NetEvents Day 1: Artificial Intelligence, Business Intelligence and CyberSecurity Intelligence
Saratoga, Calif., Sept. 21, 2016 — Reporters from around the globe wore out the keyboards on their laptops taking notes at the NetEvents Global Press & Analyst Summit, held here in a beautiful mountaintop winery resort. With keynotes and panel debates focused on everything from artificial intelligence to cloud computing to the Internet of Things, the program was packed – and the information was flowing as freely as the morning coffee.
Manek Dubash, the master of ceremonies for the NetEvents Summit, introduced the first discussion, which was on “Artificial Intelligence: Out of the Futurists’ Lab, Into the Real World.”
The first part of the AI discussion was by Kathryn Hume, president of Fast Forward Labs, who framed AI as a series of stories. She explained that just as in Aesop’s Fables, the story depends on the teller. In this case, one can have a mental model for how AI algorithms work by thinking about fables, in other words, the problem that the AI is trying to solve – and then how it solves that problems.
Ms. Hume explained that in all markets, the earliest apps are rarely killer apps – and AI is no exception. However, while analysts and experts can’t predict killer apps, they can hone their ability to spot future game-changers and pivot to take advantage of them.
An example of this, Ms. Hume explained, is with natural language generation. At first, people thought that AI-based natural language would be used to, say, automatically generate news stories in a journalistic style, such as having software write a website news story based on sports data. While there certainly have been successes there, that turned out to be a niche market. Instead, natural language generation has proven to be good at helping summarize a lot of data for human consumption. Think about how a raft of numbers (such as raw sales data) can be turned into a spreadsheet or table, which adds some meaning. That data could be translated into a chart or graph or dashboard, which adds more meaning. But what if software could translate that into the textual “Sales increased in the last quarter”? That’s AI in action.
Another example she used is of AI to analyze X-rays; today, image recognition processing can do in 104 seconds what would take a human radiologist a full month. Imagine the impact on health care when such technology becomes more widespread.
The moral of the story, she said: We are at the beginning of an AI revolution. We aren’t sure where it’s going, but there’s a massive opportunity if we focus on where the technology is succeeding in this early stage market.
Ms. Hume was followed by Stuart McClure, an inventor and founder of AI-based cybersecurity company Cylance. Mr. McClure explained that what AI is good at is categorizing things – in the case of Cylance, about whether something (such as an executable file) is good (i.e., safe) or bad (i.e., malicious).
How do you do that? Training with lots of data sets. He used the example of determining whether an adult is a male or female. There are some indicators, some features, you can train software. Hair length? Odds are that if the hair is long, it’s a woman, if short, then a man – but of course, there are exceptions. Facial hair? If there’s a beard, it’s likely to be a man, no facial hair, less likely to be male, but of course, there are exceptions (many men are clean-shaven). And so on. A human might use a few features to make decisions – but AI technologies might use hundreds, thousands or even millions of features to learn patterns. With that type of power, AI can learn to make some impressive decisions – categorizing more data much faster than humans, and potentially with greater accuracy.
Following Mr. McClure’s presentation, he and Ms. Hume were interviewed on stage by Paul Jackson, Principal Analyst, Digital Media, at Ovum.
Continuing the theme of cybersecurity and AI, the first NetEvents debate panel was “Ransomware, Spear-Phishing and Worse — Defending Against the Unstoppable.” The panel was introduced and moderated by Andrew Braunberg, Managing Director, Research, at NSS Labs.
The panelists were Bryan Gale from Cylance; Greg Fitzgerald from Javelin Networks; Greg Maudsley from Menlo Security; Greg Enriquez from TrapX Security; and Frank Weiner from Wedge Networks.
The panel discussed that ransomware is a huge problem and is becoming bigger. Technologically, there’s nothing particular new about ransomware; like many other attacks, it requires that some type of executable code gets onto a client, whether from a bad email attachment, malicious website, or another source. What’s unusual about ransomware is that the attackers have found a way to easily monetize the attack: “We have encrypted your data, pay us money or you’ll never see your data again.”
Given the success of malware the attack is spreading, and now is even available as “ransomware as a service,” where anyone can launch an attack, even without highly technical skills. The answers, the panel agreed, consist of a multilayered defense. There is no one single approach, and an organization that wants to defend itself against ransomware attacks needs to look at everything on the market.
