Reading up on AI

In my downtime before uni, between doing some hiking and coding, I figured I'd get into the Cognitive Systems vibe by reading up a bit, and this seems to end up often at some discussion of super-intelligent AI. I'll probably be writing up a bunch of these over the period of my course, but first up: one observation made at lesswrong from a GiveWell director - in particular, distinguishing between 'agent' (active) and 'tool' (passive) AI, see objection #2. The basic idea is that it is much safer to design 'tool' AI, as it just gives you data and does take action itself, and I'd recommend reading that article first and thinking about it before returning.

I think overall it's a good way to start thinking about the problem, but my take on this can be illustrated by the following tale (giving examples through stories seems popular in this area):

* * *

Investech was a company of researchers and quants, who had a small amount of success writing computational models of equity prices (shares, bonds, resources, ...) and used these to generate income from trading when the models predicted shifts. Over time, as more research was added, more patterns found and incorporated, Investech's model grew and the AI was given access to more and more information - company reports, social media posts, open health and weather data, etc. This culminated in all trading combined into a single AI that incorporated all their research and data, nicknamed the 'Oracle Oracle' (or O2) after the system it ran on. 

Not wanting to repeat some of the massive automated trading problems in the past, O2 was designed just as a tool - when provided with a certainty cut-off, it would display an equity and a price, that it forecast to the given certainty would be the case in the future. The human traders at Investech would then use this to make the actual trades, based on what levels of risk were acceptable.

The initial tests run by the QA team worked great: Given 5%, it provided a really high price for a resource future that would shoot up only if one country received unusually high rainfall. Given 60%, it had predicted the share price of a large fashion label would dip 25%, and sure enough, when the newest line was spurned by a few celebrities, support indeed dropped; much to the pleasure of both the CFO and O2's tech lead (now CTO) at Investech. Given 99%, O2 chugged along for a while, then hit a preset timeout and reported no results.

O2 was mostly then used internally to generate a number of semi-likely events which guided basic hedging strategies that ended up performing well. Due to the recent success of O2, Investech had started looking to grow their client base - and due to the same success being disclosed on their earnings report, they had been able to assemble a meeting of a number of the more powerful hedge fund managers to demonstrate the power of the Oracle Oracle.

It took a while to organize, but eventually Investech were able to assemble the meeting to show off O2 in front of these 20 or so managers who on paper controlled a sizeable portion of global markets (and off paper, probably even more). The explanations went well, the numbers for the past few quarters checked out, and most seemed happy with the techniques of converting the confidence levels into trading strategies. The main concern was over confidence in O2, and how much trust could be placed in this system that ran everything but, due to being AI that had evolved over a long period of time, was not well understood or easily explained / debugged.

The time had come for a live demonstration - Given 50% confidence, it believed price of corn in the mexico would remain stable - it seemed the market was fairly efficient, at least for corn. At 80%, the Oracle Oracle predicted shares in an energy storage startup would double; many were surprised by this, but the startup's technology was unproven so it seemed that O2's surety in the viability was quite high, having analysed a lot of research. 

While most of the managers were discussing or taking notes, one joked: "So, 99.99%?". Knowing of the earlier testing, the CTO laughed as well, keyed it in, and started to explain their corner case handling while O2 began its calculations. They were interrupted by results on the terminal:

99.99% Prediction: NYSX / Investech / $0.00

* * *

If you made it this far - that was my long way of saying that, for an AI which has some form of empathy, giving information is taking action. Providing it can predict a human reaction to the information it gives (which I'm assuming will be the case - more about prediction in later posts probably) then even without the AI-box issue, a tool-based AI can still 'take action', even if using the medium of humans rather than bits. It's much slower, but can only be avoided by not having empathy (i.e. prediction of how humans will react), which is pretty much impossible for a superintelligent AI, as humans have this to some level already... A more humorous/shorter version of this is Monty Python's Funniest Joke sketch, or a quote from The Interview.

Anyway, back to reading/thinking too deeply about this sort of stuff. Apologies if it's not the happiest of posts, I don't want to scare anyone with AI talk (Terminator/Ex Machina/Age of Ultron etc. do that enough already). I'm definitely leaning towards the AI-will-be-good-for-humans-in-the-long-run side of things, although we still have a lot of learning / understanding to go through before it's not still mostly guesswork (hence the degree!).

In other news, I think I accidentally got a part-time tutoring job over here (went to ask for details about the course and what they were looking for, and ended up being interviewed) and I'm moving into my permanent place soon, so within a month I'll finally have access to all those other clothes (+furniture etc.) I shipped over, and not just the 7 shirts in my suitcase, yay :)

Android PatFute

While reading my past blog posts due to its anniversary, I found in one of my posts from 10 years ago (mid-year 2005!) that International PatFute day is July 7th.

For some context, about 11 or 12 years ago, when first learning programming, I was teaching myself by writing a game for windows (in Visual Basic, no less). Having way more maths skills than artistic ones, it ended up being a physics simulator thinly veiled as a soccer game - the edges of the pitch are fixed solid lines, and the ball plus all the players are circles, bouncing off everything else.

Then, as you can see in my blog, I wrote it again when learning C++ and 3D graphics - this time the players were cylinders, and the ball gained height on kick, but everything else worked the same way. Little did I know, that this foreshadowed the real sport of Bubble ball many years later.

When first learning web development, I then re-wrote a javascript version using web canvas, but sadly the code for that has been lost. I later tried adapting the 2D game to touch screen, coding an iOS version (now learning Objective C) but never got it to a stage I'd put it on the app store - plus I didn't like having to borrow apple stuff to write and test it, so that version never saw the light of day, although I did end up trying to work around that with a PlayN version (old code on github) however that setup proved more painful than first hoped, and the PlayN project itself lost popularity. side-note: it seems to still be in active development though, so perhaps I should take another look.

Anyway, this leads me to International PatFute day 2015: finding some spare time over the last week, I decided to give it a proper shot, this time brushing up my Android development.

