Sunday, December 11, 2016

Some thoughts on the future of human work

After a long period of cocooned studentship and apprenticeship in academia, I finally stand here on the brink of embarking on my professional career, and for the first time, I am forced to try and comprehend the nature of the working world around me in order to figure out where I fit in. I am writing this entry to help myself organize my thoughts.

We are in a world where AI and technology are increasingly taking over human work. AI professionals and their physical engineering support are the key enablers of this paradigm shift. Almost surely, these will be one of the most sought after professions in the near future. About other professions, at this point, I think we are in a transient "cold-start" phase where we are transferring human intuition and knowledge into training AI. So I think for the next few decades, one highly important task seems to be to work with AI experts to enable this knowledge transfer in different professions. Eventually, in some fields, we may be able to achieve full automation, and human work in these fields may become extinct. In other fields, like arts and humanities, AI may never entirely displace the human element but may play an important supportive role. I think that in most cases, we will continue to need professionals to keep asking new questions and setting up new problems for AI to tackle (until maybe eventually, AI learns the trends of human curiosity and desires and asks our questions as well; but I think we are quite far from that eventuality to worry about it now). In other words, in the next few decades, there is a tremendous opportunity for professionals to be a driving force behind this paradigm shift in their respective fields. I strongly think that professionals should embrace this paradigm shift and help enable it rather than resisting it. I also think that professionals can benefit from a basic knowledge of the current state of the art in AI to be in a good position to enable this change.

Unfortunately, I do not believe there is a lot of room in the above role - and it will mostly be the top of the pyramid professionals in their respective fields, basically entrepreneurs (including ones in professions of an entrepreneurial nature like law, medicine, journalism, art etc.) and academics, who will have the luxury to play it while guaranteeing a stable income. For other skilled workers, I would say that unless one strongly believes that their skill is highly unlikely to be replaced by AI in their professional lifetime - fortunately, a substantial chunk of skills may satisfy this criterion as of now - it seems prudent to bite the bullet and invest hard into learning some other skill, or better yet, acquire AI related skills (e.g., data science), since it seems that there will always be jobs at all levels of expertise (many studies predict a shortage of this talent in the near future). Maybe there will also be an emergence of a large number of jobs related to training AI - a market where people get paid to transfer domain-specific experience to AI. Such tasks are already being posted on online crowdsourcing platforms like Amazon Mechanical Turk.  I realize that the prospect of having our skills be rendered obsolete sounds disheartening. On the positive side, it seems (and I hope) that people will have to work less to achieve a good quality of life, and may have more time to pursue happiness in other meaningful ways.

There are several small paradigm shifts that we are seeing in the short term that are precursors to the big eventuality. The traditional structures of organizations are breaking down - companies are increasingly relying on contingent and contract labor following fine-grained modularization of tasks, which can be seen as the first step towards automation. Companies like Uber, Upwork, Taskrabbit, etc. are facilitating this shift. I think that an exciting domain for AI is to help navigate humans through their professional careers in this increasingly modularized working world - utilizing data to learn different attributes of individuals in order to match them with appropriate jobs, to discover the right skills for them to learn, and to learn the right ways to incentivize them.






Thursday, November 19, 2015

Winter

Memories of spring no longer turn up unannounced like they used to.
They now have to be conjured by silly rituals.

First, carefully mould the still midnight air
With the rhythmic roll of a chosen name, till you recognize it.
Recognize your existence that results from the realization that you really do.
All the while take circuitous routes through familiar places,
Awkwardly sniffing dead leaves that you missed in their prime
(the competition was stiff back then).




Tuesday, September 15, 2015

Create

Create something. Anything.
Because the right to consume must be earned.
Because it is the greatest joy.

Thursday, May 8, 2014

Anchor

Who am I?
I am the one who _ you.
Who are you?
You are the one who _ me.

Thus fully defined, we exist,
And as if not quite satisfied, insist
On tirelessly inventing a medley of identities,
Day after day,
So that unfailingly each night
We return to the HQ
All mixed up and confused,

Until, finally undressed,
And the curious deeds of the day
Carefully shelved away,
You _ me
And I _ you.



Thursday, April 24, 2014

Dead Plant

Her plant died the other day.
She said it was not from a lack of desire
For it to live; the situation was quite the contrary,
Yet you know how silly it is to think
That one can simply
Will away inevitability.

It made me sad,
But also angry about
Yet another unfair victory for reason,
Because if I had known,
And wanted for it to live too,
Reason stood no chance against the power of two.


And now, she will never believe me,
Citing an extrapolation argument.
Funny how that works.

Sunday, April 20, 2014

Dolores

Glistening in the warm sun
That she had conjured for him alone,
A sudden, solemn drop wells up in her eye
And rushes unhindered towards a parched womb,
Rendering every game frivolous and all words empty.

Tuesday, April 15, 2014

Missed connection

The episode ended as abruptly as it had started
Birthing a single lone verse
That quickly acquired a thousand faces
Each less familiar than the other.