Go Three Dimensional and Up

The Chinese Go game has attracted the world’s attention recently when computer beat the world human champion. It also got me thinking about a higher dimensional version.

Essentially the Go game is a graph whose vertices starts with transparent color but may become black or white when stones are placed in the vertex as the game progresses. Once colored, a vertex can only become transparent again when the stone is removed due to loss of liberties. A stone’s liberty is the number of uncolored neighboring vertices on the graph and it becomes captured when its group has zero liberties. Captured vertex immediately return to transparent. Connected stones having the same color is a group and combines their liberties. All liberties of a group must be taken before any stone of the group becomes captured.

Players take turn placing stones on vertices. Stones having no liberty are removed by the board. Game ends when at least one player resigns or when there no play remaining that would not reduce players’ stone count plus liberties. Winner is the player having most stones on the graph at the end.

The traditional go board is such a game played on a 19×19 planar graph located in the square from origin to (18,18) having nodes at grid points and edges between closest neighbors.  Internal nodes have 4 neighbors. Edge nodes has 3 neighbors each and corner nodes has two neighbors each.

Therefore, we shall disregard planarity and uniformity. Consider, for example, the game played on a graph that looks like a donut when laid out in 3-d. Internal verteces have 6 neighbors, external verteces has 5-neighbors.

Shall we play the game?

The 6-2-10 system 

More aptly, a 6-14-22 system. It doesn’t roll off of the tongue like 6-2-10 but would be less confusing. Some alternatives are 0-6-12-18 and 0-4-8-12-16-20. I have come to refer to these other systems as p6 and p4–period of 6 and 4. They each have different levels of flexibility and efficacy towards collaboration, sleep cycle, social segregation, etc.

I think of this now because I reflect on my schedule: I wake around 0600 daily now, brush and dress. Commute to office immediately and achieve chair-to-chair time of less than 70 minutes. This is about the only time this commute can be done in O(1hr) with no lingering constant that add up to half an hour. I typically arrive at home after 1900 That’s an 11 hr day followed by sleep at 2100. I spend next 3 hrs sorting through a foot of past due and bear past due bills and the November ballot because I haven’t really been able to get home by 1900 for the past month. Family time is minimized…

In my futuristic world, otoh, each person is allowed to work only for exactly 8, or 6 or 4 hrs. I wonder if I counted commute time.

Where I work is too expensive to rent OR buy at my income–which I’ve been told is above median at the institution–at least half of the people whom I work with have equal or worse problems. I mean some people don’t even have families!

There feels like an invisible force pushing the state of the market to continue to be this way. Is it extraordinary for us to ask the invisible hand to grow some brains?
The question then becomes, writing this in my dreams, were we to find ourselves in this idealistic world of mine, where we are able to organize, think and act rationally as a kind–human kind–can we do better?

We can do better! right?

In such a bettered world, would commute be on company’s dime or my dime? (s/dime/time) would it matter? I guess it depends on our criterion. I would say it’s on company’s time, company would say it’s on my time, and the rule maker says: we can impose rules on maximum working hour and minimum working wage, we can certainly enforce some penalty on commuting. 

Suppose it is proven beyond a doubt that commuting is detracting from society–and that’s not a sure thing, we can simply charge a bi-tax, similar to social security tax, where for each minute of commute, the commuter and the destination pays a tax to the government. Since commuting detracts from society totally, all parties involved must pay.

Oh, and also to amend the previous design on overtime. The same bi-tax can be applied to overtime that the employee gooses to take and the company allows him to take. Again under the assumption that we can prove with sufficient certainty that said overtime is detrimental to society.

I am happy now knowing full well that I will jerk awake to my alarm shortly… And this utopic dream sequence of a government penalized, socially motivated, totally organized and advanced commute in which I am riding on air in the hyperloop in a lazy-boyish sofa, basking in lens-flare-free source free radiance, vibration-free, uv-free capsule, surrounded by merv-20 hepa rated air, working on a screen floating in front of me adjusting display to my eyes focus instead of the other way around; this dream sequence will pop and I am back in my living room(trying to not wake my family), surrounded by only merv-11 rated air, eyes still hurting from the previous day’s straining, joints hurting(and that’s completely normal for a man my age) vastly disappointed at the lost of something lovingly wonderful. Absorbing this silent morning, now that the alarm is extinguished, savoring the difference between it and thence that I have awaken just.

