By Andrew Gordon on July 1, 2015 Dear Early Investor,
Robots are like a force of nature. There's no right and wrong.
But like them or hate them, they're coming.
The question I have for you today is...
When will they have a meaningful impact?
As a consumer, sooner than you think. A few are already commercially available.
Yes, they have a ways to go. In 10 years' time, we'll be laughing at these clumsy early robotic products...
Much like we do now at the early mobile phones that were the size of shoeboxes.
But for investors, there's no time to waste.
My Co-Founder Adam and I put a robot manufacturer into our Startup Investor portfolio late last year.
It's a security robot outfitted with latest-generation predictive analytic and social engagement capabilities. It's pretty impressive technology.
We fully expect its laser scanning, wheel encoders, inertial measurement units and GPS to get much better over the next few years.
Knightscope's robot is an impressive feat of technology. But we're not waiting for traction to develop. This robot is also in demand.
CEO William Santana Li says there is a long waiting list for their robots in the U.S. And at least 25 other countries are also interested in buying them.
Emotionally Intelligent
Knightscope's latest model identifies "anomalies" that might suggest criminal activity - such as aggressive or furtive gestures.
But it's not programmed to be emotionally intelligent... not like "Pepper."
Telecom company SoftBank introduced Pepper last week in Japan. Pepper speaks only Japanese and is not available in the U.S.
It's billed as the world's first "emotionally intelligent" robot.
Pepper has real emotions, says SoftBank.
This is where the technology gets real interesting.
We get our emotions from hormones released in response to external stimuli. (I urge all of you to see the Pixar film Inside Out for a delightful, if somewhat simplistic, treatment of how this works.)
Pepper gets its emotions from a sophisticated cloud-based artificial intelligence (AI) program that mimics these same hormones.
Pepper can analyze gestures, voice tones and expressions - and can also generate emotions autonomously. It processes information from its cameras, touch sensors, accelerometer and other sensors.
Here's a YouTube video of Pepper in action from The Japan Times.
Not bad, right? I've viewed several videos of Pepper "in action." Not the ideal way to check out a product, but it's the best I can do under the circumstances.
The robot gets a mixed grade from me.
It has a nice array of hand gestures and some serious dance moves. But this technology was basically available five years ago.
Its conversational prowess is what I'm more interested in.
Sometimes Pepper sounds like Siri - for instance, when asked about the weather.
Other times it sounds like Rodney Dangerfield on a bad day. It cracks these unbelievably lame jokes far too often.
Primarily for this reason, I wouldn't want Pepper as my constant companion. |
| All in all, I'd give it a C-.
This is a preliminary grade. My ability to examine the product was quite limited. I'm not even sure if Pepper represents a strong or narrow AI. (Strong AI is a thinking machine closer to human thought, versus Narrow AI, which is more like Siri - you can read more about this important distinction right here.)
So I'd be willing to raise my grade upon further exploration.
SoftBank itself admits the robot is still early stage. It says its network will gather data that will help its robots "learn" from each other over time.
That's good. Pepper is sure to have plenty of competition.
Making Huge Strides
Costs are a big reason. They've plunged. Take the single-axis controller. It's a core part of most robots' inner workings. It has gone from $1,000 to $10.
The price of sensors has also plummeted. Navigation and obstacle avoidance sensors have fallen from $5,000 to less than $100.
At the same time, computational power is continuing its steep climb. In about seven or eight years, your iPhone should have the computational ability of a human brain. (So says Vivek Wadhwa, director of research at Center for Entrepreneurship and Research Commercialization at Duke, and distinguished fellow at Singularity University.)
All this is taking place as technology gets better and better. At the low end: sensors, memory storage and wireless connectivity. At the high end: machine learning, language processing and AI.
Both Apple and Google use artificial intelligence to translate speech to text with a high degree of accuracy, even in noisy environments. Another example: DeepMind Technologies. It has developed artificial intelligence that can figure out how to play video games with little training.
The "Force and Fit" Challenge
I made a bold prediction last year. I said the "precise calibration of force and fit" necessary to perform relatively simple manual tasks is still about a decade away.
That was before I saw a study showing how robots learn to grip various utensils from watching cooking videos on YouTube. (Click here for more details on the study itself.)
Now, that's strong (as opposed to narrow) AI. Quite impressive. And we can see below a kitchen robot holding a ladle without spilling a drop.
Still, I wouldn't expect robotic chefs to be invading our kitchens anytime soon.
But my thoughts have changed. I see "force and fit" robotic capabilities being much closer to reality than I did a year ago.
So I'm halving my prediction to five years.
Not surprisingly, startups are taking the lead in robotic products. Just to name a few: DeepMind Technologies, Knightscope, Fetch Robotics, Rethink Robotics, Savioke, OSHbot and Jibo.
Of course, in five years it'll be too late to invest. The critical technology is being developed right now.
Venture capitalists invested $341.3 million in robotics startups in 2014, up over a third from 2013 (says Silicon Valley-based robotics blogger Travis Deyle).
Most of this money is going into pre-revenue companies.
We were fortunate to have found an early-stage robotics startup with growing revenue and made that recommendation available to our Startup Investor subscribers.
Timing is everything, and now is the time to invest in robotics startups. I suggest you look for robotic technology that fills a practical need and addresses a potentially big market.
Invest early and well,
Andrew Gordon
P.S. Can't get enough of robots? Try watching BattleBots on ABC (Sundays, 9 p.m. EST). It's a reality TV show in which people from all walks of life pit their fighting robots against each other. Quite entertaining.
You May Also Enjoy... By Adam Sharp on June 26, 2015 This past Friday, the very first Reg A+ deals went into preview mode. Called "testing the waters," this period allows investors to indicate interest in a deal. Let's look at a real example. Read on...
By Andrew Gordon on June 24, 2015 A startup can have the right idea, the right team and a superior product. But without the right timing, that company will fail. Read on...
By Andrew Gordon on June 19, 2015 Diversifying your startup portfolio by these five metrics will give you a more rounded collection of startups. It lowers overall risk. And it gives you the best chance of building a high-performing portfolio. Read on... |
No comments:
Post a Comment
Keep a civil tongue.