It's now available to test in the play store, so I'm looking for people with Android phones or tablets to try it out. Just to set expectations at the right level: I'm a coder, in no way artistically inclined - so don't expect anything from the graphics side of things. I'm keen to get playability feedback though: did anything crash? does the physics work ok? are the controls as expected? was it fun?

In addition: how it works on other devices (I only have access to my Nexus 5 right now, and an N7 tablet in a few weeks, but more screen sizes and aspect ratios would be good. Plus of course, any features you'd want (more teams? different game options? scoreboards / achievements / stats?), or suggestions on what graphics are actually worth trying to put an effort into.

For those happy to test: more details are available in the "Touch Footy" Google+ community (that's how they control test app access). I plan to keep working on this a few times before uni starts, and I'm assuming some time after that too - so for the non-testers, hopefully this will eventually come to the proper play store, available for everyone! I renamed the app too, PatFute is a bit non-obvious, and I felt I had to capitalize on the 'touch' pun...

And to the coders: I have a private github repo with all the code - I plan to publish it shortly after the app reaches the public play store, but I can add you to it before as well if you're interested.

Happy International PatFute day!

New research proves that thing you agree with!

I'm assuming readers know what Facebook is (or, social media) - if you don't, this probably won't be interesting. For those that do, you may have recognized that when you click to read a shared article, underneath will appear similar articles you may also be interested in. The idea for this post came about after doing just that, and the following was the result:

It's probably trick to read at that size, so to summarize:

1. The original link was a Jezebel article about research published in September 2014 in Personality and Social Psychology Bulletin, where they quote the abstract: "On a daily basis, when people felt more insecure about their partner's feelings, they tended to make their relationships visible".
2. The first suggested link was from HuffPo about research published in July 2013 in Social Psychological and Personality Science. It doesn't quote the abstract but summarizes it, including how the study "found that participants were more likely to post information about their relationship on Facebook on days when they were feeling particularly satisfied with their marriage."
3. The second suggested link was from The Atlantic about research presented at a 'Society for Personality and Social Psychology' conference in August 2014. It covers most aspects of the research link, including both that it "found that those satisfied with their relationship are more likely to use Facebook to post couple photos and details of their relationship" as well as finding "Individuals high in Relationship Contingent Self-Esteem (RCSE) are also more likely to post affectionate content."
4. The last link is direct to the research in the first link. (yay! good to know people share this too).

I clicked the initial article hoping to read about the research methodology and precise results - e.g: how correlated is it? How many people were studied, and how was their insecurity classified? Unfortunately, all research above requires payment to view :( I might investigate that in the future, but for now, I - and presumably 99.99% of those reading the news articles - can only go on the abstracts.

Logo for http://www.spsp.org/

A few things surprised me - firstly, the bulletin/publication/conference is all run by the same group: the Society for Personality and Social Psychology (SPSP), who runs the conference and publishes both the bulletin (PSPB) and the journal (SPPS). I assume they're legit, and I guess there's just not much competition for publication in this field, but still, it is not what I had expected. What is more, the first two articles seem related - their published titles include "Can You Tell That I’m in a Relationship?" and "Can You See How Happy We Are?", and both include Amy Muise as a co-author.

On the research side, the first two articles have quite different conclusions, but when combined they say the same as #3 - that is, for the people posting more about their relationship: some are doing it because they're happier, and some are doing it because they're more anxious. My guess is you could research to find more correlations too: posting about your relationship is probably correlated to your level of facebook usage, your access to a camera phone, how much interest you take in your relationship vs other things you might post about, etc... 

The first two news articles however only bring a one-sided approach - in particular #1, which doesn't mention the contrary research that #2 had previously covered. What is more - nowhere (other than hopefully in the published research, which the average person can't access) is there any mention of the strength of correlation. From my own anecdata, people often read stuff like article #1 and now think: "See, if you post about your relationship, you must be insecure! Science says it's true!" rather than "Hmm... so this means you're slightly more likely to have RCSE than someone who doesn't post as much, and there's a bunch of other reasons you might share that many photos."

That said, I know there's nothing new to the problems with science publications and journalism (this talk is great), so consider this just a rant/friendly reminder to avoid reading too much into what a journalist summarizes from a single study. Now, time to go back to reading even more articles before starting my degree...

 

Internationally taxing

I've advocated a number of times (e.g. this post) the benefit of analysing an issue by considering multiple viewpoints, and have been annoyed how so few articles do so (i.e. I don't trust anyone's analysis if they can't name a few reasons why those who disagree could/should do so...). Time to try living up to my own standards.

Source: http://www.hhcpa.com/services/international/international-tax-compliance.php

The topic? Something that has come up a lot in Australia at least, which is the taxation of multinational internet-economy-based corporations (MNC). Full disclaimer: I used to work at one, but was in no way involved in the financial/political side. I no longer work there, and the views contained here were mine beforehand too, and not based on any knowledge internally.

To start, a brief summary: The internet is widely used in Australia (like most of the western world) and there are a number of companies founded internationally who profit from selling products or services to people in Australia. In discussions for the recent budget, the Australian government proposed that funds could be obtained by cracking down on 'tax avoidance' by multinational companies who have set up profit distribution systems in order to reduce taxation payments (places like Apple, Google, Microsoft, ...). For those unfamiliar with what sort of stuff this involves, I'd suggest checking the Double Irish wikipedia page or the multitude of youtube videos on the subject.

To structure this post, I'll try presenting points from either side - labelled 'status-quo' = corporations are ok as-is, and 'pay-more' = corporations should pay more tax - and discuss it, with my conclusion plus the problems with my approach. I'm also assuming that most readers are on the pay-more side, so you should probably watch this video first. It's from an earlier instance when something similar came up - an Australian senate hearing from 1991, with Kerry Packer being asked about tax avoidance for ACP (now PBL).

Ok, jumping in to the arguments, starting from one brought up in the video:

status-quo: MNC are just following the law, there's nothing wrong with that.
I agree with the first part, but not the second, and you can hear why in the video above: "you were doing so in ways that were contrary to the spirit of the law". It's an age-old debate - yes, it is technically following the letter of the law, but not the spirit.

