governance, and scalability. There is a lot of innovation around enterprise AI that is yet to come about, and we have not even scratched the surface yet.
When the Web came out in ’97, people rushed to have a web department in their company. Are we there, are we making AI departments and is that, like, not the way to do it?
Absolutely. I won’t say it’s not the way to do it. I’ll say it’s a required first step; to really understand and learn. Not only just AI, even blockchain—CognitiveScale calls it “blockchain with a brain.” I think that’s the big transformation, which has yet to happen, that’s on the horizon in the next three to four years—where you start building self-learning and self-assuring processes. Coming back to the Web analogy, that was the first step of three or four, in making a business become an e-business. Twenty-five years ago when the Web came about, everyone became in e-business, every process became “webified.” Now, with AI, everyone will become an i-business, or a c-business—a cognitive business—and everyone is going to get “cognitized.” Every process is going to get cognitized. Every process will learn from new data, and new interactions.
The steps they will go through are not unlike what they went through with the Web. Initially, they had a group of people building web apps, and the CEO said after a while, 1998, “I’ve spent half a million dollars, all I have is an intelligent digital brochure on the website. What has it done for my business?” That is exactly the stage we are at. Then, someone else came up and said, “Hey, I can connect a shopping cart to this particular set of web pages. I can put a payment system around it. I can create an e-commerce system out of it. And I have this open-source thing called JBoss, that you can build off of.” That’s kind of similar to what Google TensorFlow is doing today for AI. Then, there are next-generation companies like Siebel and Salesforce that came in and said, “I can build for you a commercial, web-based CRM system.” Similarly, that’s what CognitiveScale does. We are building the next-generation intelligent CRM system, or intelligent HRM system, that lets you get value out of these systems in a reliable and scalable manner. So it’s sort of the same progression that they’re going to go through with AI, like we went through with the Web. And there’s still a tremendous amount of innovation and new market leadership. I believe there will be a new hundred-billion-dollar AI company and that will get formed in the next seven to ten years.
What’s the timescale on AI going to be, is it going to be faster or slower?
I think it’ll be faster. I think it’ll be faster for multiple reasons. We have, and I gave a little TED Talk on this, around this notion of a superconvergence of technologies. When the Web came about, we were shifting from just one technology to another—we moved from client-server to Web. Right now, you’ve got these super six technologies that are converging that will make AI adoption much faster—they are cloud, mobile, social, big data, blockchain, and analytics. All of these are coming together at a rate and pace that is enabling compute and access, at a scale that was never possible before, and you combine that with an ability for a business to get disrupted dramatically.
One of the biggest reasons that AI is different than the Web is that those web systems are rules-based. They did not geometrically learn and improve. The concern and the worry that the CEOs and boards have this time around is—unlike a web-based system—an AI-based system improves with time, and learns with time, so either I’m going to dramatically get ahead of the competition, or I’m going to be dramatically left behind. What some people call “the Uber-ification” of businesses. There is this threat, and an opportunity to use AI as a great transformation and accelerator for their business model. That’s where this becomes an incredibly exciting technology, riding on the back of the superconvergence that we have.
If a CEO is listening, and they hear that, and they say, “That sounds plausible. What is my first step?”
I think there are three steps. The first step is to educate yourself, and your leadership team, on the business possibilities of AI—AI-powered business transformation, not technology possibilities of AI. So, one step is just education; educate yourself. Second is, to start experimenting. Experiment by deploying 90-day projects that cost a few hundred thousand dollars, not a two-year project with multiple million dollars put into it, so you can really start understanding the possibilities. Also you can start cutting through the vendor hype about what is product and what is PowerPoint. The narrative for AI, unfortunately, today, is being written by either Hollywood, or by glue-sniffing marketers from large companies, so the 90-day projects will help you cut through it. So, first is to educate, second is experiment, and third is enable. Enable your workforce to really start having the skill sets and the governance and the processes and enable an ecosystem, to really build out the right set of partners—with technology, data, and skills—to start cognitizing your business.
You know AI has always kind of been benchmarked against games, and what games it can beat people at. And that’s, I assume, because games are these closed environments with fixed rules. Is that the way an enterprise should go about looking for candidate projects, look for things that look like games? I have a stack of resumes, I have a bunch of employees who got great performance reviews, I have a bunch of employees that didn’t. Which ones match?
I think that’s the wrong metaphor to use. I think the way to have a business think about AI, is in the context of three things: their customers, their employees, and their business processes. They have to think about, “How can I use AI in a way that my customer experience is transformed? That every customer feels very individualized, and personalized, in terms of how I’m engaging them?” So, that’s one, the customer experiences that are highly personalized and highly contextualized. Second is employee expertise. “How do I augment my experience and expertise of my employees such that every employee becomes my smartest employee?” This is the Iron Man J.A.R.V.I.S. suit. It’s, “How do I upskill my employees to be the smartest at making decisions, to be the smartest in handling exceptions?” The third thing is my business processes. “How do I implement business processes that are constantly learning on their own, from new data and from new customer interaction?” I think if I were a CEO of a business, I would look at it from those three vectors and then implement projects in 90-day increments to learn about what’s possible across those three dimensions.
Talk a minute about CognitiveScale. How does it fit into that mix?
CognitiveScale was founded by a series of executives who were part of IBM Watson, so it was me and the guy who ran Watson Labs. We ran it for the first three years, and one thing we immediately realized was how powerful and transformative this technology is. We came away with three things: first, we realized that for AI to be really successful, it has to be verticalized and it has to really be optimized to an industry. Number two is that the power of AI is not in a human being asking the question of an AI, but it’s the AI telling the human being what questions to ask and what information to look for. We call it the “known unknowns” versus “unknown unknowns.” Today, why is it that I have to ask an Alexa? Why doesn’t Alexa tell me when I wake up, “Hey, while you were sleeping, Brexit happened. And—” if I’m an investment adviser, “—here are the seventeen customers you should call today and take them through the implications, because they’re probably panicking.” It’s using a system which is the opposite of a BI. A BI is a known-unknown—I know I don’t know something, therefore I run a query. An AI is an unknown unknown, which means it’s tapping me on the shoulder and saying, “You ought to know this,” or, “You ought to do this.” So, that was the second thesis. One is verticalize, second is unknown unknowns, and the third is quick value in 90-day increments—this is delivered using the method we call “10-10-10,” where we can stand up little AIs in 90-day increments.
The company got started about three-and-a-half years ago and the mission is to create exponential business outcomes in healthcare, financial services, telecom, and media. The company has done incredibly well, we have investments from Microsoft, Intel, IBM, Norwest—raised over $50 million. There are offices in Austin, New York, London and India. And the who’s-who, there are over thirty customers who are deploying this, and now scaling this as an enterprise-wide initiative, and it’s, again, built on this whole hypothesis of driving exponential business outcomes, not driving science projects with AI.
CognitiveScale is an Austin-based company, Gigaom is an Austin-based company, and there’s a lot of AI activity in Austin. How did that come about, and is Austin an AI hub?
Absolutely, that’s one of the exciting things I’m working on. One of my roles is Executive Chairman of CognitiveScale. Another of my roles is that I have a hundred-million-dollar seed fund that focuses on investing in vertical AI companies. And for my third thing, we just announced last year, is an initiative called AI Global—out of Austin—whose focus is on fostering the deployment of responsible AI.
I believe East Coast and West Coast will have their own technology innovations in AI. AI will be bigger than the Internet was. AI will be at the scale of what electricity was. Everything we know around us—from our chairs to our lightbulbs and our glasses—is going to have elements of AI woven into it over the next ten years. And, I believe one of the opportunities that Austin has—and that’s why we founded AI Global in Austin—is to help businesses implement AI in a responsible way so that it creates good for the business in an ethical and a responsible manner.
Part of the ethical use of AI and responsible use of AI involves bringing a community of people together in Austin, and have Austin be known as the place to go, for designing responsible AI systems. We have the UT Law school working with us, the UT Design school, the UT Business school, the UT IT school—all of them are working together as one. We have the mayor’s office and the city working together extensively. We also have some local companies like USAA, who is coming in as a founding member of this. What we are doing now is helping companies that come to us for getting a prescription on how to design, deploy, and manage responsible AI systems. And I think there are tremendous opportunities, like you and I have talked about, for Gigaom and AI Global to start doing things together to foster implementation of responsible AI systems.
You may have heard that IBM Watson beat Ken Jennings at Jeopardy. Well, he gave a TED Talk about that, and he said that there was a graph that, as Watson got better, it would show the progress, and every week they would send him an update and their line would be closer to his. He said he would look at it with dread. He said, “That’s really what AI is, it’s not the Terminator coming for you. It’s the one thing you do great, and it just gets better and better and better and better at it.” And you talked about Hollywood driving the narrative of AI, but one of the narratives is AIs effect on jobs, and there’s a lot of disagreement about it. Some people believe it’s going to eat a bunch of low-skill work, and we will have permanent unemployment and it will be like the Depression, and all of that. While some think that it’s actually going to create a bunch of jobs. That, just like any other transformative technology, it’s going to raise productivity which is how we raise wages. So which of those narratives, or a different one, do you follow?
And there’s a third group that says that AI could be our last big innovation, and it’s going to wipe us out as a species. I think the first two, in fact, all three are true, elements of them.
So it will wipe us out as a civilization?
If you don’t make the right decisions. I’m hearing things like autonomous warfare which scares the daylights out of me.
Let’s take all three. In terms of AI dislocating jobs, I think every major technology—from the steam engine to the tractor to semiconductors—has always dislocated jobs; and AI will be no different. There’s a projection that by the year 2020 eighteen million jobs will be dislocated by AI. These are tasks that are routine tasks that can be automated by a machine.
Hold on just a second, that’s twenty-seven months from now.
Yeah, eighteen million jobs.
Who would say that?
It’s a report that was done by, I believe it was World Economic Forum, but here’s the thing, I think that’s quite true. But I don’t worry about that as much as I focus on the 1.3 billion jobs that AI will uplift the roles on. That’s why I look at augmentation as a bigger opportunity than replacement of human beings. Yes, AI is going to remove and kill some jobs but there is a much, much larger opportunity by using AI to augment and skill your employees, just like the Web did. The Web gave you reach and access and connection, at a scale that was never possible before—just like the telephone did before that, and the telegraph did before that. And I think AI is going to give us a tremendous amount of opportunities for creating—someone called it the “new collar jobs,” I think it was IBM—not just blue collar or white collar, but “new collar” jobs. I do believe in that; I do believe there is an entire range of jobs that AI will bring about. That’s one.
