ORLANDO — When Hasso Plattner talks, it is best to listen. The SAP SE co-founder and current chairman of its supervisory board, can explain technical issues with the best in the world, words often leaping from his lips with a German lilt. At 73, his energy has hardly waned.
Plattner talked earlier this week — first as the Day 2 keynoter at SAP Sapphire at the Orange County Convention Center in Orlando, then in a more relaxed news conference with media and analysts — and I had the chance to listen. I even got a question or two in. What follows are some of Plattner’s comments about machine learning, which was a big focus at the three-day event, voice recognition, cybersecurity, robots and more.
Why do so many people seem to have difficulty explaining machine in straightforward and clear terms?
HP: You might have seen my title. I’ve been a professor for 15 years, but I have convince students about things I just understood — or probably not fully understood — and this is fast learning, human learning. You have to learn to explain in a way. Otherwise, they just look at you and discard you, and that’s not a good feeling for a professor. So, explaining … I judge people when I hire them by how well they can explain. There’s something we have to permanently train. Sometimes, you’re good at it, sometimes now.
Does analytics have an issue of trust that needs to be overcome in the executive suite? And how can executives trust the data?
We have never had a better relationship with top executives than when we showed them the digital boardroom. If somebody comes to SAP and we show them the digital boardroom — with a non-disclosure — we can go through any data we have in the SAP system. 99 out of 100 say, “I want this system. How long? How much?” There is work to do. It does not just come out of the box. You have to know what you want to do. You have to know how you want to drive your company. That is specific to the management style.
What is the future of voice recognition in the business world?
We’ve shown glimpses of natural language that we can understand. Others have built systems — many of you might be using Siri — so voice recognition has come a long way in the last 10 years. We have thrown all the resources we have now into machine learning for the foreseeable future to get as many projects going in order to have an impact. We started a little late with AI. … I was there 25 years ago, but it was not fast enough for our type of applications, and it was outside the system. Now we can apply AI inside the system. That is the rush.
I am not running the company anymore, I can only make recommendations by German law, and I made the recommendation that we restart voice recognition in connection with these new applications and we find a place in the world where we have capacity, whether that is in Palo Alto, or in China. … Not all components we use have to be acquired. We could work with Apple. That would be my first choice. There are many options for voice recognition.
You work so closely with machine learning, and obviously IIoT. What is your stance on cybersecurity? What sort of measures do you take? Do you have a special team?
We have to shield our data, and keeping data in one location is better than having redundant data everywhere. A lot of the business data has to be protected. … This is a real issue, and national professionals are leaking out spyware.
Obligatory robot apocalypse question: Will AI ever take over?
For many years, the autopilot in an aircraft flies better than the pilot. I once flew my aircraft for five minutes, the pilot was watching me, and it was like driving in honey. The autopilot is much better, it has much faster sensors, much better activators. There are things machines can do better, more and more — they build cars, basically — but our human brain will not be made superfluous. We will stay in control. I don’t see the robots marching in.
Ethan Miller, Getty Images/ FCC chair Ajit Pai.
R.I.P., NET NEUTRALITY: The Federal Communications Commission voted 2-1 on Thursday to begin dismantling new neutrality. Two perspectives on the issue, first from FCC chairman Ajit Pai, second fromGigi Sohn is a Distinguished Fellow at the Georgetown Law Institute for Technology Law & Policy.
Pai: This “notice is the start of a new chapter in the public discussion about how we can best maintain a free and open internet while making sure that ISPs have strong incentives to bring next-generation networks and services to all Americans.”
Sohn: This “action signifies more than a fight over Net Neutrality. What is now at stake is the ability of the FCC – the expert agency by law - to protect consumers on what is now one of the most critical inputs to the US economy – broadband networks. Instead, chairman (Ajit) Pai is proposing to abdicate the FCC’s role and give it to the Federal Trade Commission, while an important partner, cannot make rules and lacks the technical expertise to do so. Under the FTC, consumers will have no protection whatsoever until long after the harm to them is done.
Not even Basketball Hall of Famer Rick Barry can hit as many free throws as BallBot.