I am ready to commute.
The sun has not risen for it is winter in California. 

But, I will rise.

I will rise.

I will rise!

Planetism and ML

Just found an interesting google blog on equal opportunity for machine learning.

Some thoughts to head off discussions. So it would appear to me that an oracle model can never be a planetist. (Now that we are going to Mars and all, let’s reframe the problem in terms of planetist and planetism–discrimination based on planet of association)

The fact is that if a model is predicting loan defaults with 100% accuracy, then it cannot be a planetist. The challenge to that will surely be that we do not believe it is an oracle. How does the oracle know if a person will or will not default on a loan? It might be a planetist all along? We will never know know because 0 is not large enough sample to be representative when you don’t allow any Marsians to get loans.🖖🏻

A possible symbiosis

So… To take myself out of the nitty gritty for a moment it seems still possible for there to be machine-human symbiosis.
Some number of decasdes ago, while in highschool I wondered about this matter. At that time we had 80486 computers and Ram in the megabytes. My conclusion for the computer replacing many human jobs or functions or that they become more valuable than humans was that it is inevitable that we strive to live with them. Much like white people have learned to live with black people and that we care for endangered more than we care about some peoples economic welfare–we can learn to live with computers as equals and sacrifice some human pursuits to that end.

Alas that was more than two decades ago.

Last night, I heard Kaifulee’s Lee address a crowd and discuss AI businesses in China. His answer was somewhat sombering. In paraphrase, I believe what he said was that in the age of AI, where computer product managers are gradually replacing human functions that requires less than 5 seconds of human thought with AI, the future human will be in things that take more thought, but under time and audience pressure he gave art, music, appreciation of art, things that require a personal touch, as examples of new jobs.

Personally, I can empathize. As one who sees foremost advances of AI and robotics, and a person whose job is partially to make money by disruptively using this new technology to replace old systems, I can definitely see his human_replaced_counter ticking up and projections for it to grow very fast.

(Much else was discussed at his talk, of course, this is just a short question at the end of the talk)

It might just be me, but I can almost see tears as he answered this question. There isn’t a comfortable answer when you have to admit that someone, or something, else will definitely beat you at something. When that something is your livelihood, and there is even a Robo-vc now, it is harder to be objective. Now of course we do not want to be paralyzed by paranoia. But we should think hard! 

What will we do when machines take over our jobs. What do we do when machines take over our lives and live for us?

 And where is the problem? Why don’t all the Uber drivers replaced by bots go on welfare and go to free community college. They can get their degree in the comfort of their home on Ng’s coursera or Thrun’s Udacity. They can learn to do something else, perhaps learn to write programs? Learn to Code as it is now colloquially called. They can take my job after that. That way I can go get my md’s and ph.d.’s and go heal people or philosophy?
(Footnote: what we don’t want to see if a flow of talented and educated people to jobs they over qualify for: coders driving Uber, md’s ph.d.’s writing code. The prevalence of this phenomenon stirs a deep dark anxiety that I cannot name. The decisions to do so are individually very rational. However it would seem to me that society’s investment in educational infrastructure to created these md/phd’s have not achieved sufficient ROI, for the society’s sake. I.e. If they train in physics, should they not attend to physics matters that the training was designed to do? And should they not be related with completely unrelated subjects? If not why do we have so much investment in physics higher education? S/physics/another subject/; again it is worrisome if society is this way but no worries for the individual or the institutions involved in this process, each of which is arguably producing maximally and with the best of intentions)

Would that be a blast? It’ll be like 24th century of Star Trek: we will have no wars, no worries about money or scarce resources. With advancement of technologies, society is advanced. We will no longer struggle against fellow man but against a greater obstacle. We will only strive to better ourselves or humanity.

Such a grand future awaits us!