This is a tricky area, so I'll try a more tech-related analogy: I've heard the comparison before that writing legislation is like writing code - you approximate what you want, and then iterate to patch over the holes as they come up. Except that with law, it's a really old code base, changes take years, and it's rarely refactored. In this comparison, tech companies are not happy if you exploit bugs in their systems, even if you are technically 'only running the code as written'. Usually, if you profit off finding these bugs, you can be charged.

Interestingly, the MNC themselves have many contests and rewards for finding these problems. Suggestion: perhaps the ATO should do likewise? There's the question of whether people would report them, or make more privately helping others use the exploit, but this is the case with tech bug rewards too. Still, as a user you're supposed to follow the spirit of the code, so I feel the same should apply to law.

The downside is that this opens up a lot of grey areas - either you don't charge people for breaking the spirit of a law (in which case, they probably still will? perhaps there needs to be some sort of public shaming as a side-effect :)) or you need to have people decide when a breakage occurs, which seems ripe for abuse. I'm not sure a good fix here, it seems like something those more involved in the legal system will have good ideas for though.

pay-more: MNC should be found guilty of tax evasion and fined.
This is where it gets tricky - I'm assuming that the ATO knows about this loophole, so it's similar to when white-hack hackers report an issue but it still doesn't get fixed. What are they allowed to do then? There have been some examples in Australia, e.g. where someone reported a Victoria police website issue, it didn't get fixed, and so they downloaded 600,000 records of PI. I don't believe any criminal action was taken there, and while it's not exactly the same, it seems fine in the legal sense to try to patch this issue quickly and only find people guilty going forwards (which seems like will be the case in Australia). Note that public opinion != law too, so feel free to keep claiming the actions weren't the best to take. In terms of the problems with this conclusion: not having a fine or criminal conviction definitely won't act as much of a dis-incentive to continue the act, but I can't think of a good way to avoid that without breaking conclusions above.

pay-more: MNC should be taxed just like normal businesses.
Taxation for stuff like this is really tricky. Traditional businesses, especially selling physical goods, are much simpler: Person A buys a good from Shop B in Country C - you transfer tax on the profit (GST in Australia) to country C, which helps both having more happy people to buy and more useful businesses to sell things.

Compare this to something like online adverstising: A company (say, based in England with offices in France and Germany) sells advertising placements to another company (say, based in France with offices in Belgium and Germany), which is then charged only when an user in Cameroon (tunneling through an IP in Niger) clicks on it. Where should the tax be applied there, in a way that helps the governments of the user, the business, and all the connections between. Does it change if the deal between countries is made in France or Germany (or via an email exchange between England and Belgium)? While it's not the case that $27B revenue should result in $193M in tax (0.7% doesn't seem enough for the user's country to make more, happier users) it's also not going to be as easy as simply applying normal rules to online business, and will result in less tax locally (which is then countered by growing a local IT services industry). Plus, to repeat the points above - whatever the ideal taxation systems are, goverments should not be too slow in their implementation, but also it companies shouldn't require having complicated revenue relocation structures.

status-quo: Why pay more tax if the government is so inefficient?
This one's really complex, so included here will just be brief points, my guess is more will come up in comments. First: government is less efficient than corporations, this is not surprising. There are a number of things that I'm very happy to be privatised (e.g. supermarkets, banks, ...) - competition has been useful here. The problems are twofold: one is where monopolies are much more efficient - e.g. I'm happy for a national (in the most part) post handler, and national rail network. It doesn't make much sense building two parallel networks for these, which is why I'm also happy that the NBN is being done by a government. The second problem is harder: it's where being 'efficient' would ignore externalities and result in bad outcomes. People in remote areas are glad that the post system still covers them, just like how people with health problems are glad that they can get government health coverage - neither of these are 'efficient', but they seem good to have in order to avoid an external social cost.

The downside here is, when allowing inefficiency, it's hard to justify what is actually allowed, and you run a risk of becoming too inefficient. Governments certainly fall down here, due to having so much cruft build-up over time ('technical debt', in tech terms). That said, I don't find them that much less skilled at fixing it than businesses, but perhaps better measures can be put in place to ensure it doesn't build too much.

Conclusion
MNC aren't breaking the law, but they're also not acting in the interest of their customers in the countries whose goverments they are avoiding tax from (and by proxy, their own interest). A great example of this is Australia's NBN: It is in the interest of the users in Australia to have fast connections, and also the tech MNC so they can get more usage. I feel then it's crazy to both be disappointed by the cost/speed of the NBN (delays, fttn vs. ftth etc.) but also minimise paying tax that would go towards paying for a better NBN. Either these MNC can build their own (which seems even less efficient, and I'm sure wouldn't have as high coverage), or cut the tax minimization strategies. Even better would be for one or two to do so pro-actively "Hey, we acknowledge we found a bug in the tax law, and you're trying to fix it, so we'll not exploit it while helping you". Unfortunately then it's in everyone's interest to let other companies do so, so that is unlikely to happen soon...

Thanks for making it this far! I hope this has been a more interesting read that the usual one-sided analysis, and please leave feedback. I'll leave you with David Mitchell's thoughts on the subject - which takes a different line but is similar in some ways, and presented in the excellent David Mitchell manner (warning: also with his language.)

A few notes:
- I have no legal or corporate finance background, so there are likely errors. Let me know and I'll fix them!
- Related: This post is no yet finished. As more points are brought up, I'll try and work them into the analysis above as seems fit, and maybe modify a few conclusions too.
Enjoy!

Oh, and to the late Kerry Packer: I don't minimize my tax. I'm in a fortunate position to have to pay tax, and be able to live on what I do without trying to minimize it. While it is in my rational self-interest to personally minimize tax, and it would have no measurable impact on what the government has to spend, I see the problems in applying personal logic to country-level situations (which taxation is for). I feel this is the equivalent of defecting in the Prisoner's dilemma. Yearly tax is iterative, so I guess you're implying the government should play 'tit-for-tat' - i.e. fine you heavily for minimization?

A long long time ago....

