Last week, we had a great time at a Sunnyvale elementary school while promoting CS first — Hour of Code 2017. We ended up playing CodeMonkey instead because they only have iPads. Which does not support Flash. The usual “bullshit” is CS gives you super power. But, it’s indeed increasing more true each day. Of cause, don’t take it literately.
To all responsible parents: try HoC2017 with your kids. BTW, could someone please port it to a Mobile Native App, please? It’s 2017 not 2007, right ;p
It All Starts From A Game
Gaming has always been one of key factor pushing personal computing technology forward since the beginning. For example, TRS-80 (8bit CPU with 4 K RAM), one of “the earliest mass-produced and mass-marketed retail PC” debuted 40 years ago. Hobbyist was one of its key customer based, and Big Five SW turned Fun and Games into Million-Dollar business. Of cause, it is no comparison with whatever mobile phone you have today. Anyway, I still remembered when I first played Galaxy Invasion loaded from a cassette. It was really cool and fun back then.
Today, everyone plays. In 2017, the global games SW revenues is $116.0 billion according to Newzoo. Whereas, US Military spending was 599 billion in 2015. It is obvious without the gaming market, Information Technology won’t advance this quick and accelerating. Beside, making life less boring — priceless.
Furthermore, before Bitcoin Mining or Machine Learning hypes, GPU vendors, such as NVIDIA makes their fortune form Gamers. The fact is you can not build a GPU company over night or years. For example, I never know a real gamer, who will be happy to settle down with an Intel GPU.
Even Smartphone penetration is getting really high, gaming is going to pull the same magic as it’s doing in PC market. It will be even more fun for Smartphone, because there are significantly difference from PC, such as:
- Consumers replace Smartphone more often: 2–3 years
- Smartphone is at least 5x bigger: 1.5b mobile vs 260m PC on the annual shipment.
Gaming The Way To Solve Any Problem
Or provide a proxy at least. There was a story of Lydians fought hunger by games: “on one of the days they would play games all the time in order that they might not feel the want of food, and on the next they ceased from their games and had food”. Today, it maybe still a good idea, but irony to fight obesity ;p
An education scientist, Sugata Mitra talked about the child-driven education. He did interesting experiments by trowing a old PC from New Delhi to South Africa to Italy, and shown kids can learnt automatically and independently. There is also study suggests The Potential for Game-Based Learning to Improve Outcomes for Nontraditional Students. Furthermore, how to leverage Game mechanism to achieve Flow is a popular topic in education today. As Jane Mcgonigal puts it: Gaming can make a better world.
The Whole New Work
As Automation advances, human can move on to act more scalable role or do high value tasks. For example, it is fair to say flying a drone is not like a video game. Nevertheless, there are clearly a lot of benefits even if using video games as a training tool. The trend: the skills are needed for such job clearly leans toward more on video gaming in the future. Furthermore, it will inevitable to be a part of solution on pilot shortage getting worse. The crazy thing is the Top Gun 2.0 might be flying with Drone Swarm in the future. It will be hard, if not impossible for classical Top Guns or a fleet without such Tech. to complete with Top Gun 2.0. This is not even an isolated case. Wherever you check, you will see a similar pattern. Of cause, unless you happen to check a job market that labor is so cheap, so no employer has motivation to invest on Automation.
Furthermore, game is also an important part of ML as least on two aspects. Video games have been used as a playground for ML research, e.g. Deepmind: Playing Atari with Deep Reinforcement Learning. Besides, the cost of each training cycle in a virtual world is much less than a physical world. There is less restriction too, such as time.
As ML today is driven by data and based on statistical model. You can not really “test” it via existing QA methods. Simulation is likely the best substitute. Which can leverage many stuffs from simulation games.
So, what’s your game?
The opinions stated here are my own, not those of my company. Furthermore, considering them are thought experiments because the reality is more complicated & has more factors to be considered.