The second narrative was around AI being the last big innovation that we will make. And I think that is absolutely the possibility. If you even look at the Internet when it came about, the top two applications in the early days of the Internet were gambling and pornography. Then we started putting the Internet to work for the betterment of businesses and people, and we made choices that made us use the Internet for greater good. I think the same thing is going to happen with AI. Today, AI is being used for everything from parking tickets being contested, to Starbucks using it for coffee, to concert tickets being scalped. But I think there are going to be decisions as a society that we have to make, on how we use AI responsibly. I’ve heard the whole Elon Musk and Zuckerberg argument; I believe both of them are right. I think it all comes down to the choices we make as a society, and the way we scale our workforce on using AI as the next competitive advantage.
Now, the big unknown in all of this is what a bad actor, or nation states, can do using AI. The part that I still don’t have a full answer to, but it worries the hell out of me, is this notion of autonomous warfare. Where people think that by using AI they can actually restrict the damage, and they can start taking out targets in a very finite way. But the problem is, there’s so much that is unknown about an AI. An AI today is not trustworthy. You put that into things that can be weapons of mass destruction, and if something goes wrong—because the technology is still maturing—you’re talking about creating massive destruction at a scale that we’ve never seen before. So, I would say all three elements of the narrative: removing jobs, creating new jobs, creating an existential threat to us as a race—all of those elements are a possibility going forward. The one I’m the most excited about is how it’s going to extend and enhance our jobs.
Let’s come back to jobs in just a minute, but you brought up warfare. First of all, there appear to be eighteen countries working to make AI-based systems. And their arguments are twofold. One argument is, “There’s seventeen other people working to develop it, if I don’t…”
Someone else will. 
And second, right now, the military drops a bomb and it blows up everything… Let’s look at a landmine. A landmine isn’t AI. It will blow up anything over forty pounds. And so if somebody came and said, “I can make an AI landmine that sniffs for gunpowder, and it will only blow up somebody who’s carrying a weapon.” Then somebody else says, “I can make one that actually scans the person and looks for drab.” And so forth. If you take warfare as something that is a reality of life, why wouldn’t you want systems that were more discriminative?
That’s a great question, and I believe that will absolutely happen, and probably needs to happen, but over a period of time—maybe that’s five or ten years away. We are in the most dangerous time right now, where the hype about AI has far exceeded the reality of AI. These AIs are extremely unstable systems today. Like I said before, they are not evidence-based, there is no kill-switch in an AI, there is no explainability; there is no performance that you can really figure out. Take your example of something that can sniff gunpowder and will explode. What if I store that mine in a gun depot, in the middle of a city, and it sniffs the gunpowder from the other weapons there, and it blows itself up. Today, we don’t have the visibility and control at a fine-grain level with AI to warrant an application of it in that scale.
My view is that it will be a prerogative for everyone to get on it as nation-states—you saw Putin talk about it, saying, “He who controls AI will control the future world.” There is no putting the genie back in the bottle. And just like we did with the rules of war, and just like we did with nuclear warfare; there will be new Geneva Convention-like rules that we will have to come up with as a society on how and where these responsible AI systems have to be deployed, and managed, and measured. So, just like we have done that for chemical warfare, I think there will be new rules that will come up for AI-based warfare.
But the trick with it is… A nuclear event is a binary thing; it either happened or it didn’t. A chemical weapon, there is a list of chemicals, that’s a binary thing. AI isn’t though. You can say your dog-food dish that refills automatically when it’s empty, that’s AI. How would you even phrase the law, assuming people followed it, how would you phrase it in just plain English?
In a very simple way. You’ve heard Isaac Asimov’s three rules in I, Robot. I think as a society we will have to—in fact, I’m doing a conference on this next year north of London around how to use AI and drones in warfare in a responsible way—come up with a collective mindset and will from the nations to propose something like this. And I think the first event has not happened yet, though you could argue that the “fake news” event was one of the big AI events that’s happened, that, potentially, altered the direction of a presidential race. People are worried about hacking; I’m more worried about attacks that you can’t trace the source of. And I think that’s work to be done, going forward.
There was a weapons system that did make autonomous kill decisions, and the militaries that were evaluating it said, “We need it to have a human in the middle.” So they added that, but of course you can turn that off. It’s almost intractable to define it in such a way.
It sounds like you’re in favor of AI weapons, as long as they’re not buggy.
I’m not in favor of AI weapons. In general, as a person, I’m anti-war. But it’s one of those human frailties and human limitations that war is a necessary—as ugly as it is—part of our lives. I think people and countries will adopt AI and they will start using it for warfare. What is needed, I think, is a new set of agreements and a new set of principles on how they go about using it, much like they do with chemical weapons and nuclear warfare. I don’t think it’s something we can control. What we can do is regulate and manage and enforce it.
So, moving past warfare, do you believe Putin’s statement that he who controls AI in the future will control the world?
Absolutely. I think that’s a given.
Back to jobs for a moment. Working through the examples you gave, it is true that steam and electricity and mechanization destroyed jobs, but, what they didn’t do is cause unemployment. Unemployment in this country, in the US, at least, has been between five and ten percent for two hundred years, other than the Depression, which wasn’t technology’s fault. So, what has happened is, yes, we put all of the elevator operators out of business when we invented the button and you no longer had to have a person, but we never saw a spike in unemployment. Is that what’s going to happen? Because if we really lost eighteen million jobs in the next twenty-seven months, that would just be… That’s massive.
No, but here’s the thing, that eighteen million number is a global number.
Okay, that’s a lot better then. Fair enough, then.
And you have to put this number in context of the total workforce. So today, there are somewhere between seven hundred million to 1.3 billion workers that are employed globally and eighteen million is a fraction of that. That’s number one. Number two, I believe there is a much bigger potential in using AI as a muse, and AI as a partner, to create a whole new class of jobs, rather than be afraid of the machine replacing the job. Machines have always replaced jobs, and they will continue to do that. But I believe—and this is where I get worried about our education system, one of the first things we did with Watson was we started a university program to start skilling people with the next generation skillsets that are needed to deploy and manage AI systems—that over the next decade or, for that matter over the next five decades, there is a whole new class of human creativity and human potential that can and will be unleashed through AI by creating whole new types of job.
If you look at CognitiveScale, we’re somewhere around one hundred and sixty people today. Half of those jobs did not exist four years ago. And many of the people who would have never even considered a job in a tech company are employed by CognitiveScale today. We have linguists who are joining a software company because we have made their job into computational linguistics, where they’re taking what they knew of linguistics, combining it with a machine, and creating a whole new class of applications and systems. We have people who are creating a whole new type of testing mechanisms for AI. These testers never existed before. We have people who are now designing and composing intelligent agents using AI, with skills that they are blending from data science to application development, to machine learning. These are new skills that have come about. Not to mention salespeople, and business strategists, who are coming up with new applications of this. I tend to believe that this is one of the most exciting times—from the point of view of economic growth and jobs—that we, and every country in this world, has in front of them. It all depends on how we commercialize it. One of the great things we have going for the US is a very rich and vibrant venture investment community and a very rich and vibrant stock market that values innovation, not just revenues and profits. As long as we have those, and as long as we have patent coverage and good enforcement of law, I see a very good future for this country.
At the dawn of the Industrial Revolution, there was a debate in this country, in the United States, about the value of post-literacy education. Think about that. Why would most people, who are just going to be farmers, need to go to school after they learn how to read? And then along came some people who said that the jobs of the future, i.e. Industrial Revolution jobs, will require more education. So the US was the first country in the world to guarantee every single person could go to high school, all the way through. So, Mark Cuban said, if he were coming up now, he would study philosophy. He’s the one who said, “The first trillionaires are going to be AI people.” So he’s bullish on this, he said, “I would study philosophy because that’s what you need to know.” If you were to advise young people, what should they study today to be relevant and employable in the future?
I think that’s a great question. I would say, I would study three different things. One, I would study linguistics, literature—soft sciences—things around how decisions are made and how the human mind works, cognitive sciences, things like that. That’s one area. The second thing I would study is business models and how businesses are built and designed and scaled. And the third thing I would study is technology to really understand the art of the possible with these systems. It’s at the intersection of these three things, the creative aspects of design and literature and philosophy around how the human mind works, to the commercial aspect of what to make, and how to build a successful business model, to the technological underpinnings of how to power these business models. I would be focusing on the intersection of those three skills; all embraced under the umbrella of entrepreneurship. I’m very passionate about entrepreneurship. They are the ones who will really lead these country forward, entrepreneurs, both in big companies, and small.
You and I have spoken on the topic of an artificial general intelligence, and you said it was forty or fifty years away, that’s just a number, and that it might require quantum computers. You mentioned Elon and his fear of the existential threat. He believes, evidently, that we’re very close to an AGI and that’s where the fear is. That’s what he’s concerned about. That’s what Hawking is concerned about. You said, “I agree with the concern, if we screw up, it’s an existential threat.” How do you reconcile that with, “I don’t think we’ll have an AGI for forty years”?
Because I think you don’t need an AGI to create an existential threat. There are two different dimensions. You can create an existential threat by just building a highly unreliable autonomous weapons system that doesn’t know anything about general intelligence. It only knows how to seek out and kill. And that, in the wrong hands, could really be the existential threat. You could create a virus on the Internet that could bring down all public utilities and emergency systems, without it having to know anything about general intelligence. If that somehow is released without proper testing or controls, you could bring down economies and societies. You could have devastation, unfortunately, at the scale of what Puerto Rico is now going through without a hurricane going through it; it could be an AI-powered disaster like that. I think these are the kinds of outcomes we have to be aware of. These are the kinds of outcomes we have to start putting rules and guidelines and enforcements around. And that’s an area, that and skills, are the two that I think we are lagging behind significantly today.
