IBM Think 2018: The Data + AI Inflection Point

By Joe Clabby, Clabby Analytics

IBM has consolidated its many customer/partner conferences (Interconnect, World of Watson…) into a single conference, now called “Think” – a once yearly, one-stop-shop for details on IBM products and strategies, customer/user strategies and implementation, product demonstrations, strategic planning (with access to industry/product experts) and hands-on laboratories.  Now customers and partners need only take one trip each year to get access to IBM executives, product experts, deployment advisors, support personnel, planning personnel and more.

Further, a rich ecosystem of third party hardware and software suppliers, VARs, systems integrators and partners also attend Think, ready to share product information, strategic insights and implementation suggestions.  Finally, the Think agenda is rich in educational opportunities, including dozens upon dozens of customer and vendor presentations designed to share real world experiences with interested attendees.

I approached Think with a goal of understanding “who” IBM really is.  The opening keynote; a closer look at some of the technologies that I follow as well as some that I don’t regularly follow; my attendance at IBM’s 5 in 5 session; and off-the-record candid discussions with IBM customers and business partners all contributed to my new view. My view? That IBM is an ethical, forward thinking technology leader with a desire to have a strong positive moral impact on business and society.

Data + AI: The New Inflection Point

If you look back at the history of computing you’ll find that there have been technology “waves” and technology “inflection points”.  Waves are trends like eBusiness, e Commerce, the move toward distributed computing, the adoption of the Internet, the adoption of personal computing and so on).  Inflection points are radical changes in computing that have a profound effect on buyer behavior or on computer use.  Examples of inflection points include:

  • Moore’s Law – the observation by Dr. Gordon Moore of Intel that semiconductor density and, therefore, overall processing power for computers doubles every two years. This led to a huge increase in productivity worldwide as computers became more predictable from a performance perspective – as well as more affordable. This would happen about every eighteen months as the microprocessors doubled in processing power from the previous generation.
  • Metcalf’s Law – the observation that the effect of a telecommunications network is proportional to the square of the number of connected users (network squared or n2). This observation was a statement that expressed the broad value of technologies that we create – in this case the value of networks.  This “law” showed that tremendous value could be derived by networking nodes together, and helped justify the creation of social networks, clouds – and, yes, the creation of the Internet.

At Think, IBM CEO Ginny Rometty introduced a new inflection point – a happenstance that occurs about once every thirty years in the computing industry wherein a fundamental convergence of technologies creates a situation that will have a profound effect on technology use and technological innovation for decades to come.  She called this happenstance “the Data + AI inflection point”.  It is the convergence of the massive amount of data being collected across the world with artificial intelligence (AI). This convergence is leading to an era of better decision making as machines learn, producing better results using advanced analytics combined with various forms of AI.  Machines, such as IBM’s Watson question answering computing environment for example, are heralding the arrival of an era where man works on almost a peer basis with intelligent machines, a “Man + Machine” era.

At Clabby Analytics we strongly agree with IBM’s perspective and believe that the merger of data with AI will govern the computing industry for decades to come.  In 2016 we wrote a report on this topic entitled “The Invasion of the Learning Machines” that described how these technologies work together and described various artificial intelligence approaches being used to effect machine learning).

In her keynote address, Rometty went on to say that now is the time to “go on the offense”.  She suggested that information technology executives learn to leverage their digital assets, embed learning into their systems and software environments – and empower workers with learning technologies.

Again, we are in full agreement with Ms. Rometty’s perspective. In fact, two years ago, in this blog, we pointed out how IBM was embedding learning within its software to improve user productivity and decision making.  Overall, it is critical that, in order to improve productivity and to make better decisions, organizations should actively leverage data; use intelligent systems to help make better decisions (as well as to find new revenue generating opportunities); and get new intelligent system learning tools into the hands of the workforce.

