First things first
Before I dive in, a quick note that this will be the last full newsletter until April. I will probably publish a lighter issue next week, however.
IBM held its THINK conference this week, and used the opportunity to announce a litany of new products and partnerships, and to brag a bit about its Power9 systems. I can't possibly cover all the news, but here's a rundown:
- From months to minutes: New IBM Watson Data Kits mean faster time to AI value (IBM)
- IBM launches deep learning as a service inside its Watson Studio (TechCrunch)
- Apple and IBM unveil artificial intelligence service that Coca-Cola is testing (Fortune)
- IBM AI robots deployed to clean Earth’s oceans (FOX Business)
- IBM’s Watson Assistant lets any company build Alexa-like voice interfaces (The Verge)
- IBM sets tera-scale machine learning benchmark record with POWER9 and NVIDIA GPUs; available soon in PowerAI (IBM Research)
- IBM built a computer the size of a grain of salt. Here's what it's for (Fortune)
- IBM touts OpenPOWER ecosystem, announces new customers, products for AI and hyperscale (HPCwire)
Some of the cloud services and Watson Assistant seem like IBM playing catch-up (albeit a very necessary game of it) with AWS, Google and Microsoft. The Apple partnership is interesting, if only because (1) lots of people have iPhones and (2) IBM is really the only U.S. cloud provider that can partner with Apple on enterprise apps without a conflict of interest.
But what excites me here is the systems stuff, ranging from those tiny computers for tracking/securing goods to the POWER9+GPU system that IBM claims crushed a benchmark previously set by Google running TensorFlow on its cloud. Another major part of that latter story is the software library, called Snap Machine Learning, that IBM created. (They call it snap "because it trains models faster than you can snap your fingers" ...)
Looking past the usual issues with this type of record-setting performance benchmark (i.e., they're usually opportunistic and not necessarily apples-to-apples comparisons), I think it's important to not lose sight of the fact that IBM still has a core of really good researchers and systems architecture with POWER that seems to be picking up steam. And although I didn't catch any news about quantum computing coming out of THINK, IBM is one of the company's leading the charge there, as well.
However ... technology is changing fast and IBM has a lot of interesting pieces in place either in the lab or now making their way into production (I didn't even mention its early embrace of blockchain). Whether it can make the cultural changes and business changes necessary to capitalize on any of them is the trillion-dollar question, but more so than many of its "legacy IT" peers, IBM at least continually offers glimmers of hope that it can.