From that single organic nodule of package life an offspring is produced, or not.  This week has been particular enriched with insight and wisdom, some unsolicited and others remaining a bit of a quandary.

This week Hubert Dreyfus passed away.  A professor and a human being, a philosopher who challenged us to consider the practical limits of computers.  Aside from his academic acclaim and intense experience as a human being his message was far deeper than the antidotal point on message.  Yes, he asked us to consider the depth of use and application of technology in our personal lives, business and society.  He appears to have know the extent of human temptation and addiction.  Today we are drawn to the light like a moth to a candle, embracing the new and forsaking the tried and true. Is it because we are afraid of being left behind or are we considering the real vs. illusionary value of it all?  Or is it that we are sitting on the edge of restrained obsolesce and the jump seems right even if we might be stepping out into the darkness with hardly a basis of comfort.  Although I never had the chance to meet him in person I enjoy some brief ‘technological’ interchanges to better understand his stance on artificial intelligence.  During my advanced studies he provided invaluable opinions about the difference between machine learning, the need for social contact and interchange, along with some quite private discussions about risk.   It was a bit unnerving to consider to realize which was always in front of my eyes that the success or destruction is not in the device but the enablement provided by the human steward.   To me it wasn’t just a learned opinion, although gifted with experience at MIT and Rand, but his deep and profound thought given on the subject.

The Junk Drawer

Few people do not have a junk drawer.  Whether it be made up of household repair items or kitchen gadgets we all manage to eek out a space to stash away an much anticipated device of salvation.  Likewise we see the emergence of the same for abandoned technology.  Cables, cell phones, chargers and various explored ancillary devices find their way to ‘the drawer’.  We pretty much know that what goes in is unlikely to come and be used, time is not on the side of technology and the obsolescence that occurs.  But our frugal nature suggests that we or someone might find a need or use for these cast aside items.   I mention the junk drawer in a broader context that we have lots of technologies that have come and gone.  Often replaced by what appeared to be superior solutions, that later prove to be less superior that even more future ones.

I think back to my very first expose to artificial intelligence in the 1970s.  It was with a very simple but quite illustrative product called VisiEXPERT produced by the now defunct Visi Corp.  The product was a very rudimentary rule based artificial intelligent (AI) solution.  Its operational example used the pairing of cheese with wine and allowed for the addition of new elements and relationships.  At that time it was robust enough to learn or be driven by inference, it required aggressive assertions in order to advance outcome delivery.  Later in the 1980s we saw the ADA and LISP development as service languages support for data driven behavioral modification processes.  In both cases their emergence was not months but decades in the making and although solidly formed it struggles to produce a groundswell of disciples.

So our junk drawer continues to grow and with this we see the rekindling of interest.  For those in the AI community the drive is not so much from the technology as it is the promotional support of the business community.  AI is talked about in a single breath with learning machines but how does this all fit together with life?

Embryonic Appearance

This past week (May 5, 2017) I read a piece that Scott Ambler a legendary agile disciple wrote about the “Darth of Qualified Agile Coaches”.  The points reflected a condition in which labels get applied but the lack of substantive value creates an abundance of non-value.  Even though the focus was on promoting professional qualifications there remains a quite similar condition as it pertains to AI.  We see a plethora of AI involved entities who for all intense purposes are new market entrants.  But let us not also be fooled by capabilities driven by attending a course, as Scott pointed out, it involves intense and purposeful experience to fulfill the obligations as an expert in a given field.  Even in my case with over 40 years in the information technology field and intensely active engagement in AI related activities I still feel I have lots to learn.  Its from this vantage point I wish to ask the question about capabilities.

When I think in quite simplistic terms as pertaining to AI I think of ‘the seed’.  That kernel catalyst that will drive the growth of technology based learning.  I also think about risks and what level of permissiveness that we should allow the AI model to undertake.  Embedded in that kernel is data and we need to be astutely aware that data is not always clean, controlled and ready for use.  It necessitates sterilization to make it ready and all must be done in as near to real-time as possible.  Momentary lapses in time or hesitation in commitment of cleanliness jeopardizes the AI value proposition.

To further emphasize this point I will refer you back to some earlier writings on did on advanced analytics.  I find that while analytics also makes use of data, it also has the potential to become a close partner with AI and the learning machine.  As stated in this earlier article (10/15/2014) the real next generation is not in preceptor or predictive analytics, it in the real of preemptive which takes action vs. alerting us of or indicating that a condition has the potential to occur.  In short, the model reflected the raw basis of the learning machine.  While not centered on growth of knowledge and more centered on action the elements exists for the feeding the AI model through preemptive analytics.  I also contend that anything short of being soundly grounded preemptive logic, including predictions is really shaky ground for AI.  The basis of this opinion is the potential for runaway illogical reaction by the AI model paradigm.  In the most simple of examples I think of how I look at a situation and react to later discover it was not exactly as I saw it.  If this had been applied to an AI scenario who knows whether reversion would have occurred and if it did was it through intervention or a separate set of rules to deal with error management?

Conclusion… Just a Wee Bit of Fertilizer

No planting would be complete without a bit of care and accelerated nurturing.  AI is no exception.  In this context our growth enhancement hormone is a combination of pragmatic engineering, anticipatory examination and a purposeful examination of our present state of intellectual discourse.  Most would agree that humans make errors and thus anything we do is both prone to error creation but also possibly error propagation caused by what we presently do.  One cannot view human emulated thought process as simplistic.  Even the most rudimentary movements, considerations, evaluations and commitments entail literally thousands of possible paths and choices.  While technology can handle volume and responsiveness it remains the dutiful obligation of humans to craft the paths, the gates of decisions, correlation of relationships and discernment of probable paths with rational and common place or dissenting opposition.  Its for these reasons that the engagement of personnel involved with AI cannot be causal technological Spartans.  Technical proficiency will remain important but the management of the operating intellectual paradigm will remain critically essential.  This involves raw in-sources from data, progressive analytics, paradigm development and deployment to error corrective models as the minimum for sound control.   Even then understanding is bounded by experience and therefore secondary ability to communicate, examine and model will help bolster the skill set needed for application of AI.

To the consumer there may be some worry or dissension, much as was the case with the use of voice response systems.  People are hesitant to embrace what is uncomfortable or viewed as inadequate by comparison to what they have grown accustom.  So while AI proponents dabble in the science there remains a great degree of need for transitioning of people to a new world order.  Some may enjoy a more hidden affront to consumers where others will be challenged regularly by real-time consumerism exchange.  Simply remember that all things are solvable as long as we understand the nature of the best, the human condition.