It came to my attention not all that long ago that, 10 years ago today, the first ever padsterprogramming blog post was made! It's really weird to think I've been recording thoughts in here for the last ten years, and all that has happened in that time (Blogger looks different, for one, as does my blog template...).

I read through all the posts last night in order to catch up with what I was doing over the last decade; apparently there was a time before I was doing TopCoder designs and ACM contests - the first post is from first year uni, so I was still into coding, but with a lot ahead of me yet...

Amusingly, there was quite a bit similar - apparently I was watching survivor (and it was already season #10: Palau. it's still going, now up to #30). I was planning on getting a 'proper watch' - which I recently just did again, upon leaving Zurich (Aerowatch this time, not Rado though). I was starting at Uni (oooh, so many posts about uni...), which will happen again in Canada at the end of this year. I was playing indoor soccer, which I've kept up (on and off) since, and while visiting am now even playing with a spiritual successor of the same team! And I was doing coding projects (then it was Blog Watcher, now it is Line Tracker...I really haven't improved my naming scheme).

Lots has changed though - for one, my posts converted from short-text personal updates (mostly about coding) to more thought out, longer posts with pictures; probably related to the former going to the more transient social networks instead (myspace! plus livejournal, facebook, and now G+). I've done quite a lot more travelling too - in that decade, I lived in five places in two countries, plus visited seven continents and blogged about nineteen visits. I made it to 5 world programming finals, had an internship at SAAB then started working at Google, including launching Wave and working on this very product (Blogger, especially comments). And of course, there was the time Federbear first appeared (although lesson learnt: don't use images hosted elsewhere - mostly only my own picasa ones are still around).

On the more amusing side of things, this blog did at one point get killed. I don't recall how I ended up so emo at the time to do so, but I do remember feeling sheepish with every post afterwards... It was weird to read about a time before the iPod (and hence iPhone and iPad), back when I still had a USB mp3 player and thought that everything should be XML based.

I've also already done some retrospective posts (five years apart, before, and after), so know enough to realize any that predictions now of where I'll be in another ten years will be far off. Maybe starting uni again while watching survivor and playing indoor soccer? Now there's a thought...

Finally, I leave with the picture I'm most proud of in this blog over the decade. Plus, I never did find out from 2006: "When Nappy companies claim that their new products are more comfortable than their last...how do they know?? Are there professional nappy testers?". Cya in 10!

The treachery of rubbish bins 

Raspberry cheesecake in a mug

I recently found out it's possible to make a cheesecake in ~10 minutes with only four things to clean and not much to throw out - a pretty good combination!

This is mostly a note for my future self to remember how to do this, but it might be of interest to anyone who (a) likes cheesecake, and (b) likes quick and simple cooking - which is most people, right? I had bookmarked a recipe a few years ago for microwave mug cheesecake, but I didn't have a microwave at the time (whoops...) - having moved recently I now have access to said equipment so decided to give it a go for a mother's-day dessert.

Step 0: Prepare

Above is all you need, really! Full list:

• 1 egg white
• 125g cream cheese (exactly half the tub above)
• ~60ml sour cream (not sure exactly, I used ~a third of the tub, about three tablespoons)
• 2 teaspoons of sugar
• 5 teaspoons of raspberry pulp/jam (either pre-pulped, or mash it yourself).  
• 1 biscuit (Butternut snap is perfect if you can get it)

There's lots of room for adding stuff, so I'll put optional ingredients in the text, but the above is fine. For the non-edible stuff you'll use:

• teaspoon, fork and bowl (for mixing)
• microwave-safe mug (for...holding).

Step 1: Mix

Separate the egg white into the bowl (or using a plastic bottle apparently!...but that's too messy), add the cream cheese, sour cream, sugar and 3 teaspoons of raspberry. You can optionally add a few more flavours here, e.g. I added a few drops of vanilla essence, and stuff like lemon juice or spices may work too, whatever should be mixed through the whole cake.

Whisk it manually with the fork until it's all mixed and there aren't big lumps (as seen above) - this took me a bit over 5 minutes. You can use an actual manual or electronic whisk, but again, that's more to clean, and a fork is fine if you don't care about it being perfectly smooth... You can  also taste it at this point, to see if the sugar level is right, and modify it based on how sweet your tooth is.

Step 2: Microwave

Pour the mixture into the mug. Hopefully it fits :) If not, you need bigger mugs... The actual cooking is in two parts, stirring once in between to make sure the heat is distributed. With my microwave at 750w, I did it for 1 minute first (left photo for proof), then stirred the gloop around, then put it in for a final 30s. The cake itself should look like the photo on the right after both the first and second zap - wobbly like jelly but a bit more solid and textured. Essentially: if it's still liquid, do it longer, if it's completely solid (baked-cheesecake style), bad luck, that's too much. I'll let you extrapolate what to do for different wattage or number of mugs, just use cooking/physics/maths to find the correct scaling factor... 

Step 3: Wait

That's it - you made a cheesecake in 10 minutes! Don't eat it now though - it's still warm. It needs to cool, and set into the right wobbly, mousse-y texture, before it'll taste right. As I found in the first failed attempt: warm cheesecake is pretty horrid. The second attempt worked after ~30 mins in a freezer if you need it fast, but an hour or so in a fridge should also do, and the one in the photos was overnight.

Step 4: Present

I'm repeating the original picture here just to explain what happens at the end - this is where the biscuit comes in. Crush it (on the benchtop, inside a sandwich bag is recommend to avoid spillage) and the put it on the top; this replicates the normal cheesecake base, and should be crunchy and sweet(ish), plus buttery if possible. Add the remaining raspberry, and at this point I added an optional bit of grated chocolate (for mother's day flair - I'd skip this normally).

That's the whole thing - enjoy! All that needs cleaning now is a bowl, a fork, a spoon, and your mug. Plus, there's not much food waste - I got rid of the egg yolk, but everything else went to other meals (more cheesecakes, mostly...). I definitely plan to keep trying these out if my waistline is ok with that - e.g. replace raspberry with coffee/chocolate, and making it in a transparent mug for a cappuccino cheesecake should be awesome.