The OpenAI initiative is an effort to make AI so that one player doesn’t develop it—in that case an AGI, but all along the way. Do you think that is a good initiative?
Yeah, absolutely. I think OpenAI, we probably need a hundred other initiatives like that, that focus on different aspects of AI. Like what we’re doing at AI Austin, and AI Global. We are focusing on the ethical use of AI. It’s one thing to have a self-driving car, it’s another thing to have a self-driving missile. How do you take a self-driving car that ran over four people, and how do you cross-examine that in a witness box? How is that AI explainable? Who’s responsible for it? So there is a whole new set of ethics and laws that have to be considered when putting this into the intelligent products. Almost like the underwriter labs equivalent of AI that needs to be woven into every product and every process. Those are the things that our governments need to get aware of, and our regulators need to get savvy about, and start implementing.
There is one theory that says that if it’s going to rely on government, that we are all in bad shape because the science will develop faster than the legislative ability to respond to it. Do you have a solution for that?
I think there’s a lot of truth to that, particularly with what we’re seeing recently in the government around technology, there’s a lot of merit to that. I believe, again, the results of what we become and what we use AI for, will be determined by what we do as private citizens, what we do as business leaders, and what we do as philanthropists. One of the beautiful things about America is what philanthropists like Gates and Buffett and all are doing—they’ve got more assets than many countries now, and they’re putting it to work responsibly; like what Cuban’s talking about. So, I do have hope in the great American “heart,” if you may, about innovation, but also responsible application. And I do believe that all of us who are in a position to educate and manage these things, it’s our duty to be able to spread the word, and to be able to lean in, and start helping, and steering this AI towards responsible applications.
Let’s go through your “What AI Isn’t” list, your five things. One of them you said, “An AI is not natural language processing” and obviously, that is true. Do you think, though, the Turing test has any value? If we make a machine that can pass it, is that a benchmark? We have done something extraordinary in that case?
When I was running Watson, I used to believe it had value, but I don’t believe that as much anymore. I think it has limited value in applicability, because of two things. One is, in certain processes where you’re replacing the human brain with a machine, you absolutely need to have some sort of a test to prove or not prove. The more exciting part is not replacement of automated or repetitive human functions, the more exciting part is things that the human brain hasn’t thought of, or hasn’t done. I’ll give you an example: we are working at CognitiveScale with a very large media company, and we were analyzing Super Bowl TV ads, by letting an AI read the video ad, to find out exactly what kinds of creative—is it kids or puppies or celebrities—and at what time, would have the most impact on creating the best TV ad. And what was fascinating was that we just let the AI run at it; we didn’t tell it what to look for. There was no Turing test to say, “This is good or bad.” And the stuff the AI came back with were things that were ten or twelve levels deep in terms of connections it found, things that a human brain normally would have never thought about. And we still can’t describe why there is a connection to it.
It’s stuff like that—the absolute reference is not the human brain, this is the “unknown unknown” part I talked about—that with AI, you can emulate human cognition but, as importantly, with AI you can extend human cognition. The extension part of coming up with patterns or insights and decisions that the human brain may not have used, I think that’s the exciting part of AI. We find when we do projects with customers that there are patterns that we can’t explain, as a human being, why it is, but there’s a strong correlation; it’s eighteen levels deep and it’s buried in there, but it’s a strong correlator. So, I kind of put this into two buckets: first is low-level repetitive tasks that AI can replace; and second is a whole new class of learning that extends human cognition where—this is the unsupervised learning bit—where you start putting a human in the loop to really figure out and learn new ways of doing business. And I think they are both aspects that we need to be cognizant of, and not just try to emulate the current human brain which has, in many cases, proven to be very inefficient in making good decisions.
You have an enormous amount of optimism about it. You’re probably the most optimistic person, that I’ve spoken to, about how far we can get without a general intelligence. But, of course, you keep using words like “existential threat,” you keep identifying concepts like a virus that takes down the electrical grid, warfare, and all of that; you even used “rogue AI” in the context of a business. In that latter case, how would a rogue AI destroy a business? And you can’t legislate your way around that, right? So, give me a example of a rogue AI in an enterprise scenario.
There are so many of them. One of them actually happened when we recently met with a large financial institution. We were sitting and having a meeting, and suddenly we found out that that particular company was going through a massive disruption of business operations because all of their x-number of data centers were shutting down, every 20 minutes or so, and rebooting themselves; all over the world, their data centers were shutting down and rebooting. They were panicking because this was during the middle of a business day, there were billions of dollars being transacted, and they had no idea why these data centers were doing what they were. A few hours into it, they found out that someone wrote a security bot last month, and they launched it into the cloud system that they have, and for some reason, that agent—that AI—felt that it was a good idea to start shutting down these systems every 20 minutes and rebooting it. That was a simple example of how, they finally found it, but there was no visibility in governance of that particular AI that was introduced. That’s one of the reasons we talked about the ability to have a framework for managing visibility and control of these AIs.
The other one could be—and this has not happened yet, but this is one of the threats—you look at underwriting. An insurance company uses technology today a lot, to start underwriting risks. And if, for whatever reason, you have an AI system that sees correlations and patterns, but has not been trained well enough on really understanding risk, you could pretty much have the entire business wiped out. By having the AI—if you depend on it too much without explainability and trust—suggesting you take on risks, that will put your business at an existential risk.
I can go on and on, and I can use examples around cancer, around diabetes, around anything to do with commerce where AI is going to be put to use. I believe as we move forward with AI, the two phrases that are going to become incredibly important for enterprises are “lifecycle management of an AI,” and “responsible AI.” And I think that’s where there’s a tremendous amount of opportunity. That’s why I’m excited about what we’re doing at CognitiveScale to enable those systems.
Two final questions. So, with those scenarios, give me the other side, give us some success stories you’ve personally seen. They can be CognitiveScale or other ones, that you’ve seen have a really positive impact on a business.
I think there are many of them. I’ll pick an area in retail, something as simple as retail, where through an AI we were able to demonstrate how a rules-based system—so this particular large retailer used to have a mobile app where they presented to you a shirt, and trousers, and some accessories and it was like a Tinder or “hot or not” type of a game—and the rules-based system, on average, were getting less than ten percent conversion on what people said they liked. Those were all systems that are not learning. Then we put an AI behind it, and that AI could understand that that particular dress was an off-shoulder dress, and it was a teal color, and it was pairs with an open-toe shoe that’s a shiny leather. As the customers started engaging with it, the AI started personalizing the output, and we demonstrated a twenty-four percent conversion compared to a single-digit conversions, in a matter of seven months. And here’s the beautiful part, every month the AI is getting smarter and smarter, and every percentage conversion equals tens of millions of dollars in top-line growth. So that’s one example of a digital brain, a cognitive digital brain, driving shopper engagement and shopper conversion.
The other thing we saw was in the case of pediatric asthma. How an AI can help nurses do a much better job of preventing children from having an asthma attack, because the AI is able to read a tweet from that says there will be a ragweed outbreak on Thursday morning. The AI understands the zip code that it’s talking about, and Thursday is four days out, and there are seventeen children with a risk of ragweed or similar allergies; and it starts tapping the nurse on the shoulder and saying, “There is an ‘unknown unknown’ going on here which is, four days from now there will be a ragweed outbreak, you better get proactive about it and start addressing the kids.” So, there’s an example in healthcare.
There are examples in wealth management, and financial services, around compliance and how we’re using AI to improve compliance. There are examples of how we are changing the dynamics of trading, foreign exchange trading, and how a trader does equities and derivatives trading by the AI guiding them through a chat session where the AI is listening in and guiding them as to what to do. The examples are many, and most of them are things that are written up in case studies, but this is just the beginning. I think this is going to be one of the most exciting innovations that will transform the landscape of businesses over the next five to seven years.
You’re totally right about the product recommendation. I was on Amazon and I bought something, it was a book or something, and it said, “Do you want these salt-and-pepper-shaker robots that you wind up and they walk across the table?” And I was like, “Yes, I do!” But it had nothing to do with the thing that I was buying.
Final question, you’ve talked about Hollywood setting the narrative for AI. You’ve mentioned I, Robot in passing. Are you a consumer of science fiction, and, if so, what vision of the future—book or whatever—do you think, “Aha, that’s really cool, that could happen,” or what have you?
Well, I think probably the closest vision I would have is to Gene Roddenberry, and Star Trek. I think that’s pretty much a great example of a data quarter helping a human being make a better decision—a flight deck, a holodeck, that is helping you steer. It’s still the human, being augmented. It’s still the human making the decisions around empathy, courage, and ethics. And I think that’s the world that AI is going to take us to; the world of augmented intelligence. Where we are being enabled to do much bigger and greater things, and not just a world of artificial intelligence where all our jobs are removed and we are nothing but plastic blobs sitting in a chair.
Roddenberry said that in the twenty-third century there will be no hunger, and there will be no greed, and all the children will know how to read. Do you believe that?
If I had a chance to live to be twice or three times my age, that would be what I’d come in to do. After CognitiveScale, that is going to be my mission through my foundation. Most of my money I’ve donated to my foundation, and it will be focused on AI for good; around addressing problems of education, around addressing problems of environment, and around addressing problems of conflict.
I do believe that’s the most exciting frontier where AI will be applied. And there will be a lot of mishaps along the way, but I do believe, as a race and as a humanity, if we make the right decisions, that is the endpoint that we will reach. I don’t know if it’s 2300, but, certainly, it’s something that I think we will get to.
Thank you for a fascinating hour.
Thank you very much.
It was really extraordinary and I appreciate the time.
Thanks, Byron.
Byron explores issues around artificial intelligence and conscious computers in his upcoming book The Fourth Age, to be published in April by Atria, an imprint of Simon & Schuster. Pre-order a copy here

