Back in October I was invited by IBM to attend their World of Watson event in Las Vegas – I wrote a little about it at the time.
Now that I have had some time following the event, I’ve been able to percolate and put my thoughts to paper, as it were. In the interest of you, dear reader, I’m splitting these thoughts into three different posts; Technology, Business and Philosophy. Note that there are no bright lines here: I’m sure to touch on each of the three topics in all three posts.
(I predict #1 and #2 will far outshine #3 in terms of traffic – such is life, friends!)
This post is the first – Technology!
There was a lot of showing off of the kind of shiny technology that you’d expect at a conference this size.
There was a booth that leveraged Watson’s Cognitive Computing Power to help you pick a beer – you’d drink a beer, rate it out of five stars, repeat for a few more, and it would tell you what pint to choose for maximum, data-defined drinking delight.
There was the Cognitive T-Shirt Bar, where you would interact with a Watson-driven chat bot on an iPad, and it would define your personality into one of six (maybe eight? I forget) buckets – and then you got a T-Shirt with that bucket’s logo.
(I did not get Gryffindor. I got squiggly purple lightning bolts.)
Of course, there was also Olli, source of my only popular tweet of the conference:
People love GIFs!
What is Cognitive Anyway
One term that was used heavily by IBM and IBM-adjacent folks in the build-up to the conference, during the conference, and now continuing on, is cognitive computing, sometimes shortened to just cognitive.
It took me a whole day at World of Watson before it really became clear to me what they meant when they said cognitive; they meant machine learning, algorithms that, using supervised or unsupervised learning sets, could create predicted outcomes with varying levels of confidence and / or success.
Note here that I say machine learning rather than statistical learning, which is a third term for, roughly, the same work.
This may be more suited for the philosophy part of this write-up, but notice here that which term you choose to use, cognitive computing, machine learning or statistical learning does not change what it is that you’re describing, but it does serve to illustrate to the audience something about you.
Cognitive computing is what it’s called when enterprise level operators use this technology. Machine learning is what it’s called when startup developers use this technology. Statistical learning is what it’s called when it’s used in an academic setting. Language is interesting!
AI as Supplement, Not Replacement
One common thread at World of Watson was the consistent meme of Artificial Intelligence as augmentation of human work, rather than a replacement of human work. We’ll set aside for the moment the fact that there wasn’t any actual AI present – there were advanced algorithms and pretty fast NLP chat bots, but you were aware you were chatting with a non-human.
(Yes there are a lot of philosophical bibs and bobs that come out of this, we’ll get there in Part III, hold your horses)
This idea, augmentation not replacement, was clearly the party line.
IBM Chairwoman Ginny Rometty focused on it in her Keynote; Computers as helpers, as assistants, as efficiency aids. This came up many, many times, and was fascinating on a few levels, especially noteworthy in light of the exception to the party line, Sebastian Thrun, the founder of Udacity, who was given a Keynote slot on the second day.
Thrun took a radically different direction in predicting the future; he used the metaphor of the simple shovel versus a backhoe. When we invent a tool that can do a job better, faster, and more cheaply than using human labor, why not use that tool? Why not free up that labor for other pursuits? We don’t dig foundations by hand, using a backhoe only on the tricky spots. We get in there with the big machine and get it done fast.
We’re made superhuman by new technology, he says. Whether that technology is a backhoe or an advanced algorithm, it allows us to reach new heights and do things previously unavailable to us – but only if we make use of the tool.
This is a big question – what is the role of machine learning in the future of human labor, of human endeavor?
A Spectrum of Work
We can conceive of the work we do, as humans, in many different ways. Here’s one spectrum we can use to consider it:
We can think of this as a zero sum spectrum; as we move toward things that computer excel at natively, those things become more difficult for humans to do well.
Something that would fall on the far right hand of the spectrum would be producing the output of an advanced mathematical model run against a massive database. That work might take a half dozen grad students a week to finish; my laptop can do it in minutes.
