Dan has written a few posts recently (here, here, here, here) on how the semiconductor industry is adjusting to changing needs for their technology as the world becomes more wired (the internet of things). That got me thinking a little bit about the hype versus reality of some of the examples we’re already seeing in our daily lives on the product front (not the microchip inside), and what that means for investing in some of these big ideas.
Let’s start with the artificial intelligence space. This may seem like a futuristic concept many years away, but believe me, you are already seeing it in your daily life. The problem is what you’re seeing kind of sucks. The two biggest examples in my daily life are Siri on my iPhone and Amazon’s Alexa on my Fire TV. Both are kind of useless apart from very specific instructions like “Call Dan” or “play The Clash” on Siri, and “What’s the weather today?” on Amazon Echo/Alexa.
The reason for this is there are so many working parts to what’s going on when you speak to a computer. It’s such a massive leap in computing interaction that it makes Google’s first search engine seem quaint. Think about Google’s first version of their search engine. It simply scanned the internet based on your search words, and ranked results (sites) based on how popular they were (the main part of the algorithm was how many other sites linked to that site). This simple idea was a massive leap forward in our access to information and has changed the world in unimaginable ways.
Now think about what happens when you speak to Amazon Echo or Siri. First, it has to recognize what you are saying, which is no easy task with your accent, the way you phrased your instruction and any background noise. It then has to decide exactly what you want and deliver. If it is unsure, it then has to ask you questions to get to what you are looking for. This usually ends up in you wanting to throw your phone. Traditional internet search never had this problem because the interface is so different. When you do a Google search it gives you pages and pages of results. Your eyes quickly scan those pages to find exactly what you’re looking for. If you can’t find it you change your search to get slightly different results which hopefully now contains what you were looking for. Basically, you are doing a lot of the work for your search engine. And you’re doing it way faster than you probably realize because you’ve adjusted the way you think to make your experience with the search engine better (e.g. you don’t “ask” it questions, you know how to phrase searches to maximize your chances of finding what you’re looking for.)
Obviously Google has done a lot of work to meet you in the middle. But pat yourself on the back too whenever you do the perfect search and get the result you were looking for to the top of the first page. That wasn’t just Google’s algorithm.
But now go back to Siri and Alexa and think about how much harder their job is. And the issue is we’re now expecting this next leap overnight, and we’re really rushing products into this space with an expectation that this will be a quick process. This is fine if you’re talking to your thermostat or lights, but when it comes to your car being controlled by artificial intelligence things start to get a little more real. Basically what I’m saying is the real world is complicated, and humans don’t realize how smart we are (and always will be?) compared to computers.
Think about coming to a 4 way stop sign, and all the nods and eye contact that happens between drivers. Now re-read an article hyping Tesla/Uber/Google/Apple self driving cars taking over the roads. Remember that this doesn’t just happen overnight. Like one morning we all wake up and every car on the road is driver-less. So now go back to that 4 way stop sign. What the hell happens when 4 cars arriving essentially at the same time and one’s an empty driver-less car? I assume after the first couple of crashes the humans just let the empty car go through. Are we gonna let robots win that easily? Let’s at least put up a fight like Terminator 🙂
So my point is a lot of the hype of these technologies gets way ahead of the reality of our lives. And AI in particular seems way ahead of the reality of how smart humans actually are. Google search is simply a tool that humans, much smarter than the computer itself, figured out effective ways to put to use. Making the leap to where computers are doing that work for us like navigating complicated roads with humans in other cars, and speaking to us in any complicated way? That’s a much bigger leap than you’re being led to believe.
All of this will happen eventually. It always does given infinite time. But for investing in some of these ideas, there’s a reason why there’s more of a frenzy around microchips powering these products that the consumer products themselves. Smart money figured out that the chips were in demand, regardless of how crappy or good the product they’re going in were.
As far as who I’d bet my money on the long term winner in Artificial Intelligence? I forgot to mention the other example of AI in my daily life. That’s the Google app on my iPhone. I find it head and shoulders more useful than anything Siri or Alexa can do. And the reason for that is Google has been collecting all that help you smart humans have been giving them all these years. They’ve forgotten more about us than Apple, Tesla, Facebook and Amazon could ever learn. It will be hard for other companies in this space to ever catch up with that kind of data accumulation and my bet is on Google intelligence powering most of your favorite smart devices in the future.
Today, Google announced a whole new line-up of devices in that vein (phone, connected home etc.) but we’re still sort of in the “OK Google, what’s the weather today?” phase on all of this.