Voices in AI – Episode 10: A Conversation with Suchi Saria

In this episode, Byron and Suchi talk about understanding, data, medicine, and waste.
[podcast_player name=”Episode 10: A Conversation with Suchi Saria” artist=”Byron Reese” album=”Voices in AI” url=”″ cover_art_url=””]
Byron Reese: This is “Voices in AI” brought to you by Gigaom. I am Byron Reese. Today, my guest is Suchi Saria. Where do I start when going through her career? She has an undergraduate degree in both computer science and physics. She has a PhD in computer science from Stanford, where she studied under Daphne Koller. She interned as a researcher at IBM and at Microsoft Research, where she worked with Eric Horvitz. She is an NSF Computing Innovation fellow at Harvard, a DARPA Young Faculty Award Winner, and is presently a professor at Johns Hopkins. Welcome to the show.
Suchi Saria: Thank you.
Let’s start off with the biggest, highest-level question there is. What is artificial intelligence? How do you answer that question when it’s posed to you?
That’s a great question. I think AI means very different things to different people. And I think experts in the field, at the high level, understand to a degree of what AI is; but they never really posit a very concrete, mathematical description of it. Overall, our goal is… We want computers to be able to behave intelligently, and that’s really the origin of how the AI field of computer science emerged.
Now along the way, what has happened is… Starting from really classical applications— autonomous driving or image recognition or diagnostics… As lots and lots of data has been collected, people have started to develop numerical tools, or statistical methods, or computational methods that allow us to leverage this data, to build computers or machines that can do useful things that can help humans along the way. And so that then has also become part of AI.
Effectively, the question that as a field we often ask ourselves is: Does AI really mean useful tools that help humans, automating a task that humans can do, giving computers the ability to also do? Or is it going at properties like creativity and emotion, that are very interesting and unique aspects of what humans often exhibit? And do computers have to exhibit that to be considered ‘artificially intelligent’? So really there’s a debate about what is intelligence, and what does it really mean; and I think different experts in these fields have very different takes on it.
I only ask it because it’s a really strange question. If you ask somebody at NASA, “What is space travel?”, maybe there’s… “Where does space begin?” or “How many miles up?”—If you ask all these different fields, they kind of know what the field is about. You said something I never heard anybody say, which is: “Those of us who are researchers in it, we have a consensus on what it is.”
I would say at a very high level, we all agree. At a high level, it is the ability to systematize or help computers behave and reason intelligently. The part that is left to be agreed upon is ‘behave and reason intelligently the way humans do’. The way humans do things is important because, in some fields, we should study humans; we should understand the way humans do it and biological systems do it, and then build computers to do it the way humans do it.
In other fields, it’s not so important that we do it exactly the way humans do it. Computers have their own strengths, and effectively, perhaps what’s more important is the ability to do something, rather than the process by which we’re getting there. So we all agree that the goal is to build intelligent machines.
Intelligent machines that crunch a lot of data, intelligent machines that can reason through information that’s provided, produce what needs to be done, interact intelligently—and by that, we mean understand the person that’s in front of you, and understand the scenario that’s being presented to you, and react appropriately.
Those are all things we’ll agree on. And then, effectively, the question is: Do we need to do it the way humans are doing it? In other words, is it in the making of human intelligence, or is it about giving this capability to machines by whichever way the machines are able to learn that?
I won’t spend too much time here, because it may not be interesting to everyone else, but to say artificial intelligence is teaching machines to reason intelligently—you’re using ‘to reason intelligently’ to define the term ‘intelligence’. Doesn’t that all obfuscate what intelligence is?
Because at one extreme end, it’s defined simply as something that reacts towards its environment; a sprinkler system that comes on when the grass is dry is intelligent. On another extreme end, it’s something that learns, teaches itself; it evolves in a way that your sprinkler system doesn’t. It’s a learning system that changes its programming as it’s given more data.
Isn’t there some element of what intelligence is that we all have to circle around, if we are going to use this term? And if we’re not going to circle around it, is there a preferred way to refer to this technology?
Yeah, I think the preferred way is the way we think about it. I think the other aspect of the field that I really love is the fact that it’s very inclusive. The reason the field has moved forward so quickly is because, as a field, we’ve been very inclusive of ideas from psychology, from physics, from neuroscience, from statistics, from mathematics—and of course, computer science.
And what this really means is as a field, we move forward really quickly and there’s really room for multiplicity of opinions and ideas. The way I often think about it is: What’s the preferred way people like me think about it—and others might give you different opinions about it, but fundamental to all this is the idea of learning.
Rather than building brittle systems that effectively have hard-coded logic which says, “If this happens, then do this. If that happens, then do this”—what’s different here is that effectively these systems are more designed to program their own logic, based upon data. They’re learning in a variety of different ways—they learn from data. Data where in the past, people have presented a scenario.
Let’s say in this scenario, you might consider how another intelligent human or an expert human is reacting to the scenario, and you’re watching how the human behaves or reacts; and from that, the computer is trying to learn what is optimal. Alternatively, they may learn by interacting with their environment itself. For instance, if the environment has a way… Like in the game of Go, the environment here being the board game itself—had a way of giving feedback… A version of feedback would be, if you make a move, you get a score attached to whether or not this is a good move, and whether or not it will help you win, and they’re basically using that feedback.
It’s often the type of feedback we as humans use all the time in real life… Where effectively you could imagine kids… If there’s a pot that’s too hot and they touch it, next time they see a similar object, they’re much less likely to touch it. And you know as adults, we go and we often analyze scenarios around us, and see if something has a positive or a negative feedback.
And then, when we see negative feedback, we sort of register what might have caused it, the reason about what might have caused it, and try to do that process. The notion of learning is pretty fundamental. The way by which it learns is really a huge body of work which has focused on that—which is, how do we develop more general purpose methods by which computers can learn, and learn from many different types of data, many different types of supervision and effectively, learn as quickly as possible?
You used the word ‘understand’ the person, ‘understand’ the situation. There’s a famous thought experiment on that word, and what the implications are. It’s called The Chinese Room Problem, and just to set it up for the listener… There’s a man who speaks no Chinese—we call him the librarian—and he’s in this giant room with thousands and thousands of these very special books.
And people slide questions under the door to him, and they’re written in Chinese. He doesn’t understand them, but he knows to match the very first symbol in the message to the spine of a book, pulls that book down, looks up the second symbol that directs him to another book, and another one, and another one… Until he finally gets to the end of this process.
And he copies down the characters that he sees, slides that back out, and it’s a perfect answer in Chinese. And of course, the man doesn’t know the meaning of what it was about, but he was able to produce this perfect answer using a system. The question is: Does the man understand Chinese?
And of course, the analogy is obvious. That’s all a computer is doing; it’s running a deterministic program, and so forth. So I put the question to you: Does the man understand Chinese? Can a computer understand something, or is understanding just a convenient word we use, but [where], clearly, the computer doesn’t understand anything?
Let’s shift our attention for a second away from computers and to humans. I often think hard about… I try to pull out scenarios where I’m wondering, am I effectively running an algorithm? And what is my own algorithm?—and even considering scenarios where it’s not so prescriptive. Perhaps I needed to be creative.
My job often involves being creative, and coming up with new ideas frequently. And the question I ask myself is: Am I just deriving this idea out of previous experiences that I already had? In other words, am I effectively just engaging in the task of…
Let’s say I have A and B, and this really creative idea… But what my brain has become really good at is, in new scenarios, quickly figuring out what are the relevant elements like A, B and C in my past which are pertinent, and then from that, coming up with something that looks like a combination or variation. In other words, it’s not as big a leap of faith as it [might seem] to someone who doesn’t have my experience, or doesn’t have my background.
And then, I think hard about it, and perhaps it really is just derived from the things I know. What this is getting at is me being a little cynical about my own ability to—my own assessment of how much do I really understand. Is understanding effectively the ability to quickly parse information, determine what’s important, apply rules of logic and a bit of randomness in order to experiment with ideas and then come up with a new idea?
I don’t really have an answer to this. But I’ve often wondered, this is maybe what we do; and our ability to do this really rather quickly is sort of what distinguishes different humans in their ability to understand and come up with a creative idea quickly. And so, if I think about it from this point of view, it doesn’t seem to me a complete stretch to imagine that we could teach computers to do these things.
So let me give you an example. For instance, going back to the very popular news story around AlphaGo, when the AlphaGo started to explore new moves. Many individuals who are not familiar with the topic of AI thought, “Wow, that’s amazing! It’s being creative, it’s coming up with brand-new moves altogether”—that humans, human experts hadn’t really known. But really, all it was doing is doing search, in some super large space.
And its ability to do search is pretty expansive. And the other thing it has is really clever ways of doing search, because it has heuristics that is has built up from its own experience of doing learning. And so, in a way, that’s really what humans are doing, and that’s really what experience gives us. So let’s go back to your question of: Does the [Chinese Room person] really understand?
I think my personal issue is that I don’t know what understanding really means. And the example I gave you… If you were to define understanding that way, then I think in today’s world, we would say maybe that man didn’t understand what he was doing, but maybe he did. I’m not sure. It’s not obvious to me.
Do we measure understanding by the output—the fact that you give an input and they give a reasonable output? Or do we measure it by some other metrics? It’s a really great question though.
You captured the whole debate in what you just said, which is… The room passes the Turing test. You wouldn’t be able to tell—if you’re the Chinese speaker outside the room passing in the messages—you wouldn’t be able to tell if that was not a native speaker on the other side.
And so the machine ‘thinks’. Many people in the field had no problem saying the man understands Chinese, but at a gut level, that doesn’t feel right. Because the man doesn’t know if that was a message about cholera or coffee beans, or what—it’s just blinds on paper. He knows nothing, understands nothing, just walks through some old thing that gets him to copy these marks down on paper.
And to say that is understanding trips people up. The question is: Is that the limit of what a machine will ever be able to do? I will only say one thing, and then I would love your thoughts. Garry Kasparov kind of captured that when he lost to Deep Blue back in ‘97. He said, “Well, at least it didn’t enjoy beating me.”
His experience of the game was different from the computer’s experience of the game. I only think it’s a meaningful question because it really is trying to address the limits of what we can get machines to do. And if, in fact, we don’t understand anything either, then that does imply we can build AGI and so forth.
I agree with you. I think it’s a very meaningful question. And I certainly think it’s a topic we should continue to push on and understand more deeply. I would even go back, to say that I bet there are people around you—maybe not as holistic and expansive a context as the Chinese man you described—but you could imagine scenarios where somebody is really good… their whole job is sort of like, they learned numerous algorithms like this.
And you could imagine colleagues like that… Where they’re effectively really good at fielding certain types of questions, and pushing data out. And maybe they have not built the algorithm, but they understand what the person in front of them is asking, and they understand what kinds of answers they need to hear in order to be able to answer questions in a satisfactory manner.
Effectively, my point is even though we think that in the example… [If] somebody told you he doesn’t understand, that [conclusion] is very possible. If nobody had told you that, and he always was able to produce something that was acceptable or of a high quality, everybody else would always think of this person as, “He understands what he’s doing.” And we probably have people like that around us. We’ve all experienced this to some extent.
It could be. And If that is the case, it really boils down to what the word ‘artificial’ means in ‘artificial intelligence’. If ‘artificial’ means it’s not really intelligence—like artificial turf isn’t really turf —if it really means that, then you’re right. As long as you don’t know that he doesn’t understand, it doesn’t really matter.
I would love to ask one more question along these lines. I’m really intrigued by what we will need to do to build a machine that is equivalent to a human; and I think your approach of, “Let’s start with what humans do and talk about computers later” is really smart.
So I would put this to you… Humans are sentient, which is often a word that is misused to mean intelligent. [But] that’s actually ‘sapient’. ‘Sentient’ means you’re able to feel things—usually pain—but you’re able to feel something, to have an experience of feeling something.
That’s kind of also wrapped up in consciousness, but we won’t talk there yet…
Is it possible for a computer to ever feel anything? It’s clearly possible to set up a temperature sensor that, when you hold a match to it, the computer can sense the temperature; and you can program the computer to scream in agony when it passes a certain temperature. But would it ever be possible for a computer to feel pain, or feel anything?