Something that might fall on the far left would be navigating the conversation at Thanksgiving dinner – a nuanced task that requires language input and output, the reading of social cues from multiple sources and their possible impact or influence upon one another, the use of memory from dinners past, the understanding of a complex social and historical network of the folks at the table as well as those (maybe intentionally!) not present.
For a long time we’ve leveraged computers and what they’re good at down there at the right hand side, in their territory. One way to interpret what is happening at IBM (and elsewhere), is using this scale.
We’re trying to push computers out of their comfort zone (as it were) into tasks that they don’t natively excel at. From things like computation toward things like conversation.
And, frankly, they have a long way to go. But they’re getting better!
There were a lot of really impressive things at World of Watson that were impressive in a right-hand-of-the-spectrum way, in a sort of not-sexy, not-glamorous way.
One thing that I was especially excited by, being a student of remote work, was the work of Joe Russo and his team at IBM. I spent probably an hour talking with Joe, and the work they’re doing to build a collaboration platform for distributed work teams was really remarkable – the front end resembled your classic Slack interface, but it leveraged the Watson back end to detect and present the most important conversations, and the most important pieces of those conversations.
Imagine you take a week off to attend a chair-building seminar deep in the Adirondacks, away from cell service and wifi networks. The fresh mountain air! The calming whir of a band saw. The satisfaction of a chair built with your own two hands!
Upon your return, you ask your collaboration platform, “What were the most important conversations about my projects? Who do I need to reach out to?” – and getting an accurate and actionable reply.
For someone who has spent hours of his life catching up on back scroll, this sounded pretty fantastic. Leveraging an enormous enterprise platform to solve this kind of small-daily-problem makes me happy.
Another use of the cognitive technology that I found promising was in providing better tools for support and sales organizations – being able to very quickly create a heads-up about every customer you’re chatting with is an obvious win.
In the time it takes to open an email, imagine beside the email are automated interpretations of that customer’s recent activity, danger to churn, similarity to other customers, possible outstanding and automatically detected errors or behavior patterns associated with common problems, etc.
Especially when combined with a chat-bot system, that sort of smart customer data collection and recollection could be a serious game changer.
That type of use of Watson and other IBM programs was the big talk at the conference – as amazing as natural language processing and leveraging self-learning algorithms to improve GPS routing is, what resonated with folks were integrations, especially with Salesforce.
Have you heard of Salesforce? It’s a pretty big deal* – and using IBM’s systems to get more actionable insights out of existing customer data was ringing a lot of bells for a lot of folks.
It’s absolutely worth noting that the things that most impressed me, and seemed to resonate the most with others, are still on the right hand side of that spectrum. Gathering data, applying a formula or function, and presenting the output at the right time – something computers are very good at indeed.
Wrap It Up
The technology on display at this event was amazing; a self driving bus, social-media responsive apparel, a robot you could talk to.
The stand-out technology wasn’t the sexy stuff – it was the back-end work, the solving of old problems using new solutions. The greatest strides were in turning our new tools back toward our existing work, our existing points of friction, rather than using them to seek out new problems.
Machine learning, cognitive computing, is only going to grow as part of the way we work. Staying perceptive, setting aside the ego, and figuring out how to ride this new disruption to the way we work and live, will be a huge component of success for technology professionals in the next decade.
I’m Simon Ouderkirk, I write about small data, remote work, and leadership at s12k.com. If you liked this first part of my World of Watson recap, please do follow my blog via your favorite RSS reader – Parts II and III, on Business and Philosophy at World of Watson, will be published soon! I’m also on LinkedIn & Twitter
* Salesforce is possibly the biggest deal
IBM has paid for me to attend World of Watson and provide unbiased coverage of the event. They have not provided content for me to publish, but ask that I do publish regarding the event on blogs and social media in exchange for free admission and travel expenses. My thanks to the Watson Analytics team for inviting me.