A Discarded Paragraph From a Machine Learning Paper

Cut from the introduction:

Finally the economic incentive. By using machine learning for this task we are freeing up the man-hours of dozens if not hundreds of astrophysics PhD students whose sole existence is to categorize galaxies, quasars, and red giants for their thesis advisers. In this manner we follow the footsteps of Bertrand Russel who advocated for shorter working hours in American industrial revolution of the 20th century \cite{russell-idleness}. Today, thanks to the increasing plurality of machine learning tools we are again questioning what is a necessary weekly hourly allocation of work.

AI Could be the Social Atomic Bomb of the 21st Century

I’ve seen some impressive AI demos across the web where researchers teach an AI to reproduce human faces or lip sync audio over pre-existing videos. Impressive, and unimaginable, several years ago for sure. Similarly unimaginable today is the way in which social networks and other tools have been used to undermine democratic systems through the propagation of misinformation. I fear that these two things have a shared future, where fake news is generated by AI’s is indistinguishable from the real thing.

One of my first classes in graduate school was artificial intelligence, after the course summary the instructor and the TA’s held a bit of a Q and A session. One of the questions that stuck out to me was along the lines “people like Elon Musk and Stephen Hawking are giving us really big warnings about AI, could their predictions be accurate, should we be worried?” The panel looked at one another for a moment and they gave several answers in so many words saying: “they’re not in the field and don’t know what they’re talking about” and concluding “there is always an off switch.”

In college there was an article assigned as reading in my introductory CS class, published in Wired magazine in 2000: “Why the Future Doesn’t Need Us” by Bill Joy. In which, Joy stresses that we could be on the verge of creating the social or existential equivalent of the atomic bomb for the 21st century if scientists (computer and bio-scientists in particular) forego ethics or if standards are relaxed in anyway. I make it a point to reread it every couple of years, it rang true when I first read it in 2009, today that ring is deafening.

NRA slogans are simple logical blanket statements but they do not directly refute criticisms of gun control. A recent one like “the only thing that can stop a bad guy with a gun is a good guy with a gun” is based on a lot of assumptions like how a good guy can never become a bad guy. Banning guns nationally –as Australia did in 1996– also stops bad guys with guns. An older one which I don’t see very often anymore is “guns don’t kill people, people kill people.” This is true, guns -inanimate objects- don’t just walk around autonomously going off, but can be more accurately phrased “Guns don’t kill people, people with guns kill people.”

I think the TA’s of that introductory AI class are largely right in the way that the NRA slogans are right. There is always an off switch, and AI’s do not possess any physical autonomy no matter how much digital autonomy we give them. Yet the larger point is being missed about the dangers of over-developing AI. “AI doesn’t harm people, bad actors using AI harm people.” Social networks don’t spread propaganda, bad actors, AI algorithms–designed to make us click as much as possible–and unsuspecting users on social networks spread propaganda. When it comes to these complex social systems which AI can run on top of the off switch is difficult to find, and the on switch is often tripped without fully recognizing it.

In another 20 years, will AI be a tool for empowerment, justice, and productivity? Or a tool for suppression, discrimination, and distraction? I therefore believe it’s critically important that AI researchers think critically and ethically about what they are building. It’s no longer a question of what can AI do but of what someone can do with it. In the context of some tool or model, what malicious actions can be performed with using it, how to recognize it, and finally, how to turn it off.

“Game X” Is Only as Fun as…

Keystone gameplay elements, called so because if they’re taken out the game isn’t that much fun anymore. I started thinking about this during a bitter game of Overwatch and I’ve applied it here to bunch of other games:

Player Unknown’s Battlegrounds (PUBG) is only as fun as the loot you find in the first five minutes. If you’re in the top-20 on Miramar without a a scope you’re going to have a difficult time going forward. The same is somewhat true for the Erangel map but there’s much more consistent cover and loot is a little more dense from the start too.

Fortnite is only as fun as the rarity level of the guns you find. Unlike the staggered –textile mixed– loot spawns of PUBG, the potions (armor) and type of weapons you find in the beginning do not so much influence your fun in the long run.

Overwatch is only as fun as your teammates are willing to cooperate. This is independent of game mode, there’s this strange contrast between Quick Play and Competitive game modes. In competitive, people can be resistant to changing characters which put the team at a deficit, even to the point where they’ll spitefully throw the game. In Quick Play, people play the characters they want to play, leading to deficient and unusual compositions of all DPS and tanks or or all DPS and one secondary healer. Usually, the team to take some composition initiative get’s momentum and wins.

DOTA 2, LoL, Heroes of the Storm (MOBA’s) are similarly fun to Overwatch. It’s only as fun as your team mates are willing to cooperate. Unlike Overwatch though, it’s much harder for a single person to carry the team.

Hearthstone is only as fun as the cards you initially get and how well they align with the current meta. However, you will always be outspent by the people who take the game seriously.

World of Warcraft is only as fun as your character’s level and/or your guild’s ability to organize and raid. There are some truly amazing sights through out the game but until you’re of a certain level you’ll never get to see them. The even larger content needs to be done through raids which need a guild to do. Once, I was lucky enough to raid Molten Core with 39 other people and it was very exciting.

Path of Exile is only as fun as the quality of your build. The game has a staggeringly complex skill tree, while most basic builds can work through the first and second difficulty levels the third one will really test it, and it’s usually where I’m forced to scrap my characters. There are plenty of sample builds on the forums but they too often require unique equipment which is extremely expensive and drop very, very, rarely.

Kerbal Space Program is only as fun as the length of your patience and determination to reach other planets or build space stations. With a little experience missions to the Mun and Minmus can become routine. However missions to the Jool and the far outer/inner planets consist of lots of time waiting for comparatively small maneuvers only to get there and find out you don’t have enough fuel to complete the orbital burn or you forgot to deploy the solar panels and you’re out of power. Being honorable, you revert the save to vehicle assembly and redesign your rocket only to go through that whole process again.

Rocket League is only as fun as the mean skill all players compared to your skill individual. If your skill is lower than all other players it can be a bit daunting to have cars of all variety flying around while you drive in circles just trying to get some boost. This does not apply to 4×4 modes, anything can happen there.

Minecraft is only as fun as your desire to explore and create. The caverns are deep and endless, but what is over that hill? The sun is low, the monsters are coming, are there enough torches? To my 1:1 scale model of Helm’s Deep!

Starcraft 2 is only as fun as you know your build order and it’s the right build order to counter your opponent. Contrary to popular belief, unless you’re Master league or higher, the game is not so heavily depended on super-human actions per minute (APM) and micro ability.

X-Com[1] is only as fun as your ability to best position a squad of units against an unknown enemy. Though ultimately, it relies on the secret RNG which determines the actual probability of your shots to hit the target.

Civilization (chose your version I through VI) is only as fun as the AI is able to be even to you point-wise. An experienced player playing on Settler (easiest) difficulty won’t have much fun as they stomp across the globe. On higher difficulties, it feels as though each move is precious and the struggle against your digital strategic equals will not be firmly decided until the last move!

No Man’s Sky is only as fun as you can convince yourself you’re having fun collecting resources in order to build better tools used to collect more resources to collect more tools.

  1. This is only applicable to the current iterations of the game, the original games in the 90’s have some mechanical differences which do not apply.