On the brighter prospects of a world with more tasks handled by machines

2-Color-Design-Hi-Res-100px-widthSince the advent of the world wide web in the early 1990s it has been possible to craft viable business models from highly specific — and limited — market niches. Now, in 2015, with the promise of an expansion in the capabilities of computing machines to handle more tasks of, perhaps, a mundane nature, this opportunity horizon has widened even further. (If you would like more information about why I have specifically connected the enormous popularity of web pages exposed over Ethernet networks for the general public as an important milestone leading to an enormous expansion in the range of viable tech business notions, please contact me as I offer consulting services in this area).

I think it makes sense for readers to keep this factor in mind as they witness public debate about the notion of just whether or not the proliferation of robots, hardware computing machines powered by algorithms, and even what are colloquially referred to as “smart” applications (and apps) will, in sum, result in a net positive, or negative, result for the sheer number of people employed.

An OPED piece published on the CNN web site on March 18, 2015 communicates the seriousness of this debate and adds a raw edge to it: Silicon Valley to millennials: Drop dead (http://www NULL.cnn NULL.com/2015/03/18/opinions/wheeler-silicon-valley-jobs/index NULL.html). The piece is written by David R. Wheeler. I could not find any information about him, beyond his picture on the CNN web site. So I can provide no background on why CNN decided to post his article.

The raw and right-to-the-point flavor of Wheeler’s chosen title for his piece certainly captures one’s attention. When this factor is combined with CNN’s decision to go to press, and prominently, with this piece, I would hope my readers will agree the topic has a lot of interest behind it, as it should given what I take to be Wheeler’s core point: “The commonly held belief is that with hard work and a good education, a young person in America can get a good job”.

Given the statistics Wheeler provides in his piece, he is probably correct in his conclusion the employment horizon has darkened. But if I replace “can get a good job” in the above quote with “can achieve financial security and even wealth”, then the horizon opens up for another phenomenon we are all witnessing today: an explosion in the number of small businesses and, particular, technology startups.

As recently as Sunday, March 29, 2015, an article appeared on the Financial Times web site about an entrepreneur by the name of Bart Van der Roost. Mr. Van der Roost has started a business by the name of neoScores (http://www NULL.neoscores NULL.com). I hope readers can share my appreciation for Van der Roost to craft what may become a very promising business from an especially narrow niche market — musicians requiring scores on digital devices. Perhaps we can extrapolate from his notion an opportunity for literally millions of these niches just waiting for entrepreneurs to expose.

Sure code is required. But isn’t code one of the skills people can go to college to learn? I hope we can all take a more sunny view of a new world of computing with hardware devices (powered by algorithms) capable of executing a widened vista of tasks.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2015 All Rights Reserved


Just how accurate are advertising predictions produced by machine learning systems?

2-Color-Design-Hi-Res-100px-widthThanks to Mikio Braun, who on Thursday, January 2015 published an article on the InfoQ website. Braun’s article includes mention of a Google acknowledgement about the role played by machine learning (also known, at least in part, as data processing by algorithms) as a predictive tool in its ad placement technology for its click ad business. Readers interested in this topic should read Braun’s article, which is titled Google on the Technical Debt of Machine Learning (http://www NULL.infoq NULL.com/news/2015/01/google-technical-debt-ml).

I have written about the inaccuracy of click advertising in earlier posts to this blog. To quickly summarize my opinion on this topic, I found the systemic tendency towards poor ad placement to be especially difficult to overcome when the items to be promoted provide subjective, intangible benefit. So gaining a perspective on just how much of the ad placement technology behind Google Adwords and, in all likelihood, its direct competitors (principally Microsoft’s Bing advertising system), as Braun points out in this short article is very helpful.

What is also very helpful in Braun’s article is the manner in which the Google researchers (Braun’s article is really a news report on a presentation at a recent conference event held in Montreal, the Software Engineering for Machine Learning workshop, part of the annual Neural Information Processing Systems, NIPS, conference held in Montreal) shed light on the precariousness of proper performance for machine learning systems, in this (online advertising) context, given the effect they have on other related computer processes. These researchers make clear how the basic assumptions powering Neural Networks can actually adversely affect these siblings, and, thereby, produce erroneous results along with very little value to people depending on them. Readers should note this conclusion is my own, and not a conclusion expressed anywhere in Braun’s article.