Let’s step back and ask the following question… Two parts: First is, “To make computers that feel something—can that be done?” The second question is, “Why do we need computers that feel things?” Is that really what separates artificial intelligence from human intelligence?
In other words, is that really the key distinction? And if so, can that be built? Let’s talk about how do we build it. Have you heard, or have you seen, any of the demos out of this terrific company—I think it’s called Hanson Robotics. If you go online, you can Google it, you can search for it. David Hanson is one of the founders, and effectively, what they build is a way to give a robot a face; and he has these actuators that allow very fine-grained movement.
And so, effectively, you see full facial features and full facial expressions projected onto a robot. The robot can smile and the robot can frown, and it can get angry and it can stare and express excitement and joy. Effectively, he’s sort of done a lot of the work of—not just what it takes to build mechanically those parts, but also thinking harder about how it would get expressed, and a little bit about when it would get expressed.
And then independently, there’s great work from MIT—and you know, other labs, too—but I’m just thinking of one example: They looked at learning and interpreting emotion. For example, you might imagine [that] if the person in front of you is angry, you might want the robot to react and respond differently than if the person was happy and excited.
Effectively, you could imagine putting a camera, seeing the stream coming in, [and] the computer processes it to do classification for whatever type of emotion is being expressed—you could specify a list of emotions that are commonly expressed. From that, the computer can then decide what human emotion is being expressed, and then decide what emotion it wants to express.
And now, you can imagine feeding it back into Hanson’s program that allows them to generate robotic facial motions that are effectively expressing emotion, right? So if we had to build it, we could build it. We know how to think about building it. So mechanically, it is not impossible. So now the piece here is—the second question is: If we could do this, and in fact there are studies that…
For instance, when I was with Microsoft Research, there was a robot that would greet you, and it would basically see where you were standing, and it would turn its head to try to point to you. And many, many individuals who weren’t familiar with robotics—many visitors who would come to Microsoft, people that weren’t in the technology industry, but were just visiting—would see that and get really excited, because the idea of a robot turning its head and moving its eyes in response to where you’re standing was cool, and seemed very intelligent.
But effectively, if you break down the mechanics of how it’s doing it, it’s not a big surprise. Similarly, you could augment it by also showing facial expressions, and I think CMU— Carnegie Mellon—has a beautiful robot that’s called the robot receptionist; her name is Valerie. They worked on it at the drama department at Carnegie Mellon.
And they basically filled the robot with lots of stories, and it was really funny… As a graduate student, I was visiting, and met Valerie for the first time… You could ask her for directions, and she would give you directions on where to go. If I could say, “Where’s Manuela’s office?” the robot would point me to where it is.
But in the middle, she would behave like a human, where she would be talking on the phone to her sister; and they’d be talking about what’s going on, what’s been keeping them busy, and they’d hang up or she’d put people on hold if a new visitor came in, and so forth.
So what I’m challenging is this concept of, is it really the lack of human emotion, or what you consider to be human-like emotion—to be very special to humans? Is it that? Is it mimicking that? What does it mean to feel pain? Is it really the action-reaction—somebody’s poking you and you react—or is it the fact that there’s something internal, biological that’s going on, and it’s the perception of that?
That could be. You asked a good question: Does it matter? And there would be three possible reasons it would matter: First, there are those that would maintain that an intelligence has to experience the world, that it isn’t just this abstract ones and zeros it-lives-in-a-computer thing—that a true intelligence would need to be able to actually have experiences.
The second thing that might make it matter is… There was a man named Weizenbaum who famously created a program in the ‘60s called ELIZA, which was a really simple program. You would say, “I’m sad.” It would say, “Why are you sad?”
“I’m sad because my brother yelled at me.”
“Why did your brother yell at you?”
And Weizenbaum turned against it all, because what he saw is that even people who knew it was just a very simple program developed emotional attachment to it. And he said… When the computer says, “I understand,” as Eliza did, he said it’s just a lie. There is no ‘I’ in there, and there’s no understanding.
But really the reason why it might actually matter is another thought experiment, that I will put to you and to those listening: It’s the problem of Mary.
Mary is a hypothetical person who knows everything about color. She knows literally everything, like at a god-like level. She knows everything about photons and cones and how color manifests in the brain. She knows everything that there is to know about it, but the setup is that she has never seen it. She lives in this room that’s all black and white, and only has black-and-white computer monitors.
She walks outside one day and sees red for the first time. And the question is: Did she learn something new? Is experiencing something different than knowing something? And if you say yes… It’s one of those things where most people, at first glance, would say, “Yes, if she’s never seen color and she sees it for the first time; yes, she learns something.”
And if that is the case, then a computer has to be able to experience things in order to learn past a certain point. Do you think Mary learned something new when she saw color for the first time? Or no, she knew exactly what it would look like, and experiencing it would make no difference?
So, you know what Mary knew. Did she know ahead of time what red would look like when she stepped out?
Well, she knew everything about color. She never saw it, but she knew exactly what it would do to her brain—at a molecular level, atomic level—every single thing that would happen in her brain when she saw a color, but she’s never seen it.
As a computer scientist, when you say that you me, I would say that the representation of what Mary understands or know is ambiguous. What I mean by this is, I don’t know what it means to say—I understand what it means to say “she knows at the molecular level what happens.” I understand what it means to say she knows, perhaps, about the relationship between different primary colors, and the derivative colors and so forth.
But are you saying that she knows… Is it the case that she receives an image using her eyes, and her eyes represent it using some form of internal neuronal format?—Are you saying she knows that? Because if she doesn’t know that, then effectively, she still has a partial understanding of what knowing everything about color means.
So this might be an interesting place… Where we think her knowing everything about color…
If you tell me: Somebody presented a red image to her, and she knew what it meant to take that red image and convert it—and these are really hypotheticals; I’d have to understand this more deeply and really study it, and perhaps bring in someone who understands human perception really well—but my first step-check would be: What does it mean for her to know everything about color?
And what if we present her with an image, her visual cortex processes it, and effectively, she is getting data, and she is seeing it internally. Is it stored in RGB format? Is she storing it in some format that she understands? Is she aware? Has that core process happened in her head before? It may not have been due to her stepping out, but the question is: Is that something that she is privy to, or has knowledge of?
And if so, then I would say that when she steps out… And if all she is doing is focusing on the color red, and that is the only sensation that’s being generated in her head; then yeah, this is going to seem familiar to her because it’s something she’s seen before. The word ‘experience’ at that point is a really interesting word. And it would be fun to sit down and try to write down formal definitions for what it means.
And generally, we think of having ‘seen’ and having ‘experienced’ as two different things, in human emotions. But I think from a computer point of view, they don’t seem different. Even as a human, if I think hard about it, I don’t know really what the distinction is. I don’t know what it means to kind of know it, to know it, and then experience it. What is the difference between those things?
It may be that the question imperfectly captures it, because it’s formed very casually, but… Humans experience the world.
You taste a pineapple, and what that pineapple tastes like… Tasting it seems to be a different thing than knowing something. If I know what it tastes like, it’s a different thing than actually having the experience of tasting it.
Knowing how to ride a bicycle is different than having ridden a bicycle, and knowing how you feel balanced when you get on one. Touching something warm feels a certain way that knowing all about warmth does not capture.
And so, the question is: If a machine cannot actually feel things, touch things, taste things, have any experience of the world—then whatever intelligence it has is truly fake. It really is artificial in a sense that’s completely fake.
And you’re right, I think, in asking the question… Why we ask these questions… And a lot of what people are often doing is asking questions about people. Are people machines? Are we…
But then they have this disconnect, to say: “But we feel, and we experience, and we know, and those seem to be different than things my iPhone can do.” So I think I’m trying to connect those dots to say, experiencing something seems to be different than knowing something.
But you’re right; it’s imperfectly formed. I’ll let you comment on that, and then let’s move on to your research, because there’s so much there I would love to hear more about.
Sure! So I think I am going to continue to push back a little bit on… I feel that people’s experience of what they believe a machine or an iPhone can do is very much based on… I think it’s easier to think about a single narrow task.
You could take the task of eating a pineapple, or the task of going and experiencing a warm day… But effectively, the way I think about it is [that] a lot of these capabilities don’t exist because most people haven’t thought that building a machine that eats a pineapple is a very useful thing, so people haven’t bothered to build it.
But let’s imagine I decided that was important, and I wanted to build it. Then, what I would do is much like—going back to David Hanson… I would try to first identify what do I mean by ‘experience eating a pineapple’, and if the idea is that every time I am given it—a tasty pineapple—I can eat it and it’s delicious, and my eyes light up. And if I eat a rotten pineapple, then I’m visibly upset.
Then I could imagine building the sensor to which you feed the pineapple. It runs chemical tests that check, effectively, what’s in the pineapple and… You could start by version one. Version one tests what’s in the pineapple, and based on that—and it’s hooked up to David Hanson’s robot—and it generates the reaction, which is excited, or sad, or unhappy, and visibly unhappy, or sad, depending on how tasty or not-so-tasty the pineapple is.
And you could even take it a step further by saying, “You know what? I’m going to give lots of humans things to eat; and based on that, I will watch what the humans are doing. And then effectively, the computer’s just learning by taking the same fruit and eating it itself. And you didn’t even program anything about how to react. All it did was watch humans eat it, and based on that, it learned that when certain molecular compositions exist in the thing it’s tasting, then it tends to get happy or less happy.
And you might imagine it starts to mimic. In fact, we could take it even another step further and say, “Let’s give a group of robots the same set of sensors, and they have to figure out a way by which they communicate and barter with each other.” So effectively, there’s an objective function, and the objective function—or the goal for the group of robots—is to figure out an effective way to trade.
The trade is such that one group of robots loves apples. The other group of robots loves pineapples. And the way you know that is, effectively, they’ve each lived in different environments and—I don’t like the word ‘live’, because it’s over-interpretive…
What I mean is, they’ve been trained in different environments, and the ones that love to eat apples have learned to get an excited expression to good apples, and the other set of robots get an excited expression to good pineapples. And you want them to work together to trade, such that everybody is as happy as possible.
Then it’s completely possible they’ll be able to effectively learn, on their own, a trading strategy where they say, “You know what? The people who don’t like pineapples should give away their pineapples, and the people who don’t like apples should get rid of apples.” So, effectively, what I was giving you was an example where…
If we understand what is the objective we’re after—which is, what does experiencing a pineapple mean—then very often, you can turn it into some mathematical objective by which the computer can learn how to do similar things, and very quickly… ‘Very quickly’ depends a lot on the complexity of the task—but it can mimic that behavior or goal—and now I use the word ‘mimic’ lightly…
But effectively it can, be it similarly, or—and one could argue, “What does ‘similar’ mean?” and, “What does ‘behave similarly’ mean?”… But for the most part, we would look at this and be pretty satisfied that it’s doing something that we would consider to be intelligent. We would consider it to be experiencing something.
Unless, the only block in our head is we think it’s a machine… So it’s hard because we think humans experience things and machines don’t… But what I think would be really cool is to think about, “Are there tasks where we really experience something, that we think there is no way to build a machine to experience the same thing?” What does it mean to experience in that setup?
I think that would be interesting, and I would love to hear [from] our listeners who have ideas, or want to send me ideas. I would love to hear that!
Well, I think the challenge, though, is that in civilization we’ve developed something called ‘human rights’, where we say: “There are things you can’t do to a person no matter what. You can’t torture people for amusement, and you can’t do these things.”
So we have human rights, and we extend them—actually broadly—to other creatures that can feel pain, so we have laws against cruelty to animals because they feel pain.
It sounds like you’re saying the minute you program a computer to be able to mimic a frown, or to scream and mimic agony, that that is somehow an equivalency; and therefore, we need laws that… Once the temperature hits 480 degrees, the computer screams, and we need to outlaw that; we need to grant those things rights, because they are experiencing things.
And then, you would push it one step further to say, when I am trying to get my car out of the mud, and it’s smoking, and the gears are grinding… That that too is experiencing pain, and therefore that should be…
You run into one of two risks. You would either make the notion of, “Things that feel have rights not to be tortured”—you either make that ludicrous, by applying it to anything that can make a frowny face…
