Started reading The Inevitable: 12 Technological Forces that Will Shape Our Future (affiliate link).  I am only on page 83, and I have already bought 9 copies for friends. It’s written by Kevin Kelly, co-founder of Wired magazine, and does an amazing job of explaining what artificial intelligence (AI) is, what it is becoming, and how it will completely change everything. Fascinating, frightening, and – well – inevitable. This is the first blog post of many on this topic.

There are 12 chapters on each of the technological forces. Interestingly, he describes characteristics of new generation of technology: becoming, cognifying, flowing, screening, accessing, sharing, filtering, remixing, interacting, tracking, questioning, beginning. Let’s start with #1 BECOMING.

Showing the author’s quotes in blue italic.

Endless Newbie. This is how we should view ourselves in relation to technology. It is changing so incredibly fast that the most important technologies that will dominate life 30 years from now have not yet been invented. Lots of examples in the book – but it’s safe to say that it is more than just the speed of change – it’s the second and third derivative. Let’s call it the acceleration and jerk.

Upgrading is a new hygiene. Information flows are continuous. The concept of version 1.0, 2.0 is very antiquated. Too slow, too focused on product shipping cycles, and shrink-wrapped CD-ROMs of the shelves of Circuit City. Software updates are going to flow . . like the Tesla Model S updating its windshield wiper algorithm when it’s early users complained of a glitch. Even if you don’t want to update, you will, you must. Hence, inevitable.

Protopia. It is not utopia (perfect heaven, nirvana) and not dystopia (broken, ruined world). It’s a protopia (state of becoming, progressing, changing). The subtle progress is not dramatic, not exciting. It is easy to miss because a protopia generates almost as many new problems as new benefits.” It will be an imperceptible grind of progress with intelligence flowing into everyday objects “like electricity.” Curious what the author would say, but he seems to be arguing that it will be less of B2C technology adoption lifecycle (refer: Geoffrey Moore’s Crossing the Chasm) because it will be so ubiquitous that you cannot even avoid it.

The internet was not like TV. People first thought that the internet would be like TV on the computer. Wrong. Who could have predicted that users would generate so much of their own, free, widely-available content?  YouTube, Instagram, Facebook Twitter, TripAdviser, Yelp, etc. . .Kelly notes that 300 hours of video is posted on YouTube every minute. . . uh, (18,000 to 1).

Web 2.0 will not be like Web 1.0 In the same way, we might foolishly think that the web in 2050 will just be a bigger, more intuitive internet. Wrong. The future web will 1) have connection to all kinds of physical items (internet of things) 2) be able to view earlier versions of the same website (e.g., J Crew catalog during last year’s black Friday sale) 3) will anticipate your actions, needs, requests.  It will be a low-level constant presence like electricity; always around us, always on; subterranean.”

Bad at predicting. In 1994, Time magazine said the internet would never go mainstream. I actually remember also telling a friend (KK) that the internet was a bit silly, and a waste of time. Uh – epic bad call.

The key takeaway is that we are at the beginning of the beginning.” We think we know what to expect, but we don’t.  Famously, www.mcdonalds.com was available in 1994, so a writer for Wired magazine bought the domain and tried to GIVE IT to McDonalds. This proved difficult because they struggled to find someone who knew what the internet was, and whether they should take this random caller up on his offer.

Old Watson is not the New Watson. IBM’s famous supercomputer beat the Jeopardy champion in 2011. It was the size of a room full of computers. The new Watson is spread across a network of distributed computing “running several hundred instances of AI at once.” For Star Trek fans, think of the BORG: We are WATSON.

“This kind of AI can be scaled up or down on demand. Because AI improves as people use it, Watson is always getting smarter; anything it learns in one instance can be quickly transferred to the others.” 

Crazy fast learner. Google bought a startup in 2015 called DeepMind based out of London. Researchers there taught an AI how to LEARN HOW TO PLAY videogames. At first it plays randomly, then 30 minutes later its better, then 1 hour later it plays almost perfectly. In fact, within a few hours – without any additional coaching from creators – “could beat humans in half of the 49 Atari games they mastered.”

Investment in AI is booming.  Private investment in AI is growing at 70% CAGR for last four years.  Kelly notes that the next 1,000 starts ups will simply take X (think: anything) and add AI (as if you were adding intelligence to it). Smart cooking, smart laundry, smart lawn mowing, smart sports, smart driving, the list is endless.

Cheaper computing. This all requires cheaper and more powerful computing. Even a neophyte like me knows that. What I did not know: graphics processors (GPU) historically used for XBox gaming and Hollywood CGI allow for parallel processing needed in the neural networks which gird AI architecture. Nvidia, the leading chipmaker for GPU seems to have benefited from this demand. Look at this stock chart.

Big data. For the longest time, I did not (perhaps still don’t) know what this meant. Andrew Ng, the ex-chief scientist at Baidu, now an adjunct professor at Stanford explained it like this: AI is akin to building a rocket ship.  You need a huge engine and a lot of fuel. The rocket fuel engine is the learning algorithms but the fuel is the huge amounts of data we can feed these algorithms.” In short, artificial minds need a lot of data to learn stuff.  Thankfully (for the AI minds), there is a lot of data out there (books, pictures, videos, unstructured data) to learn from.

Learning algorithms. This part you gotta read for yourself. Lots here.

Many kinds of smartness. This reminds of me a quote for Albert Einstein, which does feel a little bit odd applying to artificial intelligence, ““Everybody is a genius. But if you judge a fish by its ability to climb a tree, it will live its whole life believing that it is stupid” Yes, AI are smart in many different way.  Kelly says, We don’t know the full taxonomy of intelligence right now.”  Here is the scariest thing I read in the book so far, “Indeed, we need to invent intermediate intelligence that can help us design yet more rarefied intelligence that we could not design alone.” 

Death knell to repetitive work. This is something that Seth Godin says, and I repeat in my classrooms: If your job is easy, you should be worried. Kelly estimates that 70% of the jobs today will be replaced by automation of some form. While that sounds scary – it doesn’t have to be.  That means that we (all of us) need to gravitate to work that is more creative, value-added, unique, human, emotive, fun, and different. Repetitive = robot work.

Race with the machines.  From what you have read, I know this sounds dire.  Seems like it is a race against the machines, but Kelly explains that that the majority of the AI work will be things that only robots can do.  There will be some displacement; heck, Vikram Pandit, ex-CEO of Citigroup, said just a few months ago that 30%+ of banking jobs will go away.  On the surface that seems bad, but Schumpeter would say that was inevitable, and we could better spend our time figuring out higher-value things to do. Run with the machines.

Related posts:

Share This