Kindling #3

There's a bit of a short lull while I work out how to write the next few posts and what exactly they should be about, but I've been doing a lot of reading recently due to travel/leaving work; as I was about to clear some books off my Kindle I figured it worth reviving an earlier trend of mine and captured notes for what I made it through recently for those interested now or in the future. So, to summarize my thoughts for the novels just removed:

Dark Places (Gillian Flynn) - Having seen Gone Girl and immediately thinking it'd be awesome if it were a Palahniuk book focussed on Amy, it seemed a good idea to read another from the same author. Dark Places is Flynn's second novel (before Gone Girl) and already is being turned into a movie, which surprised me a bit as it's a lot less marketable than first movie. The characters are certainly strange enough to be from a Palahniuk novel (extreme family history, unusual behaviour in normal situations, add some kleptomania and money issues) but the underlying plot seems fairly...sane. Not to say it's bad, and probably much more approachable for those not used to the Damned/Doomed-style absurdity, but not quite what I was after. Hey, at least I get emails from Amazon titled 'Sharp Objects' :)

Endgame: The Calling (James Frey) - This is the first in a trilogy of YA fiction around a teenage battle royal; nothing new there, after hunger games, divergent etc this genre is really being done to death, but in the case of Endgame, it includes a real-world puzzle solving and cash prize component, along the lines of Ready Player One (which I enjoyed). Endgame is certainly airport fiction - not for invoking any deep contemplation of challenging ideas, but still easily consumable. The writing style I found pretty weird (third person present, Patrick says) and the characters are too well-trained and super-human-y plus the Endgame rules too unclear to make much sense out of why anyone is doing what they are, so perhaps skip it unless you've read all the other good YA or want to compete.

Thief of Time (Terry Pratchett) - RIP Terry! After March 12, I felt I had to read a new Discworld novel, and a time-related one seemed ideal; I'll keep this short, and just say that if you haven't read Discworld, you should! If you have, you'll probably like this, or know enough to figure out yourself if you'll want to read it.

The Gemini Effect (Chuck Grossart) - Amazon recommended this one, as it was a finalist in their 2014 Breakthrough Novel award. Usually the recommendations are pretty good, but this seems to have really missed the mark. The premise sounds like a Crichton novel I'd enjoy: some old army biological weapon gets into the public, infects animals, and humans try to fight back to contain it. Except every character is some genius-attractive-caring-gritty-scientist-general-leader who is in love with / attracted to / falling for all the other genius-attractive-caring-gritty-scientist-general-leaders around them. The clichés galore removes all the actual interest in how these people will respond to the situation and makes much of it fairly predictable. Unfortunately, I was reading mostly to figure out how they'd resolve everything (I couldn't think of how to do it, and my internal bet was some Deus-ex-machina like Signs or War of the Worlds) but no, not really any satisfying ending. So I'd suggest avoid this, unless you're looking for how to write stereotypes.

The Buried Giant (Kazuo Ishiguro) - another recommendation, probably after reading previous Murakami or Fantasy novels. Certainly a pleasant book to read, but maybe more in the style of watching some arty short movie where I feel I should be enjoying the craft moreso than the content. The idea of memories being removed, and how much you'd want to get them back, is really interesting, but the slow pacing means a lot of the characters and world can't be explored - perhaps I'm supposed to fill in those parts myself? The fantasy setting itself was unusual too - I couldn't tell if it's capitalising on a middle-ages fiction revival, but that didn't seem to add much, it could have been set anywhere with low population density (like...Iceland is also popular? or space?) and have worked; but as with Endgame, I found I didn't learn enough about the rules of the land or time to know why the characters were deciding to do what they were. Still, enjoyable nevertheless, and it was nice to change pace from the stuff like Endgame or Gemini Effect I was reading while travelling.

Revival, and Mr. Mercedes (Stephen King) - Two of the more recent releases by King, I think the former is a lot closer to the style of writing that I read him for (other than Dark Tower): a New England setting that is always clearly on the dark side of normal, you know bad stuff will happen (or is happening, off-page somewhere) but the pacing keeps leading you on until everything is revealed much later. Mercedes is a bit like that too, but more on the crime-dark rather than fantasy/mystery-dark. It's still quite good as a crime novel, and the characters end up rather interesting and likeable, although for I prefer my bad guys a bit more...psychologically adept (Hannibal-style). Still, it looks like there's a new one coming out with the same protagonist and two support characters, so that's on my list.

Currently reading: 

In Cold Blood (Truman Capote) - on my list of good novels to read for a while, it's still pretty early but I'm appreciating his story telling style. Still in the character-development phase but there's been enough plot to be interested in what happens, but also random back story to explain it when it does - plus, generous comma usage, which matches how the voice in my head reads things as you can probably tell from my blog posts too, and even this sentence...

I hope that helps someone! Note: These are just the ones I was clearing off the device, there's a big gap back to the last post which included some interesting stories too: The City & The City, The Circle (lol!), A Song of Ice and Fire #3-#5, The Luminaries, Amnesia, ... If anyone wants to know about these, let me know too, but otherwise, I should probably just do this more often.

RtW Cities: Paris

This is probably my final one of these for a while due to leaving Zurich and no longer being employed, but after 2.5 years of living only a few hours away, I finally made it to Paris! It's one of the most popular tourist cities in the world, as quickly made clear by the lines of non-french folks everywhere. Having only been once before, when I was about the age of two, it was definitely high on my list of places to visit - and so as I was leaving, I felt now was my last opportunity.

First up - Paris is huge, over 12 million in the Paris metropolitan area, and it certainly feels like it. I came in by train, so managed to avoid the much maligned CdG airport, but still, getting from Gare du Nord station to my hotel (about 10 mins walk) in the dark and rain was a serious introduction to Paris traffic. For most of the trip I'd resigned myself to doing completely touristy stuff rather than trying to get a better feel for the city - partially for language constraints, but mostly for time ones - though the various pâtisseries I visited for breakfast seemed a bit more normal, and lived up nicely to the french cuisine rumours.