From the tech suite from the exec suite with the second debate panel, “From Rule-of-Thumb to Smart Data-Driven Businesses: The CEO/CIO Software Toolkit for Success.” Introduced and chaired by Dean Takahashi, a reporter with VentureBeat, the panel looked at the problem with information overload. From financial and inventory reports from the ERP system and a flood of roll-up spreadsheets, to documents and an unmanageable flow of messages, it’s too much. If information is the lifeblood of business, decision makers are drowning. It’s gone from helping make decisions and driving productivity, to harming efficiency and making it harder — not easier — to understand the business.
The panel consisted of Guillaume Amaud of Anaplan, Jim McNiel of NetScout, Rob Pickering of ServiceNow, and David Gurle of Symphony Communications.
The solution, the panel discussed that real-time decision support needs tools, like the cloud. It needs security for trust and compliance. It needs to go beyond dashboards to help employees know what they need to know when they need to know. Promising technologies include AI techniques such as machine learning, predictive forecasting and natural language. As the panel said, if you can frame your business problems, you can use those technologies efficiencies to help humans do their job better.
A key is size. The panel quoted the “law of big numbers” – you need a scale, a quantity of data to help use technology to make decisions. But if smaller companies have the data, they can go head-to-head against the biggest players.
Jean-Baptiste Su, technology columnist for Forbes, moderated “Unicorns, Baby Unicorns and Other Tech Leaders — Who Are They, Where Do They Go From Here?” His panel had Guillaume Amaud of Anaplan, Gregg Holzrichter of Big Switch Networks, Stuart McClure from Cylance, and David Gurle of Symphony.
In his introduction, Mr. Su explained that there are 175 unicorns – companies with over a billion dollars in valuation – most of which are in Silicon Valley. He noted, however, that the number of unicorns is shrinking – and this is a challenging time for companies to grow and succeed at that scale.
To get to size quickly, companies need hypergrowth, and the panelists insisted that if you want hypergrowth, you have to hyperspent. That means, of course, getting big fast, driving revenue. This might mean always flirting with profitability, but at this stage, the goal is to plough the money back in, rather than take money out.
The panel advised always that you never say “no” to more money, even if you aren’t actively looking for funding – because you never know when you might need the money. And when you do need more money, it may not be available at that moment. So take it when you can get it.
Some investors want their companies to double-down on growth and get as big as possible as fast as possible. Others advise executives and founders not to burn up their nest egg. In all cases, founders need to seize opportunities, look for partners, and never let something good pass you by.
Why is Silicon Valley special? Sure, there are opportunities elsewhere, where there’s little competition from the likes of Google and the huge array of companies in Northern California. That said, there is still something unique here in the Valley – institutions, capital, human capital. Sure, there is competition for talent, but the nexus of capital and entrepreneurial spirit gives Silicon Valley and its companies a competitive advantage.
“Let’s Redefine the Internet of Things: IoT Means Internet of Profits!” That’s the catchy name for the debate panel chaired by Tam Dell’Oro, Founder and President of the Dell’Oro Group. Her panel was Milind Pansare of Aerohive; Tom Ramar of H3 Dynamics; and Will Wise of the IoT Institute.
Tradition service-provider and carrier businesses are shrinking – which can mean declining revenue and declining spending. This panel talked about new opportunities for unlocking new markets and opportunities, such as package delivery, smart connected cars, and convenience businesses – all driven by the Internet of Things.
The challenge is organization dynamics: How quickly can service providers (and all companies, in fact) adapt to these new technologies? They need to understand how it will change their life and their customers’ lives. How it will create threats and opportunities. How it will redefine their business model. In some cases, there may be regulatory issues, like there are with cars and drones.
The IoT is all about devices talking to devices, often using Internet protocols. The IoT can allow startups to disrupt established giants with innovative business models, because the established company can’t threaten its established revenue streams – but a startup or new market entrant can do so.
IoT is also about connecting networks of networks. For example, a car has an internal network with many sensors and local intelligence tied together with its own LAN. In the IoT world, you can leverage that and powerful algorithms to create new capabilities – especially when you link it to other networks, such as the cloud. Then you can add in analytics, AI, and new services.
For service providers, there’s an opportunity to go beyond providing connectivity to leverage those IoT networks at scale to create new managed services.
you don’t know today what you’re going to need for network features and capacity a year from now – so design and architect that way. A challenge, of course, is to measure the ROI for these initiatives. The panel pointed out that there aren’t good academic models that can be applied to disruptive technologies like the IOT – so the best advice is to deploy and try and learn as you go.
Awards and More!
The remainder of the first day of the NetEvents summit included the Shark Tank presentations and awards for companies in the IoT and Cloud Innovation categories. More about that in a separate entry!