You either try to elevate everything that’s mechanical to that, or you end up debasing people, by saying: “You don’t actually feel anything. That’s just a program. You’re a machine, and you don’t actually have any experience. And you reporting pain, it isn’t real. It’s just you, kind of, programmed to say that.”
How do you have rights in a world where you have that reductionist view of experience?
Personally, I think it’s pretty liberating that computers don’t get tired, and they don’t feel pain. When I say the word ‘feel pain’, I mean feel pain in the sense that, if you ‘hurt’ me a lot in a certain way using a pin, I may screech. But also, I could shut down, I could stop being productive.
But if you take a computer, and it has a hard, metal shell… And you take a pin and you effectively poke it too hard, it doesn’t really do much to the computer because it’s fine.
But then, there are other things… For instance, if you unplug the computer, it’s dead. And there’s an equivalent notion of unplugging me. So for me, I kind of find it liberating that we don’t have to try to do all the same things. The thing that is very exciting to me about it is that this has its own strength. A machine is effectively a very… I think there’s two takeaways for me, personally: One, the fact that it makes me think harder about, “What do I have to do to be special?”—about myself.
So effectively, there are lots of things that I used to consider to be very special—I’m still special, of course [laughs]—but what I mean is, I would attribute this mystical sense to—which is maybe not so necessary… Like the whole task of programming computers and developing these learning machines has really made me a little bit more humble about what I consider to be very hard, and not-so-hard; and effectively realizing that maybe some of these properties that humans exhibit can actually be demystified, right?
I understand a little bit more about, what does it mean to do X and do Y? It makes me think harder about something that comes so naturally to us—how is that we do it? How is it that different beings do it? And the fact that computers can do it, and maybe it’s not exactly the same way, and it’s a slightly different way…
So just having that awareness is actually pretty exciting, because it makes things that are everyday around us, which are pretty rote… not so rote anymore. It’s fun to watch people walk. You’re sort of saying, “Ah, it’s so natural and easy for them,” but if you really think about it, there are just so many complicated things we are doing. And then, you try to make and teach computers how to walk, you sort of very quickly realize how complicated it is, and it’s kind of cool that we as human beings can do it.
So effectively, one of the aspects of it is it teaching me a little bit more about myself, and realizing the complexity and also the steps or procedures it takes for me to do some of the things that I’m doing. The second aspect of it is realizing that perhaps it’s a good thing [that] there are certain things a computer is good at, and things that it’s not good at… And perhaps, taking advantage of that in order to build systems that are useful in practice, and can really make us, as a society, better off is pretty exciting to me.
So I think the idea of trying to exactly mimic humans—or whether we would be able to exactly mimic humans—is sort of interesting; but practically speaking, I don’t think of it as the most interesting consequence of this… or the area of debate for most experts in the field.
We think more of it as, what are areas where we can really build useful things that could then help us make humans faster, make everyday life better, save us work—that would be better to pass off to a computer to do, so that it frees up time for us to do other things.
But… Does that answer your question a little bit more, about human rights? So effectively, I think the issue was, if you are concerned about pain, then perhaps there should be rules about when humans experience pain, we ought not to do X, Y and Z. Maybe computers could have different sorts of rules, because they experience different sorts of things, and they’re good and bad at different sorts of things.
And I think we just haven’t come to a place where there’s a general agreement among scientists building it, about what is and isn’t useful, and we work around those principles. And that has really dictated what gets built.
Fair enough! So tell me about… You have an unusual professorship at Johns Hopkins. What is that? Can you talk about your work there?
Yeah, sure! I’m a faculty in Computer Science and Stats, but also, I’m a faculty in Public Health. Hopkins is one of the largest schools of public health in the country; and in particular, I am in the department of Health Policy and Management. So what’s unique about my appointment is that…
Hopkins has a very large School of Public Health, a very large School of Medicine. And I effectively interact—on a day-to-day basis—not just with engineers, but also people who are clinical experts and public health experts who design policy… [Which brings] a multifaceted view into the kinds of questions we’re trying to answer around using computers, and using data-driven tools to improve medicine and improve public health.
And so, what does that look like on a day-to-day basis? What kinds of projects are you working on?
Let’s see… Let me give you a concrete example.
One area of study that we spend time on is detecting adverse events in hospitals. They’re called ‘hospital-acquired complications’. One example of this is sepsis. And effectively, what happens is, let’s say a patient is coming into the hospital for any condition; and sometimes they come in because they have an infection, and the infection goes undetected, and turns into what’s called sepsis.
Sepsis is effectively when your body is trying to fight the infection, [and] it releases chemicals, and these chemicals start attacking your [own] organs and systems. This has happened in some fraction of the cases, and if it does happen, it ends up causing organ damage, organ failure, and eventually death if it goes untreated.
And so, this is an example where individuals who have sepsis at the moment… Physicians are relying on visible signs and symptoms in the patient in order to be able to initiate treatment. And what our last work has shown is [that] it’s possible to identify very early, based on lots of data… So when they come in, as part of routine care, they’re taking tons of measurements, and these measurements are getting stored electronically…
And so, what we do is we analyze these measurements in real-time, and we can identify subtle signs and symptoms that currently the physicians miss all the—you know, it’s a busy unit. In a 400-bed hospital, there’s persons coming in, there are lots of other patients; it’s a distributed care team. It’s tough. And if the symptoms are not really visible, or are subtle, they sometimes get missed.
And so, an example area where we’ve shown is—with sepsis, for instance—you can identify very early, subtle signs and symptoms, and identify these high-risk patients and bring this to the caregiver; so that they can now start to initiate treatment faster. And so, this is exciting because it really demonstrates the power of computers: They’re tireless; they can sit there, process data from 400 patients continuously, all the time.
We can learn from expert doctors what are signs and symptoms, but not just that! We can look at retrospective data from 10,000 or 70,000 or 100,000 patients, and understand things like what are the subtle signs and symptoms that happen to appear in patients with sepsis and without sepsis, and use that to start displaying this kind of information to physicians.
And now, they’re better off, because suddenly, they are missing fewer patients. The patients are better off because they can go in completely happy that they’re going to be cared for in the best way possible, and the computer is sitting there, and it really has no reason to complain because all it’s doing is processing the data, and it’s good at that. So that’s one example. And there are lots of other areas.
Another area we’ve been spending time looking at is complex patients, patients of… the word ‘complex patients’ is a little… Let me demystify that a little bit. So looking at diseases where there’s a ton of diversity or heterogeneity in symptom profile; so for example diseases like lupus, scleroderma, multiple sclerosis, where the signs and symptoms vary a lot across individuals. And really understanding which person is going to be responsive to which treatment [in these cases] is not so obvious.
So again, going back to the same philosophy: If we can take data from a large patient population, we can analyze this and start to learn what—for a given patient—is their typical course going to look like, and what are they going to be likely to be responsive to. And then [we can] use that to start bringing that information back to our physicians at the point of… They can now use this information to improve and guide their own care. So those are some examples.
I was just reading some analysis which was saying that before World War II, doctors only had five medicines. They had quinine for malaria, they had aspirin for inflammation, and they had morphine for pain… They had five medicines, and then, you think about where we are today. And that gives one a lot of hope.
And then you think about… We kind of have a few challenges. I mean even all the costs, and all the infrastructure and all of that, just treating it as a mental problem… One, as you just said, no two people are the same, and they have completely different DNA; and they have completely different life experiences. They eat different food for lunch, all of this stuff.
So people are very different, and then we don’t have really good ways to collect that data about them and store it and track it. And so it’s really dirty data over a bunch of different kinds of patients. So my question is: How far do you think we’re going to be able to go? How healthy will we be?
You can pick any time horizon you want. Will we cure aging? Will we eliminate disease? Will we get to where we can sequence any pathogen, and model it with the person’s DNA in a computer, and try 10,000 different cures at once, and know in five minutes how to cure them? Or do we even have a clue of what’s eventually going to be possible?
So I think one of the interesting things, when I first joined Hopkins, that I learned very early, is that when we dream of what an ideal health system ought to look like… Wouldn’t it be great if we had cures for everything? [But] one of the most surprising and disappointing facts I learned was that even in cases where we know what the right treatment is; even in cases that we know—where we could have treated them, had we known upfront, who they were and what was the appropriate sort of therapy for them…
Right now, we have many such cases we miss. So I don’t know if you’ve seen this Institute of Medicine report that came out in 2011 or 2010—I can’t remember the date—where they talk about how a third or a quarter of the amount of money that’s spent in healthcare, they think of it as ‘unnecessary waste’.
Unnecessary waste means waste because we are over-treating; waste in cases where we’ve kept people longer than was necessary; waste because there were complications that were preventable; waste because we gave them treatments that weren’t the right treatments to begin with, and we should’ve given them something else.
And I don’t think the answer is as simple as, “Oh, why isn’t our health system better? Is it because we’re not training the most competent doctors? Is it because our medical educational system is broken?” No. I think if you actually sit inside a hospital, and you watch what’s going on, it’s such a multi-disciplinary, multi-person environment…
That every decision touches many, many people, including the patient. And there’s all this information, and all these decisions have to be made very quickly. And so what to know about any given individual, at any given time, to determine the right thing to do is actually very complicated. And it’s pretty amazing to me that we’re as effective as we are, given the way the system is built up.
So effectively, if you really think about it… To me, a part of it is the system’s problem, in the sense that if, going back… Our delivery of healthcare has very much come out of the era where there were only so many medications. They kind of knew what to do, there were only so many measurements, the rules were easy to store in our head, and you could really focus on execution—which is making sure we’re able to look at the individual and sort of glean what is necessary, and apply the knowledge we’ve learned in school very quickly.
And then the top challenge is… Medical literature is expanding at a staggering rate. Like you noted, the number of treatments has expanded at a staggering rate, but much more so, our ability to measure individuals has expanded. And as a result, even sort of knowing our notion of what is a disease…
It’s not just the case that… The rules aren’t so simple anymore. It’s much more challenging. Rather than saying, “For every person with sepsis, give them fluids.” No.
Some are very responsive, and some are not responsive, and the obvious one is if they have any kind of heart failure, don’t give them fluids because it’s going to make the condition worse. What I’m effectively going to is…
I feel there’s a huge low-hanging fruit here, which is… I think we can make human health a lot better by even thinking just harder about even all the treatments we already have, as we start taking many more measurements, and as these measurements are becoming visible to us in ways that they’re accessible.
Improving the precision at which we prescribe these measurements will make a huge difference, and I think that’s very tangible, very easy to… I think something we’ll get to within the next five to ten years. There are lots of areas of medicine that will see a huge improvement, just from better use of lots of data, that we already know how to collect. And thinking about the use of that data and improving how we target therapy.
I’ll give you an example: An area study that I am familiar with is, as I mentioned earlier, these complex diseases—like scleroderma.
They used to think of scleroderma as one disease, and any expert who treats scleroderma patients knows that there’s tremendous diversity among individuals when they come in. Some have huge impact on the kidneys, others have a huge impact on the gastrointestinal tract, and yet others have huge impact on the heart or lungs.
And effectively, when the persons come in, you’re kind of wondering, “Well, I have an array of medications I can give them. Who is this person going to be? And what should I be treating them with?” And our ability to look at this person’s detailed data and understand who this person is likely to be… And then, from that, targeting therapy more effectively, could already influence and improve treatment there.
So I think that’s one area where you’ll see a huge amount of benefits. The second area that I think… is basically increasing our ability to measure more precisely. And you can already see whole genome sequencing, microbiomes, and there are specific disease areas where being able to collect this much more easily will make a big difference.
And then, effectively, they’re going to give rise to new treatments because there are pathways that we are unaware of, that we will discover in the process of having these measurements, and that will lead to new treatment. So, I think the next ten years are going to be very, very exciting in terms of how quickly the field is going to improve. And human health is going to improve from our ability to administer medication and administer medicine more precisely.
That is a wonderful thought. Why don’t we close on that? This has been a fascinating hour and I want to thank you so much for taking the time to join us.
You’re welcome and thank you so much for having me! This was really fun.
Byron explores issues around artificial intelligence and conscious computers in his upcoming book The Fourth Age, to be published in April by Atria, an imprint of Simon & Schuster. Pre-order a copy here