As with many of these European cities, in particular the bigger ones, the density of interesting stuff is very impressive, in particular around and just south of the Seine, which is where I spent most of my time. For example, the Eiffel tower (of course!), Arc de Triomphe, Musée d'Orsay, Sacré-Cœur Basilica, ... so many attractions, and most seemed like things I'd be happy to visit multiple times or if I were living there, e.g. the park out the back of Notre Dame was an awesome place to kill some time reading (once the weather had improved). Top of the list though was the Louvre - I'm not a huge fan of Early-modern-France paintings, and have never been amazed by the Mona Lisa, but you can easily spend a day in that place - for me it was mostly the statues plus roman/egyptian areas. Plus, for anyone a fan of Monet, make sure to visit the Musée de l'Orangerie - a small place, easy to miss and it doesn't have much, but I really like the idea of dedicated galleries for certain works, and in this case, the Water Lilies there are amazing (for those that can't make it, check street view)! And then there's Versailles - where most castles are proof of extravagence beyond utility over normal houses, Versailles was proof of extravagence beyond utility over other castles...crazy.

As mentioned, getting a feel for Paris as a city was tricky - it's just too big, and I didn't get to do the things you'd normally do while living there. The bus/metro/train system seemed well equipped to cater for the size (albeit crowded, confusing, and a bit dirty, but I have the high bar there of Zurich). I enjoyed the residential/commercial mixing (i.e. everyday stores and restaurants were mixed well, which makes living easier) although I feel like I missed out on the famed french dining, most places looked more like what I'd consider italian restaurants, and served stuff like lasagne or burgers... It seemed that apartment density was a bit high centrally, I guess population and demand is high so those who don't want to spend heaps on housing end up crammed into small area; there must have been about 100 people standing outside the hairdressers in the area near me each day, waiting for partners to get haircuts which also suggested high people density. Although some things were skewed because it was easter long weekend, and from the sheer tourist numbers. Smoking is quite big there (it seemed more than zurich or berlin) and I did get asked for spare change quite often - though usually only in French. Perhaps I looked local? As I also got asked for directions three times in the three days I was there, and there was that 2011 french eurovision guy... :) But probably just the higher number of tourists (I got asked three times in five days in London too, the week after).

Photos available in this album. As mentioned at the start, this is probably the last of these for a while as I'll be moving to Canada and going back to being a student; I have a few more blog posts planned though, some for pysch/neuroscience, some for other musings, and a few review-style. Stay tuned!

Changing vessels

Some of you may have heard it already, but I haven't really broadcast it yet: 31st March was my last day working at Google, and probably my last day as a professional software engineer! I wanted to hold off announcing it until after April 1st, to avoid any suspicion of an elaborate hoax, but now it's the 4th, I think it's safe: officially, I'm now unemployed. Some more details to answer what people have been asking me most often:

Why? I like to think that it was inspired by this wonderful SMBC comic. While not following exactly, I had been doing coding for around 10 years (4 at university, 6 at Google), so it was time for a change. Additionally, it's a good time to change - I was finishing a project at work, I had to move out of my apartment anyway, I have no dependents (partner/kids) and working at Google has made me financially stable, at least for the next SMBC 'lifetime', 

There were a few other (relatively minor) things at work which made the decision easier although I can't go into much detail. They were certainly not the deciding factor, but having started to do many non-CS MOOCs, and finding this (and personal projects) much more interesting than work coding, I decided it was time to switch full-time. It's possible I'll give up after a year or so and coming crawling back to free food and stable employment, but hopefully this next lifetime will be as long and enjoyable as most of the last!

What next? My plan for life #2 is more along the looking for facts/truth part of the comic: I'm hoping to switch into Psychology/Neuroscience - or more particularly, an undergraduate degree in Cognitive Systems at UBC in Vancouver (if they accept me...). This has the best mix of Psych/Neuro I could find in a city that seems really nice to live in, gives me a little while to pick a specialty (both the philosophy and psych tracks look good). This also means about 5 months of 'holiday' until it starts - some of that is actual holiday (like now - in Paris then London! Probably NZ at some point too), definitely expect more blog updates and coding on my github, plus catching up with folks in Adelaide / Sydney, and a few attempts at more creative writing (stories/music/...).

Assuming all goes well - let me know if you're visiting Vancouver!

RtW Cities: Vienna

For Christmas this year I decided to continue my tradition of heading towards the German Weihnachtsmärkte, or Christmas markets. Two years ago it was Stuttgart, last year Berlin (sorry, I forgot a RtW post for it!) and this year I decided to make a bit more of a trip out of it: first a train to Stuttgart again for two days to catch up with a friend, then train to Munich for a few, then finally continuing the rail journey to end at my main destination of Vienna for five days, including the 25th. Also: one warning, being from Australia, I apologise if at any point in this post I write that instead of Austria - it's hard to retrain my fingers, but I mean .at, not .au.

Vienna had been high on my to-visit list for a long time, mostly because (a) it's an old European city with a rich (and musical!) history, and (b) it consistently ranks as one of the top cities to live in, so I had high hopes. Note that 'liveable' doesn't necessarily mean good to visit - Zurich also ranks highly, but is low on attractions outside of shopping and locality to other things. Vienna has the advantage of a much larger population however (~1.7 million, about the same as Perth, Hamburg or Montreal), and the cultural centers it has accumulated left me with a filled week and wishing I had made it a longer stay.

First up: I wouldn't recommend starting any visit there on the 24th. I made this mistake - I was able to make it to one Xmas market (at the Rathaus) but most other things were closed, including Vienna's main shopping street (Mariahilfer Straße) - I guess similar to Germany, the 24th (in particular night) is a time for celebrating with family, but I managed to find a Turkish place in the end :)

The city was much more alive on Christmas day however - the markets were full, this time Maria-Theresien-Platz which included a short history lesson on who Maria Theresa was; I had never heard of her, but she seemed quite good at leading the Habsburgs, improving education, progressive civil rights, while also having 16 children, 11 of whom were called Maria. For me, the 25th was also topped off by an excellent lunch (Aperol spritz, and my family's traditional smoked salmon, at Huth Stadtgasthaus) and rounded out with a quartet playing Mozart and Haydn at Mozart house. Similar to when watching the Vivaldi concert in Venice, I'm not sure I'm as much of a fan of these really small close-in performances, but it's still enjoyable (and comforting to see/hear that even professionals make mistakes), and something I felt I had to do while visiting. Sadly, no snow though, it came close, but didn't progress past a drizzle of rain.