We expect that this new era of Man + Machine will not only impact workforce productivity, it will also have very significant impacts on society.  Some jobs previously in the domain of humans will now be executed by intelligent machines.  On the other hand, machines will need guidance – and that paves the way for a new generation of workers capable of guiding their machines to the desired results (IBM referred to this new generation of workers as “new collar” workers – a play on the older “white collar” and “blue collar” worker descriptions).

Rometty also claimed that intelligent machines will make “inclusion” a reality for everyone in society.  Intelligent systems are blind to race, color, sex, nationality and other qualifiers that have caused social biases in the past, suggesting that intelligent machines don’t care who their human partners are.  And removing such biases should lead to more open societies in the future.

Where I Spent Most of My Time at Think 2018

I am a research analyst with specialties in microprocessor and systems designs, infrastructure, management, security and certain applications (such as blockchain).  Accordingly, I gravitate toward sessions of systems hardware, infrastructure, security and systems/storage/network management.

From a systems perspective, I looked closely for new news in OpenPower, Power Systems, in z Systems (mainframes) and GPU-based systems:

  • OpenPower – OpenPower is an open consortium dedicated to building systems/solutions based around IBM’s POWER microprocessor architecture. At Think 2018, IBM held a separate, off campus update session on progress within the OpenPower Foundation.  New news included the announcement that the OpenPower Foundation has over 300 members at present, a far cry from the 59 members we described in this 2014 report.  A consortium member reported that some members, who originally thought that they would build their own POWER processors using the completely open IBM POWER design, had rethought their original focus and are now more focused on building servers and specialty servers using IBM’s POWER processors and chipsets.  The consortium member pointed out how Moore’s Law had come to an end, and now the focus is on accelerating system performance using faster busses, faster memory communications, new technologies (NAND, SSDs, etc.) and faster networks to drive-up overall system performance.

    One member of the consortium, Google, announced a new POWER-based platform named “Zaius”  Based-on POWER9 processors, the new platform uses on-chip accelerators, PCI-Express gen4, cache-coherent co-processors and an open firmware stack.  With a system this powerful, Google could enter the computer systems business (if it so desires) or use the technology internally (in volume) to create competitive advantages for its own services.  All-in-all, this is a pretty powerful design from a major vendor that could have a big impact on the systems/server marketplace.


  • IBM Power Systems – Power Systems are the POWER microprocessor-based servers designed and manufactured by IBM that offer 10X the I/O bandwidth and 3-4X the processing power of comparable Intel-based servers. In recent years IBM retooled its offerings, working hard on bringing Linux to the platform, capturing Linux applications using Little Endian bit order.  The company built a whole new low-end of Linux-based servers, then expanded upward with racks.  The focus in years past was on scaling for compute power.  But the new Power Systems are now also being focused on “Insight Scaling” – providing the processing power and memory needed to support advanced analytics environments.  Further, the new Power Systems support very large memory environments such as SAP Hana (POWER9-based servers can support up to 24 TB of main memory – enough to easily address the needs of @90% of the current SAP Hana installed base).  In fact, IBM claims that over 1,000 SAP Hana customers are using Power Systems.

    As a side note, I must mention that the large stage area in which the Power Systems update session that I attended was filled to the brim and beyond.  That reflects a tremendous amount of interest in Power Systems. That’s not surprising given the amount of processing power that Power Systems brings to bear on complex analytics, scientific computing and other applications that need the kind of processing power that POWER9 systems can deliver.


  • IBM z Systems – New on the z Systems front included an announcement that can extend z-level security out to distributed systems (HyperProtect). Imagine bringing advanced mainframe-class encryption and other security strengths to Intel-based servers – it would be like putting a security umbrella using the most advanced commercially available security server in the industry over entire systems environments.  This opportunity for IBM with HyperProtect could be huge.

    At Think, IBM focused heavily on the success of its mainframe/Linux servers known as LinuxONE (see our report here that describes the differences between LinuxONE and x86 servers).  Initially successful as a Linux consolidation server due to superior virtual machine handling, LinuxOne has gained strong acceptance among IBM’s mainframe customer base – as well as opening a slew of new deployments in numerous “new to mainframe” customers.  Over 50% of IBM’s customers now use Linux on their mainframes – and many have also deployed LinuxOne servers to run open source applications.