Four Questions For: Sara Spivey

In what ways is technology changing marketing today?
In every way. I can’t think of an area in marketing that technology hasn’t changed in some way. From research to social media to everything in between, technology is omnipresent and essential to today’s marketers. Technology has enabled marketers to better identify our target audiences, track their behavior across the spectrum of touch points, provide customized offerings to our clients, work faster to move our audiences through the sales funnel, build better products based on real-time feedback, engage with customers on their preferred medium and much more. However, for all its benefits, I think technology can also be a crutch that we rely on too much at times. Like any tool, it should be used for a specific purpose, but not over relied upon to do everything for us. The human element still provides the competitive advantage to take all the benefits and insights technology offers and decide how and when they should be used.
How do you see big data influencing marketing in 2017? Any trends that you expect to see?
Big data has already played a tremendous role in helping marketers provide personalized offerings for their clients and respond quickly to changes in the industry. Yet I think it has also led to a loss of creativity in marketing. In today’s world, marketers are measuring every single thing with tools and systems trying to find an underlying analytical theme. This is leading to an abundance of data that doesn’t necessarily have any meaning or practical application. With a profusion of automated services collecting information, I think marketers have reached a point of saturation. For instance, we don’t need 25,000 data points about a consumer; we need the 12 that matter in their purchase journey. I think marketers will soon realize the need to distill the most important insights from data being collected and to use it to develop thoughtful, actionable, creative strategies to engage and reach their target audiences more effectively.
What role do you believe artificial intelligence will play in marketing?
Let’s take chatbots as an example of artificial intelligence in marketing. While there will still be a place for chatbots, companies will be much more selective about using them. Many companies have implemented these AI programs, but in some areas, they hurt the customer experience. Though automation technologies will still be used in marketing, I don’t think chatbots will be applied unilaterally. Companies are realizing that human interaction is still crucial to engaging consumers and sustaining relationships with the ones they already have. As far as virtual reality goes, it definitely has made a major splash, but I’m not sure how it leads to sales conversion, particularly in retail. For example, when consumers use Oculus devices in-store, they might be interested in purchasing that specific VR device, but it rarely leads to lateral purchases for clothes, electronics, etc. There’s a gap between correlating the VR experience between general interest and purchase intent. It’s also an incredibly expensive marketing tactic to employ.
How can marketers better incorporate data analytics into their strategies? Are consumers still as willing to share personal facts about themselves in hopes of receiving more personalized experiences?
Marketers must go the extra mile to provide more value to their companies by using their data to reach the customer when it matters most. I believe the industry will only continue to build and sharpen its data focus. The more marketers can learn about their consumers through data, the easier it is to personalize the outreach for a customized experience. That’s how they go the ‘last mile’ with consumers – by reaching them when they are actually in-market with personalized offers.
I am amazed by consumers’ continued willingness to share data, even with growing concerns around privacy. Consumers are sharing personalized facts about themselves at a much higher rate, and while they know companies are using this data to market to them, they are hoping that it will lead to more personalized experiences. It is crucial for companies, brands and retailers to take the data customers are sharing and apply it in a meaningful way that leads to better, more personalized experiences. Whether these organizations offer coupons based on consumers’ past purchases or exclusive access to certain products due to brand loyalty, these little factors can make a significant difference. Our recent study found that more than 70 percent of US internet users said their purchase decisions were influenced by coupons and discounts. What’s more is that 54 percent of consumers claim to have made a purchase as a result of brand outreach regarding abandoned shopper cart items or recommendations based on past purchases. When we have so many tools and data at our disposal, we owe it to our customers to get it right when reaching out to them and providing content that matters to them.