The 26th was mostly visiting the amazing Schönbrunn palace - as can be seen at the start of this post, the weather was perfect for wandering the palace gardens, plus touring inside and finally, enjoying the markets out the front (yes, some places still had xmas markets open after the 25th!). What also surprised me was how recently it had still been used - e.g. Franz Joseph died while living there in 1916, and it was used as offices post-WWII (meaning that the fancy older-style rooms inside were occasionally connected to a modern bathroom en-suite...). As it was boxing day, I also continued the old Australian Lord-of-the-Rings release day tradition by going to see Hobbit #3, although accidentally in German (the dubbing/languages are done differently than in Zurich, I realize now...if it helps anyone in the future, OmU = original language with subtitles).

Then, on the 27th, I got quite a bit of proper snow! So spent some time outside just enjoying it, but after getting cold, ended up visiting the impressive Museum quarter - in particular Mumok, the modern art museum. The main exhibition ("Hippies use side door") was a bit hit and miss for me, but the stuffed animals were used well, and I can now say that in real life I've seen both a Mondrian and in particular, my favourite painter Magritte as they had La Voix du Sang). Additionally, as a psychology fan, I made sure to check out the Freud museum (pictured above). It's a bit unusual as it's just a few small rooms where he used to live and work, but it can be interesting to imagine while there at what it would have been like when in use, and see works like The Interpreation of Dreams in their natural environment. Plus, all the non-psych political goings on at the time - the late 1930s was not a good time to be Jewish in Austria, and he eventually exiled to London.

Overall, Vienna definitely was a great place to be a tourist, and I imagine living there would be quite enjoyable too. Things like politics/health care/social security seem reasonably well run, public transport is great - trains, busses, trams (with heated seats! great for snowy days...) and a good subway system suggest it will survive future growth too. It also seemed to have a suitable mix of old and new stuff - while there's a large number of older impressive buildings (the national library is amazing!), it didn't seem to present much of an impediment to modern expansion - e.g. the Donau area definitely feels much more 20th/21st century.

For all the photos, see this album, and I leave you with Federbaer's most recent acquisition - he decided to join in on my guitar playing, but one-upped me by getting an Austrian instrument.

To other updates that I didn't want to put in their own blog post. First is that Katie and I broke up - check various social networks for more details. It was a friendly split though, so we're both doing ok, but that explains why she's not in this (or future) posts. The other much smaller one is that I turned off Google plus comments and went back to blogger ones - mostly to remove the need to have a google account to add comments, plus this model is a bit easier to understand (comments here don't also appear as top-level posts in G+). If you have any problems posting comments in the new system, let me know (as I wrote much of the code :p).

RtW Cities: Rome (+ Florence)

Travelling over the winter mostly slowed down due to skiing instead, so this is the first round-the-world cities post in a while. Other travels in that time included a trip home to Australia (Sydney, Melbourne and Adelaide - maybe I should add them to the series?) and London (already written up), but otherwise, just sticking within Switzerland. Anyway, to Rome (via Florence).

Florence had been recommended by family as a good place to visit, and as it lies between Rome and Switzerland, it seemed a worthwhile place to add a stopover on our trip. Arriving pretty late on a Thursday night, we only got sneak peaks of all the fancy buildings on the way to the hotel; which turned out to be interesting, it appeared to have only just opened (i.e. still under construction, although not near the rooms) and the walk up the stairs to our room past a giant old fresco was a good reminder we were in Italy. After a good half-hour introduction from the extremely friendly Cesare, and a quick peek out the rooftop across the town, we made it through the streets to the Ponte Vecchio, which reminded me quite a bit of the Rialto bridge in Venice - equally shop-y and randomly built over many many years to the point where I'm surprised it's still standing.

That wasn't the only similarity though - Florence was a much more tourist-packed town than I was expecting. The city itself isn't all that big, and has a huge density of attractions in the middle, around which tourism seemed to be almost the only thing (I wouldn't want to live near there...). Unfortunately we couldn't fit it all in a day, but the Cathedral (picture at the top) was amazing - such happier looking compared to e.g. Koln's, and the view from the top of the tower lends itself for great photos (but be warned, not the easiest hike!). The Uffizi gallery was also a good stop, near the Piazza della Signoria (with all the statues, including that on the right), however I've realized that Italian Renaissance paints aren't really my thing - scultures maybe, and architecture yes, but endless Madonna-and-child, not so much...

Florence was followed by the main stop for the trip - Rome. The visit got off to an awesome start, a fancy dinner at Lo Stil Novo, including some wine that even Katie liked (not sure the last time I ate at a place with a sommelier...) but the best cheesecake I have ever had. I kid you not - if you're in Rome and you like cheesecake, you must try it. And if you don't, it'll change your mind :)

Anyway, we had all of Saturday to explore Rome on foot - starting with the Altar of the Fatherland which is a truly impressive sight. Next up was the Roman Forum and Colosseum; as someone who did Latin at school, I particularly enjoyed these parts, remembering the history behind them. It's quite a sobering thought, especially for an Australian (with European settlement in the 18th century), that many of the buildings in the area were from two thousand years ago. The Colosseum, built in 10 years and used for numerous spectacles, is mostly still standing and even has ancient 'graffiti' in parts; I can't imagine what it'd be like to write on e.g. Adelaide oval seats and have people read it in two millennia.

Last, but not least (lest I be struck down by lightning...), on the Sunday morning before leaving we got to cross an international border and check out the Vatican. I regret not organizing going inside, as I hear it's quite spectacular, but even just the novelty of the smallest country (2.3 Popes per square km, and a useful explanation video), the Swiss guard, and the huge crowd out the front was worth visiting. This was followed by the very pretty trans-alp rail journey back (with no missed connections, despite the warnings!) and I'll be eating bread with olive oil & vinegar for a while yet!