    IBM reiterated that its Blockchain Service is run on backend LinuxOne mainframes.  This is important because z Systems, with their advanced security as well as their orientation for transaction processing, z Systems are the best systems in the industry for Blockchain processing, as we state in this report.


  • Other – I did attend a session on NVIDIA’s Volta graphic processing unit (GPU) architecture. GPU-based systems are now being used for rapid processing as search assistants, for translation environments, for shopping, for detect/diagnose/treat applications, as well as in various manufacturing, EDA and agricultural environments.  I was fascinated to learn about a new GPU-driven farm implement that can march through fields, identify weeds and zap them, identify crops and fertilize them – all in the same pass.  This was a good example of how AI works for mankind.

Very Noteworthy: 5 in 5 – Women in Technology

I have three daughters, a wife and seven sisters, so women play a big part in my life – and women’s rights are important to me.  I like seeing women in the computer field – there are far too few, in my opinion.  Happily, there was no shortage of women during IBM’s presentation its annual “5 in 5” prognostication – a view of five innovations that will take place over the next five years that will have a significant impact on society over the next 5 years.

At Think 2018, IBM’s 5 in 5 indicated that:

  1. Crypto-anchors and Blockchain will help block counterfeiters in years to come – creating a formidable wall of defense against bogus goods;
  2. Hackers are going to have great difficulty when they encounter IBM’s “lattice” cryptography technologies;
  3. Small autonomous AI microscopes that will be networked in clouds will be able to monitor the condition of a very precious natural resource: water;
  4. Biases that influence AI results will slowly be eradicated; and,
  5. Quantum computing will become a reality.

What impressed me about this years’ 5 in 5 was that the lead researchers in three of these categories were women.  I don’t know if IBM did this on purpose, but I was delighted!   And let’s not forget, this company is being ably led by one of the tech industry’s few women CEOs.  Impressive.

Summary Observations

To understand “who” IBM is, what their strategies are, what their product offering are, who their customers and partners are – and more – the best place to start is with the keynote speech by Ginni Rometty, the company’s Chief Executive officer.  In her speech entitled “Think 2018 Chairman’s Address; putting Smart to Work” Rometty talked about a new inflection point in the industry, the kind of confluence that only comes along one every twenty-five years or so – and that inflection point is the convergence of data and AI.  She likened this inflection point to Moore’s and Meltcalfe’s Laws, and I concur – the convergence of AI and data have created a profound change in the computing industry.  It is now, indeed, the era of Man + Machine.

According to Rometty, this inflection point is changing the way business is conducted – and has major implications for IT customers and society at large.  In business, companies that figure out ways to exploit their internal data, as well as data from external sources, are creating distinct competitive advantage by being able to “outlearn” their competitors.  The problem is that only 20% of world’s databases are searchable (a problem that also presents a huge opportunity).

From a societal perspective, Rometty described an era of man PLUS machine, noting that the sharing of platforms and intelligence (yours and others) is creating new approaches to solving problems.  But the magic word in the previous sentence is “sharing.”  Sharing is important – the more data analyzed, the more accurate the results.

To maximize the positive social benefits of this new inflection point, Rometty emphasized the needs for trust and responsibility, for the creation of new skill sets that will lead to “new collar” jobs for the masses and making inclusion a reality for everyone.  This final point really hit home – as societies enabled by this data/AI inflection point, we need to include people of all races, colors and genders in our efforts to use AI for the common good of all people and the planet.

IBM’s contribution to this Data/AI inflection point is that of a provider of innovative technology (cloud services Watson and AI, Trust and Security), as well as a provider of strategic and implementation services.  All these points were clear at Think 2018, and these points described, in a nutshell, “who” IBM really is.

Leave a Reply

Your email address will not be published. Required fields are marked *