ArchiTECHt Daily: IBM's cool—and confusing—cloud story

IBM certainly understands the value of marketing, but I think it might have finally outdone itself. T
ArchiTECHt Daily: IBM's cool—and confusing—cloud story
By ARCHITECHT • Issue #42
IBM certainly understands the value of marketing, but I think it might have finally outdone itself. The company announced some interesting new products at its InterConnect conference on Monday, but between the sheer volume of press releases it issued (20 since Sunday); the breadth of the product lines involved; and IBM’s, shall we say, ambitious claims, it’s really hard to tell what matters and what’s for real.
Here are what, I think, are the highlights:
Most of these products are in some form of beta, and at least one of the storage options is not yet available.
Barb Darrow at Fortune (my podcast co-host and former colleague) addressed some of IBM’s business issues in a story on Monday—including its being late to the cloud party and its enduring dependence on legacy systems—but I would say the situation goes even further. At a high level, IBM has too many divisions tackling, seemingly, the same space, and might have done itself a disservice with all the Watson marketing. 
On the latter point: it has been 6 years since Watson competed on Jeopardy!, and thanks to all the permutations, partnerships, and grandiose claims followed by skeptical reporting, it is hard to tell what Watson actually means or does. Analytics is not AI, and mere voice recognition is not curing cancer. 
I would also point out the IBM was very early in talking about the promise of the cloud and even branding a cloud lineup, Blue Cloud, in 2007. It took until about a year ago that IBM actually delivered, with Bluemix, any sort of cloud technology that people take seriously.
I think IBM’s embrace and commercialization of Blockchain is both progressive and potentially very important. I think its embrace of Kubernetes is very smart from a business perspective. I think its recently announced efforts around commercializing quantum computing in the next few years could be revolutionary. But at this point with IBM, I’ll believe it when I see it.

Sponsor: Datos IO
Sponsor: Datos IO
Around the web: Artificial intelligence
Some folks from Google, OpenAI and elsewhere launched aninteractive journal for machine learning research. It’s a great idea, especially to help non-researchers understand what’s happening in the space. For more, check out editor Chris Olah’s blog on Distill.
Seems like a good application of deep learning, even if it is a bit of overkill. Also, I didn’t realize this is where former Google exec Vic Gundotra ended up, as CEO. 
From Quartz, so it’s a little easier to digest. This really could be a big deal in terms of optimizing models, but as the article points our there are some issues and limitations at the moment.  •  Share
Ogma has built its own deep learning framework, including a technology that it calls “exponential memory.” If it works, I’m sure Google or Microsoft will take a look. This type of tech is very hard to commercialize.  •  Share
I have never heard of and its website doesn’t go into specifics on the tech, but it’s apparently working with a steel mill to optimize operations. Physical world case studies are a good thing.
Around the web: All things data
A clever new tool called Weld compiles code from systems like Spark, TensorFlow and Pandas, and optimizes it to look more like a single job and take advantage of parallel processing and multithreading.
This is a thorough, yet plain-English rundown of the new product Cloudera announced last week. I’ll add that I saw a demo at the company’s analyst conference this week and it looks pretty useful. 
Did you enjoy this issue?
The most interesting news, analysis, blog posts and research in cloud computing, artificial intelligence and software engineering. Delivered daily to your inbox. Curated by Derrick Harris. Check out the Architecht site at
Carefully curated by ARCHITECHT with Revue. If you were forwarded this newsletter and you like it, you can subscribe here. If you don't want these updates anymore, please unsubscribe here.