Joe Wikert's DisruptorFest

How Tesla is quietly building a disruptive AI advantage

Here's an interesting tidbit from Martin Ford's terrific book, Rule of the Robots: How Artificial Intelligence will Transform Everything:

Every Tesla is equipped with eight cameras that operate continuously, capturing images from the road and the environment around the car. Computers onboard the cars are able to evaluate these images, determine which ones are likely of interest to the company and then automatically upload these in a compressed format to Tesla’s network. Over 400,000 of these camera-equipped cars are driving on roads throughout the world, and that number is increasing rapidly. In other words, Tesla has access to a truly massive trove of real-world photographic data.

The author goes on to say, "Tesla’s ability to incorporate massive amounts of real-world data is a potentially disruptive advantage." Tesla is certainly an innovative vehicle and it's fascinating to think about the many ways it could lead to other marketplace disruptions.

Are we on the verge of a learning system breakthrough?

I've been experimenting with a popular video-based training platform as I continue my journey as a lifelong learner. The platform is excellent and offers a variety of resources; in fact, there are so many options to choose from it's sometimes hard deciding where to go next.

And that's the problem. The platform doesn't really know me. It's designed as a one-to-many approach where the content is extremely broad and it's up to the user to figure out where to invest their time.

There's nothing wrong with that approach but it doesn't take advantage of today's technology capabilities.

If we could combine three distinct technologies I think we could take a huge leap forward in learning systems. I'm talking about (1) voice UI's (think Alexa), (2) modern text-to-speech services and, (3) artificial intelligence (AI) where the platform learns about me and therefore delivers a custom, one-to-one, solution.

Voice UI's are all around us but they're still in the new, experimental stages. Early TV is sometimes referred to as "radio in front of a camera" and before too long we'll describe today's voice UI's in a similar fashion.

One way voice UI's will move forward is by having access to enormous libraries of richly tagged content. I say "richly tagged" because the content will need to be granularized so that it can be searched and reconstituted in an infinite number of ways depending on each user's needs. Also, we shouldn't rely exclusively on Amazon and their capabilities, hence the need for one of the more modern text-to-speech solutions which are often indistinguishable from an actual person.

The third leg of the stool is the AI to power the conversation, learn about me personally, understand how to answer and where to take me next.

It all adds up to a user experience that feels like I'm receiving one-on-one training from an expert on the topic. Over time the system learns about me and my strengths and weaknesses, just like any good teacher. It also builds successful learning paths based on different user skill and learning preferences thereby making the system even more useful for future users.

I'm curious if something like this already exists, even if it's on a small scale or in the early stages. The pieces of the puzzle are already available so it's just a question of pulling them together, managing the IP rights/income streams and offering it at a compelling price.

Experimentation and paid search

The next time you do a web search take an extra moment to see how the paid results compare to the top organic results. Sometimes the top link is the same for both.

This article, from the authors of The Power of Experiments, drives it home with this excerpt:

“Evidently, users who Googled ‘eBay’ (or another eBay-related search term), who had been clicking on the ad because they saw no reason to scroll down to the organic link just below it, were now instead clicking on the first organic search result. For these searchers, eBay essentially swapped in free organic clicks for each advertising click lost,” explain Luca and Bazerman. “In other words, much of the money eBay was shelling out to Google each year was a waste.” After the results of these experiments were published, 11 percent of large companies that were buying search ads in the same way as eBay discontinued that advertising.

It reminds me of that classic quote: "Half my marketing spend is wasted...I just don't know which half."

More importantly, it illustrates the need to continuously monitor and analyze data, all the while maintaining a strong culture of curious experimentation.

I wonder how much of Google's income is derived from advertisers who never bother asking if their high organic ranking might perform just was well as the results they're paying for...

Amazon, Alexa and podcasting

When the Spotify-Joe Rogan news hit last week I wondered again about Amazon's role in podcasting. Sure, you can ask your Alexa device to play the latest edition of just about any podcast but is Amazon only looking to act as a pass-through agent, serving up streams from Apple?

If the latest rumors are any indication, the answer is no. It appears Amazon doesn't want to simply enter the podcast market...they want to totally disrupt it.

The local content angle may seem somewhat narrow at first but think about the possibilities. Amazon certainly has the resources to curate the best of the best as well as fund development of new local content while newspapers, local TV, et al, are declining. And if I'm going to Amazon for my local podcast content I'm also shifting all my non-local podcast subscriptions to their one app/service as well. All of this, btw, will be accessible through the countless Alexa devices in all our homes (and ears, as my Bluetooth earbuds are also Alexa-enabled).

The most interesting element of this Amazon podcasting story is the advertising angle. Most of the ads I hear in podcasts today are still very mainstream, trying to cast as wide a net as possible. That's why 99% of those ads don't resonate with me. Amazon, however, is loaded with data about my preferences, buying habits and more. They're uniquely positioned to serve up programmatic advertising for an audience of one: you. That doesn't exist in the podcast world today but it definitely will tomorrow.

Advertising engagement and conversions would both be exceptionally high in this environment. And if you like what you hear, buying/subscribing/opting-in to whatever the ad is promoting will be as easy as saying, "Alexa, sign me up for..."

This model makes privacy advocates cringe, of course, but it's also likely to create an entirely new ecosystem of streaming content driving significantly more revenue for plenty of parties, not just Amazon.

Disrupting and improving communication with machine learning

The topic of artificial intelligence (AI) is generating a lot of buzz these days and it's often difficult separating fact from fiction. For example, what are the most interesting AI applications today and where is the technology heading tomorrow?

I recently started reading a good book on the topic called Prediction Machines, by Ajay Agrawal, Joshua Gans, and Avi Goldfarb. Prediction Machines offers a solid overview of AI fundamentals while also providing plenty of real-world examples. One of my favorite examples is Grammarly, a tool to help improve written communication. Here's how the authors describe the service:

Grammarly achieved these corrections both by examining a corpus of documents that skilled editors had corrected and by learning from the feedback of users who accepted or rejected the suggestions. In both cases, Grammarly predicted what a human editor would do. It goes beyond the mechanical application of grammar rules to also assess whether deviations from perfect grammar are preferred by human readers.

Years ago there were a few grammar-checker software products that tried to solve the problem the old-fashioned way, with brute force. They certainly helped fix a lot of grammatical errors but they often didn't produce the results you'd get from a good human editor.

I'm using the free Grammarly service, both as a standalone app and as a Chrome plug-in, so this article was made better thanks to Grammarly. I'm also going to let Grammarly have a look at some of the documents I write at work.

There's a danger in all of this. Google has dumbed us down, making us over-reliant on their search and map services, for example. I spend less time thinking about the best route and instead simply plug the address into Waze and let it tell me. The same thing could happen with Grammarly where my writing skills decline as I get lazy and rely on the service to fix my errors. My plan is to stop and think about each correction Grammarly recommends and do my best to avoid making the same mistake again but we'll see...

I hope you'll try out the Grammarly service as well. If you're interested in where AI is heading, be sure to read Prediction Machines and think about how this rapidly changing technology is likely to impact your business and your job.