From Braun’s article, and the technical précis of a research paper on the algorithmic process behind machine learning, which was also published by Google researchers, online advertisers should be careful to set realistic expectations about paid placements. Perhaps it will make sense to horizontally structure these campaigns, with a panoramic reach wherever possible, if they are to produce any meaningful results.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2015 All Rights Reserved


The job of classifying large amounts of text data becomes easier with JSON

The final cloud-like computing theme contributing to the unfortunate fog around the notion of “big data” is JSON (http://www NULL.json NULL.org/). In my opinion, enterprise consumers of big data solutions built with NOSQL databases aren’t going to be able to connect the dots from the presentation on the JSON open-source project homepage.

More intelligible information about JSON for the non programmer can be found on the web site of the Apache CouchDB project (http://couchdb NULL.apache NULL.org/). “CouchDB is a database that completely embraces the web. Store your data with JSON documents. Access your documents and query your indexes with your web browser, via HTTP” (quoted from the first sentence of editorial content published on the site). Quering indexes with your web browser, hmmm . . . might this have something with Chrome’s Omnibox (https://www NULL.google NULL.com/search?sourceid=chrome-psyapi2&ion=1&espv=2&ie=UTF-8&q=omnibox%20json)? In fact, as any reader following the link just provided will note, it does.

So now with this flexibility in mind, it might provide enterprise computing consumers with more of a rationale for calling for the implementation of databases conforming to JSON, which will lend themselves to analytics built with NOSQL tools. If the process of collecting data on some aspect of a business process can be reduced down to little more than punching some keywords into Chrome’s Omnibox (a version of which is now available for Firefox and Internet Explorer), then Lines of Business (LoBs) can count on their personnel getting to the data they need, when they need it, from any device (mobile, desktop, laptop) whenever they need it without the need for any proprietary solution.

Pretty cool. The cool factor increases when one reads more about the CouchDB project. JSON represents an alternative to XML, which requires substantially more verbosity (meaning lines of code) to express the same programming statement. Lots of lines of code contribute to a slower web, where pages can take forever to load. So the comparatively lighter weight promised by using JSON to express steps in a program makes a lot of sense. The intention of JSON and XML are the same, namely to provide a method of data exchange (http://www NULL.idealware NULL.org/articles/data_exchange_alpha_soup NULL.php).

JSON produces “JSON Documents”. Here’s an example of what IBM© is doing with JSON: Search JSON documents with Informix (https://www NULL.ibm NULL.com/developerworks/community/blogs/idsdoc/entry/search_json_documents_with_informix?lang=en).

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2014 All Rights Reserved


The NoSQL notion suffers from some of the same ambiguity plaguing the notion of big data

Readers interested in finding out what NOSQL is all about will benefit from simply developing some familiarity with the definition of this acronym. NOSQL stands for “not only SQL”. I found this definition to be very helpful as it helped me correct my first misunderstanding about this notion. I thought NOSQL referred to a set of software tools designed to work with text, document, databases lacking the columnar table structure their Structured Query Language (SQL) siblings thrive upon.

But my understanding was wrong, which, unfortunately for businesses championing a NOSQL approach, may be the case of a lot of the enterprise user segment of the enterprise computing market for NOSQL analytics and the tools required for their delivery. mongoDB (http://www NULL.mongodb NULL.com/nosql-explained) is an example of a database built to conform to NOSQL.

But as the cliche goes “the best of all intentions” can go astray, as is the case, in my opinion, for the mongoDB definition. The average consumer of enterprise computing solutions built to work with social media conversations culled from lots of web pages, likely a chief marketing officer for a popular consumer brand-name, isn’t likely to be able to understand how “Document databases pair each key with a complex data structure known as a document. Documents can contain many different key-value pairs, or key-array pairs, or even nested documents” (quoted from the mongoDB web page presentation).

Further, characterizing the choices facing the enterprise consumer as an either “RDBMS” or “non RDBMS” isn’t going to be helpful if the literal definition of the NOSQL acronym is applied. As MapR© points out on its web site, an optimum approach to implementing NOSQL analytics is to combine SQL and text query tools built with JSON components to digest the same data, which, admittedly be incorporated into a mongoDB database, but came, originally from an RDBMS.

What’s even more surprising about the page on the mongoDB website is the light it sheds on a programming effort by a much larger, and much more mature ISV, namely Microsoft: “Graph stores are used to store information about networks, such as social connections. Graph stores include Neo4J and HyperGraphDB”. Hmmm . . . Now “Office Graph”, which is the predecessor of “Delve”, makes a lot more sense.

Ira Michael Blonder

© IMB Enterprises, Inc. & Ira Michael Blonder, 2014 All Rights Reserved