While there are rules, like less metabolic cost, people bend them, and we are rather hard to predict
Our final pessimism bites us in the ass last, but the most. “We could do X…if only we had funding.” That if only is as important as whether we can do it. A great product which nobody can or wants to pay for is an irrelevant product. Cost cannot be an afterthought.
There is a clear financial incentive for autonomous cars. It’s one reason the private sector is able to do them. Is the incentive as clear in every profession? There’s a lot of money to be made in gambling, yet the poker supercomputer was funded by a grant, and nobody is letting that thing sit at a real table. IBM funded an omnipotent chess player. How much money have they made from that the last 20 years? They then funded an unbeatable Jeopardy player, which they claimed would have endless applications. Yet their revenue has fallen 20 quarters in a row. Just because it can be done doesn’t mean people will pay for it once it is.
The Microsoft Kinect could actually do a fairly impressive version of personal training. And it’s cheap! Around $100. Games aren’t pricey either.
Two years ago I’m in Lake Arrowhead and I start thinking how far this approach could be taken. Having a bit of an existential career crisis with my girlfriend, I deliberate on the thing for hours.
- Thermal vision to assess truly how hard something is (hotter the person the harder the exercise)
- MIT found a way to use the Kinect to measure heart rate
- Pupil dilation to assess how your verbal feedback is going
- X-Ray vision to see when bones are deforming too much i.e. when injury risk has gone up too much
Yet the Kinect has all but been discontinued as a product. Not the personal training aspect, but the entire Kinect! The fitness portion has been fully canceled. The Kinect has barely had a game released for it in the last two years. As primitives like CrossFit have taken off.
Where now we’re at remedial business.
1. Does anybody even care if this AI is made?
Often you don’t know until after you make it. Warren Buffett famously doesn’t invest in technology stocks. Not because he thinks technology is not important, or that it won’t have an impact. He doesn’t invest there because tech based stocks, while no doubt influential on society, are on balance terrible for the capitalist. See: all the airlines which have gone bankrupt.
“I like to think that if I’d been at Kitty Hawk in 1903 when Orville Wright took off, I would have been farsighted enough, and public-spirited enough–I owed this to future capitalists–to shoot him down. I mean, Karl Marx couldn’t have done as much damage to capitalists as Orville did.”
2. Do people care enough? Is it profitable / worth it?
The Kinect sold 8 million in its first 60 days, a world record for a consumer electronics device. Even more than the iPhone. Within a few years 24 million units.
This is why we spent so much time going over costs for AI in part 8. If Kinect sales numbers weren’t enough!? Maintaining those customers wasn’t helping the bottom line. (Remember our data centers.) Xbox Fitness was such a problem for Microsoft they stopped allowing new downloads last year, and you can’t even play it as of July, 2017. They would rather refund customers than continue maintaining the old games. Mere maintenance is causing them that much of a headache. For a digital product!
Mark Zuckerberg thinks bots are going to be big in arenas like customer service. Other AI people aren’t so sure:
“Note that none of my reasons for [bots will go bust is] ‘the AI isn’t good enough yet.’ The issue with most systems like siri is more that they’re poorly implemented. We can build many interesting bot interfaces using modern techniques, the bigger issue in my mind is that its not clear humans want to use them.”
Yann Lecun, the head of AI at Facebook, agrees,
-> This is where we remember to be careful listening to predictions on human behavior from the valley. What did people do in response to music being easier to listen to? They didn’t spend less money on it, they spent more!
I did veer towards thinking the Kinect trainer wouldn’t be too big, at least not a threat to trainers. I now call these types of products “glorified home workout DVDs.” A home workout video is not necessarily futile. But they are what they are.
My favorite anecdote from The Revenge of Analog: Real Things and Why They Matter is Thomas Edison thinking school teachers would become obsolete due to motion pictures. History truly repeating itself in 2017 with talk of online lectures replacing teachers.
I’m not at all confident in anybody’s ability to predict human behavior. I certainly didn’t think the Kinect would die, which it’s on the verge of. (Check out The Upstarts to see how many purportedly smart people, who work in tech (!), who thought Uber and AirBnB were stupid ideas.) However, I wouldn’t mind putting some money on the next bust, the next “cool shit nobody wants to pay enough for,” being virtual reality. I would bet a lot of money there is or soon will be companies working on virtual reality exercise apps.
-> Wearables are already going through their downturn. Pebble sold to FitBit, who then laid-off 6% of its employees. There is about to be a graveyard of wearable companies.
On one side, I imagine people a hundred years ago saying “Wait until you see how good we get photographs. Rather than go to the Grand Canyon, people will pay to look at photos of it!”
Say anything photograph oriented to a person who as been to the Grand Canyon and you’re likely to hear “Oh no no no no no. You HAVE to go. It’s indescribable. No photograph can do it justice.”
Nine part series-
- AI is neat. People are messy.
- Are computers really as good as humans in chess / Go / poker?
- Classification is done by people. Not AI.
- And people are fallible.
- Liability / Are we ok with machines telling us what to do? / Loneliness
- It’s not all sunshine and rainbows.
- The more expensive the gym is the less incentive the gym has to keep us there
- Why the gym hopes you never show up.
- Improved performance methods aren’t too relevant / Improving performance filtering is a dangerous, superfluous, endeavor
- Ed Sheeran laughs at predictive analytics.
- Why you’re unlikely to find the perfect program yourself, and why AI might not be better than a trainer
- Machines can tell you what to do, but can they tell you why?
- Thinking about why other countries have more trust in their healthcare system
- The electricity bill could be insane
- Incentives still matter. Voo doo economics.
- Is any company that says it’s green, actually?
- While there are rules, like less metabolic cost, people bend them, and humans as a market are rather hard to predict
- What happened to Xbox Kinect?
Posted on April 19, 2017