As Chief Marketing Officer, Sara Spivey is responsible for overall leadership of Bazaarvoice’s global marketing programs, including demand generation, solutions marketing, brand strategy, and communications. Sara has more than 30 years of marketing, strategy, and leadership experience with industry-leading organizations.

Report: Hybrid application design: balancing cloud-based and edge-based mobile data

Our library of 1700 research reports is available only to our subscribers. We occasionally release ones for our larger audience to benefit from. This is one such report. If you would like access to our entire library, please subscribe here. Subscribers will have access to our 2017 editorial calendar, archived reports and video coverage from our 2016 and 2017 events.
Data - generic
Hybrid application design: balancing cloud-based and edge-based mobile data by Rich Morrow:
We’re now seeing an explosion in the number and types of devices, the number of mobile users, and the number of mobile applications, but the most impactful long-term changes in the mobile space will occur in mobile data as users increasingly interact with larger volumes and varieties of data on their devices. More powerful devices, better data-sync capabilities, and peer-to-peer device communications are dramatically impacting what users expect from their apps and which technologies developers will need to utilize to meet those expectations.
As this report will demonstrate, the rules are changing quickly, but the good news is that, because of more cross-platform tools like Xamarin and database-sync capabilities, the game is getting easier to play.
To read the full report, click here.

Who needs traditional storage anymore?

The traditional enterprise storage market is declining and there are several reasons why. Some of them are easier to identify than others, but one of the most interesting aspects is that there’s a radicalization in workloads, hence storage requirements.
Storage as we know it, SAN or NAS, will become less relevant in the future. We’ve already had a glimpse of it from Hyperconvergence, but this kind of infrastructure is trying to balance all the resources – at the expense of overall efficiency sometimes – and they are more compute-driven than data-driven. Data intensive workloads have different requirements and need different storage solutions.

The Rise of Flash

sign-flash-trash1All-flash systems are gaining in popularity, and are more efficient than hybrid and all-disk counterparts. Inline compression and deduplication, for example, are much more viable on a Flash based system than on others, making it easier to achieve better performance even from the smallest of configurations. This means doing more with less.
At the same time, All-flash allows for a better performance and lower latency and, even more important, the latter is much more consistent and predictable over time.
With the introduction of NVMe and NVMeoF, protocols which are specifically designed to access flash media (attached to PCI bus) faster, latency will be even lower (in the order of hundreds of microseconds or less).

The Rise of Objects (and cloud, and scale-out)

At the same time, what I’ve always described as “Flash & Trash” is actually happening. Enterprises are implementing large scale capacity-driven storage infrastructures to store all the secondary data. I’m quite fond of object storage, but there are several ways of tackling it and the common denominators are scale-out, software-defined and commodity hardware to get the best $/GB.
Sometimes, your capacity tier could be the cloud (especially for smaller organizations with small amounts of inactive data to store) but the concept is the same, as are the benefits. At the moment the best $/GB is still obtained by Hard Disks (or tapes) but with the rate of advancement in Flash manufacturing, before you know it we’ll be seeing the large SSDs replacing disks in these systems too.

The next step

High Self Efficacy Level - Efficiency ObjectiveTraditional workloads are served well by this type of two-tier storage infrastructure but it’s not always enough.
The concept of memory-class storage is surfacing more and more often in conversations with end users, and also other CPU-driven techniques are taking the stage. Once again, the problem is getting results faster, before others if you want to improve your competitiveness.
With new challenges coming from real-time analytics, IoT, deep learning and so on, even traditional organizations are looking at new forms of compute and storage. You can also see it from cloud providers. Many of them are tailoring specific services and hardware options (GPUs or FPGAs for example) to target new requirements.
The number of options is growing pretty quickly in this segment and the most interesting ones are software-based. Take DataCore and its Parallel I/O technology as an example. By parallelizing the data path and taking advantage of multicore CPUs and RAM, it’s possible to achieve incredible storage performance without touching any other component of the server.
This software uses available CPU cores and RAM as a cache to reorganize writes while avoiding any form of queuing to serve data faster. It radically changes the way you can design your storage infrastructure, with a complete decoupling of performance from capacity. And, because it is software, it can be installed also on cloud VMs.
A persistent storage layer is still necessary, but will be inexpensive if based on the scale-out systems I’ve mentioned above. Furthermore, even though software like DataCore’s Parallel I/O can work with all existing software, modern applications are now designed relying on the fact that they could run on some sort of ephemeral storage, and when it comes to analytics we usually work with copies of data anyway.

Servers are storage

Software-defined scale-out storage usually means commodity X86 servers, for HCI is the same and very low latency solutions are heading towards a similar approach. Proprietary hardware can’t compete, it’s too expensive and evolves too slowly compared to the rest of the infrastructure. Yes, niches good for proprietary systems will remain for a long time but this is not where the market is going.
Software is what makes the difference… everywhere now. Innovation and high performance at low cost, is what end users want. Solutions like DataCore do exactly that, making it possible to do more with less but also do much more, and quicker, with the same resources!

Closing the circle

Storage requirements are continuing to diversify and “one-size-fits-all” no longer works (I’ve been saying that for a long time now). Fortunately, commodity x86 servers, flash memory and software are helping to build tailored solutions for everyone at reasonable costs, making high performance infrastructures accessible to a vaster public.
Most modern solutions are built out of servers. Storage, as we traditionally know it, is becoming less of a discrete component and more blended with the rest of the distributed infrastructures with software acting as the glue and making things happen. Examples can be found everywhere – large object storage systems have started implementing “serverless” or analytics features for massive data sets, while CPU intensive and real-time applications can leverage CPU-data vicinity and internal parallelism through a storage layer which can be ephemeral at times… but screaming fast!

Serverless-enabled storage? It’s a big deal

The success of services like AWS Lambda, Azure Functions or Google Cloud Functions is indisputable. It’s not for all use cases, of course, but the technology is intriguing, easy to implement and developers (and sysadmins!) can leverage it to offload some tasks to the infrastructure and automate a lot of operations that, otherwise, would be necessary to do at the application level, with a lower level of efficiency.
The code ( a Function) is triggered by events and object storage is perfect for this.

Why object storage

Object storage is usually implemented with a shared-nothing scale-out cluster design. Each node of the cluster has its own capacity, CPU, RAM and network connections. At the same time, modern CPUs are very powerful and usually underutilized when the only scope of the storage node is to access objects. By allowing the storage system to use its spare CPU cycles to run Functions, we obtain a sort of very efficient hyperconverged infrastructure (micro-converged?).
Usually, we tend to bring data close to the CPU but in this case we do the exact opposite (we take advantage of CPU power which is already close to the data), obtaining even better results. CPU-data vicinity coupled with event triggered micro-services is a very powerful concept that can radically change data and storage management.
Scalability, is not an issue. CPU power increases alongside the number of nodes and the code is instantiated asynchronously and in parallel, triggered by events. This also means that response time, hence performance, is not always predictable and consistent but, for the kind of operations and services that come to mind, it’s good enough.
Object metadata is another important key element. In fact, the Function can easily access data and metadata of the object that triggered it. Adding and modifying information is child’s play… helping to build additional information about content for example.
These are only a few examples, but the list of characteristics that make scale-out storage suitable for this kind of advanced data service is quite long. In general, it’s important to note that, thanks to the architecture design of this type of system, this functionality can boost efficiency of the infrastructure at an unprecedented level while improving application agility. It’s no coincidence that most of the triggering events implemented by cloud providers are related to their object storage service.

Possible applications

Ok, Serverless-enabled storage is cool but what can I do with it?
Even though this kind of system is not specifically designed to provide low latency responses, there are a lot of applications, even real time applications, can make use of this feature. Here are some examples:
Image recognition: for each new image that lands in the storage system, a process can verify relevant information (identify a person, check a plate number, analyze the quality of the image, classify the image by its characteristics, make comparisons and so on). All this new data can be added as metadata or in the object itself.
Security: for each new, or modified, file in the system, a process can verify if it contains a virus, sensitive information, specific patterns (i.e. credit card numbers) and take proper action.
A businessman or an employee is drawing an analytics optimisation chart on the glass screen in a modern panoramic office in New York.Analytics: each action performed on an object can trigger a simple piece of code to populate a DB with relevant information.
Data normalization: every new piece of information added to the system can be easily verified and converted to other formats. This could be useful in complex IoT environments for example, where different types of data sources contribute to a single large database.
Big Data: AWS has already published a reference architecture for Map/Reduce jobs running on S3 and Lambda! (here The link)
And, as mentioned earlier, these are only the first examples that come to my mind. The only limit here is one’s imagination.

Back-end is the key

There are only a few serverless-enabled storage products at the moment, with others under development and coming in 2017. But I found two key factors that make this kind of solution viable in real production environments.
The first is multiple language support – in fact the product should be capable of running different types of code so as not to limit its possibilities. The second, is the internal process/Function scheduler. We are talking about a complex system which shares resources between storage and compute (in a hyperconverged fashion) and resource management is essential in order to grant the right level of performance and response time for storage and applications.
One of the most interesting Serverless-enabled products I’m aware of is OpenIO. The feature is called Grid For Apps while another component called Conscience technology is in charge of internal load balancing, data placement and overall resource management. The implementation is pretty slick and efficient. The product is open source, and there is a free download from their website. I strongly suggest taking a look at it to understand the potential of this technology. I installed it in a few minutes, and if I can do it… anyone can.

No standards… yet

Contrary to object storage, where the de facto standard is S3 API, Serverless is quite new and with no winner yet. Consequently, there are neither official nor de facto standards to look at.
I think it will take a while before one of these services will prevail over the others but, at that time, API compatibility won’t be hard to achieve. Most of these services have the same goal and similar functionalities…

Closing the circle

Data storage as we know it is a thing of the past. More and more end users are looking at object storage, even when the capacity requirement is under 100TB. Many begin with one application (usually as a replacement of traditional file services) but after grasping its full potential it gets adopted for more use cases ranging from backup to back-end for IoT applications through APIs.
Serverless-enabled storage is a step forward and introduces a new class of advanced data services which will help to simplify storage and data management. It has a huge potential, and I’m keeping my eye on it… I suggest you do the same.

Originally posted on