ROBOT OF THE WEEK: We are deep in the NBA playoffs (and likely headed for Cavs-Warriors III, the first trilogy in NBA Finals history and the first in any North American major pro sports league in 61 years), so why not turn to the hardwood for our latest RotW? BallBot is a robotic ball launcher designed by Geva Patz, a technologist who works at the National Museum of Mathematics, and it almost never misses a free throw. Can it teach Shaq or DeAndre Jordan some proper form?
MORE NEWS FROM SAP SAPPHIRE: What else happened during three days in Orlando? One of the bigger stories involved IBM joining forces with SAP Ariba “to develop and deliver next-generation applications that redefine the source-to-settle process.” That means IBM will use Watson and SAP will use Ariba and its new Leonardo platform to deliver cognitive procurement solutions for users. It blends procurement data and predictive insights to produce better decisions.
“We’ve built a cognitive procurement platform trained specifically to understand procurement transactions and unstructured data such as weather, non-standard part numbers in contracts and complex pricing structures,” said Jesus Mantas, general manager of Cognitive Process Transformation for IBM Global Business Services. “By combining the power of IBM Watson on the IBM Cloud with SAP Ariba, we are leaping existing procurement benchmarks and delivering unprecedented value to our joint clients.”
Plattner, meanwhile explains Leonardo … as only he can: “I learned a new English word today: bounding box. It’s a math term, a box around a set of objects, and in this case, it’s a box around a set of tools to build a system that then, with machine learning algorithms, finds insights and builds algorithms we can attach analytics and transactions, et cetera. It is mainly a set of tools we can use to produce a machine learning system that produces software.”
THE REACHING EFFECTS OF A CYBERATTACK: Hacking and cyberattacks have been back in the news for more than a week, of course (thanks, WannaCry ransomware). Online security company Centrify aimed to break down some of the related numbers you might not normally think about in a new report that surveyed more than 1,300 IT and security professionals, senior marketing professionals, and consumers.
Take company stock price, for, example, which drops about 5% on the day a company’s breach is exposed. The percentage of customers who leave in the wake of a cyberattack, too, is particularly high, with 31% of consumers affected by a breach ending their relationship with the company. Oh, and 56% of IT practitioners say they aren’t confident in their own abilities to prevent, detect and resolve an attack. There are more numbers, but those are probably the most jarring.
HOW SMART ARE SMART HOMES?: According to new research from Parks Associates, more and more consumers are grinding through problems with their smart home tech. During the first quarter, 34% of folks who owned a smart door lock experienced issues when trying to fix a problem. (That figure is up from 22% year over year.) 35% of owners of smart doorbells with a camera experienced problems with their devices.
Consumers and manufacturers use some different tech, of course, and devices will almost always be more secure on a factory floor than near a dining room floor, but this is still an issue to keep in mind. No tech is perfect, and no tech is a panacea. Technical problems have increased over the last year with networked cameras, thermostats and plenty of other smart home devices. Be smart about what you have in your plant.
COMINGS, GOINGS and MONEY MATTERS: More machine learning this week! The California company FogHorn Systems, which develops edge analytics and machine learning software for IIoT applications, announced $15 million in Series A funding, notably from Dell Technologies Capital. (Saudi Aramco Energy Ventures, the venture capital arm of the Saudi Arabian Oil Company, also poured into this round.) CEO David C. King called it “the largest initial round of funding for a startup focused on fog computing and edge analytics technology.” … Stratasys released its quarterly financial report Tuesday, with revenues totaling $163.2 million, a GAAP net loss of $13.9 million (about 26 cents per share), and $25.4 million worth of cash generated from operations.
He wore suspenders back then, too. (Hey, it was the ’80s.)
QUOTE OF THE WEEK: “We think that in four to five years, the Japanese will finally figure out how to build a decent computer. And if we’re going to keep this industry one in which America leads, we have four years to become world-class manufacturers. Our manufacturing technology has to equal or surpass that of the Japanese.”
That was Steve Jobs, way back in February 1985, in a Playboy Interview still worth reading more than three decades later. (And, yes, that link is SFW; it’ll take you to the story aggregation site Longform, not Playboy.)
