The San Francisco Report
9/30. High impact innovation in two days
Sometimes it’s just in front of you. In the last tow days I witnessed two events which will have a great impact on the world. First, I attended a machine learning meet up in Silicon Valley. Andrew Ng, one of the main academic figures int he field, was there. Others such as Jike Chong, the leader of the meet up was there, too. Next day I visited the Tesla factory in Fremont. They didn’t show cars. They showed batteries for energy storage. Both experiences have the potential to become “I have been there when it happened” events for me.
Energy storage is a key component for electricity to free itself form the grip of slow moving and sometimes corrupt utilities. The industry is large. If Tesla can change that market, their impact could be even bigger than what they are doing to the auto market.
Their approach is simple. First, they took the batteries from computers and laptops and stacked them up to make a car. Then they made a better car with more battery. Now they are planning to make even more batteries, more cars and make everything cheaper. Sounds easy? Well, it’s not. But the approach CEO Elon Musk is using is simple to understand. I appreciate that. The impact is huge. This is a high impact, low complexity, high technology investment. The expected returns are high.
One key question is, why are they so cost competitive? Why can they offer something for less than others? The answer to that question is not clear yet. Maybe they are not much cheaper. Maybe they are cheating. Maybe they hope to cheat their way right into volume production. Some people question their numbers. This all might be true. But what about the product.
But the product is great. Execution is great and costumers love the product. So, there is not much to complain about the product. When the product is right usually many good things happen. Apple has lots of hick ups, but the products are good. Nest makes things we really don’t need, but the products are good. Whole Foods is self righteous but the product is good. When the product is right, many good happen.
Machine learning is even better. I see some of the most dramatic advances in computer science right in front of my eyes. Problems like speech recognition, computer vision and machine reading are ready to be solved. In fact, companies such as Google, Facebook and Apple have already applied some of the techniques of machine learning to problems such as speech and vision.
It’s always hard to predict where exactly a technology will strike. But this one for sure will have an impact. Speech recognition is huge. Imagine devices able to understand your speech. Imagine writing and translating text to speech and the other way around. This is huge.
Imagine computer vision where cameras take a visual and commuters interpret the content. Tell you who it is, warn you if you’re driving the wrong way or even drive themselves. Machine learning is the key to self driving cars. Other applications such as automated reading are possible. Computers can read text and interpret. Reading at scale has never been done before. Reading at scale offers extremely new ways of learning.
Imagine a hedge fund reading one company filing. The company says “Our major customer from China is not paying the bills”. That doesn’t mean much.Now imagine reading at scale. After having zipped through 100,000 filings this morning the computer says that interestingly over 100 companies have filed revenue warnings because of slow paying customers. Now, this is powerful.
We are excited.