More complete photo album here.

MOOC^3

Hey all - in case you're curious at what's been occupying my time recently: www.coursera.org
In particular, while in Zurich I've been filling a bunch of my spare evenings and weekends with some online courses. While I spend most of my day doing computer-science-y stuff (or, I guess 'software engineering' to be more exact), I feel the amount of new stuff that I'm learning now is pretty limited and there's only so much coding I can do before wanting to learn something else... so while my first was a programming course (Scala), the remainder have been a mixture of economics (Economics for Scientists, Competitive Strategy, Advanced Game Theory), med (Computational Neuroscience, Diabetes: Diagnosis, Treatment and Opportunities) and psychology (Beginner's Guide to Irrational Behaviour, Social Psychology, Moralities of Everyday Life).

In particular though, three of those courses mentioned above (Comp Neuro, Game Theory, Moralities) were all offered at the same time at the start of year, when Katie was away and my other commitments (i.e. travel) were low. My plan was to sign up for all of them, and when the workload increased, drop the least interesting one (or two, given I'd only done one at a time beforehand). To cut a long story short, they turned out to be all worth completing, well presented and interesting, and the relatively low workload - a few hours per week for each (thank you, variable speed video playback!) and game theory only lasting four weeks - meant that I finished each. I'd say the Moralities one was the most interesting (probably due to being furthest from comp. sci) and the best taught course I've had so far, but also probably the least challenging out of the three, which I guess suited me as I'm more familiar with the pure maths of game theory, or the coding/modelling for the neuroscience course. Hilights from each?

Game Theory

The Revenue Equivalence theorem - roughly speaking, when distributing goods to rational, risk-neutral people, then given two distribution mechanisms (e.g. different types of auctions), if they both distribute goods in the same way given the same inputs (e.g. bids), and don't distribute anything to someone who values the goods as zero, then the expected revenue of both systems is equivalent. This is extremely general, but means that e.g. any auction of one item that gives the item to the highest bidder will have the same expected revenue as a second-price auction (i.e. the winner pays the amount of the second highest bid). Which is why most bidding systems (e.g. auctions like ebay) are essentially this. In a close second place, the Vickrey-Clarke-Groves auction is a nice auction that both maximises global utility ('socially optimal'), but also makes it optimal for people to value things at their true personal utility (and has a nice proof).

Computational Neuroscience

The most 'woah' style moment came right at the start - one of the early pracs included data from the firing rates of cricket neurons which were attached to leg hairs which sensed wind direction. It had been discovered that these four neurons fired related to the angle of the wind relative to 45/135/225/315 degrees respectively, and so by applying some algorithms, you could calculate the direction the wind was blowing on the cricket in the lab at the time. Which is pretty trippy once you think about it! Other neat things included examples of how simple neuron models turn out to implement quite complex algorithms (calculus, stats, even principal component analysis / k-means clustering), and a somewhat freaky guest lecture where scientists were able to train a monkey to move their hand by sending neural signals through an external computer, while the normal link (brain -> spine -> arm muscle nerves) had been cut. Future implications for this are both impressive (e.g. fixing people with spinal damage) but scary (digitally 'enhanced' neural processing).

Moralities of Everyday life

Not really any wow! moments here, but still interesting - I think the experiment that I remembered the most, other than ones I knew already, was the following progression. First, you start with the Ulimatum Game: given $X (in the order of $10), you get to offer some to someone in another room, who can't see you and can't talk back. If they accept your offer, they get it, and you keep the rest. If they reject it, you both get nothing. How much do you offer? (average is about 50%). Then you progress to the Dictator Game: same as above, but they can't reject it - you get to keep everything not offered. How much do you offer in this case, where there's no threat of retaliation by the other party? (average is about 20-30%). Finally, it progressed to a variant of the dictator game: before playing, you're given the opportunity to walk away with 90% of the money - i.e. in the $10 example, you can either take $9, or play the dictator game (and get $0 to $10). Without external factors, it is irrational to take the $9 - either you don't care about the other person, in which case you are better of playing the dictator game and keeping all $10, or you do care about them, so you should play and at least offer them $1. That said, many people (me included!) would take the $9 option, and the theory is that it's worth the difference just to not have to decide in the dictator game.

That's all for this update - hopefully travel starting again soon (London and Australia booked!). In the meantime, I have started another course (Ethics, by the Australian Peter Singer), and am otherwise now keeping myself busy with this awesome game (2048, pictured above) plus finally getting Gran Turismo 6 as a reward for finishing the three courses - custom music, mount panorama plus the return of the red bull concept cars means it'll occupy my spare time for a while!

Using Psychology

Ψ

Just this week, I finished my fifth Coursera MOOC - Social Pyschology, from Wesleyan University. Somewhat interesting, I didn't really learn as much as the behavioural psych and it was a lot of watching videos from experiments, although that still had some hilights like original shots from the Milgram Obedience and Stanford Prison Experiments. Plus, going back to writing assignments took a bit of adjusting to :)

One thing is for sure though, I think Social Psych (or at least, the people running this course and discussing the most on the forums) seemed too closed-mindedly left-wing for my liking. While that might not be a bad thing on average, it's a bit unfortunate in a hopefully scientific, educational environment if logical reasoning is dropped and people just follow oversimplified arguments.

Anyway, the course is over, enough of the mini-rant. The actual interesting thing that seemed worth posting about is something I noticed when doing part of the required readings. In particular, one of the weeks (covering 'Conflict, Peacemaking, and Intervention' - amusingly not the week on 'Persuasion') included reading The Psychology of Climate Change Communication - 'A Guide for Scientists, Journalists, Educators, Political Aides, and the Interested Public'. An interesting how-to guide for anyone involved in marketing, it's mostly a collection of social-psychology-ish findings and how to apply them to climate change communication to best get the message across; e.g. don't call it 'global warming' because that suggests that every part of the world is warming, and lessens the link to natural climate.

I'm not sure how to attribute this image to the original source,
as none of its current users seem to either.

Reading this made me a bit uneasy - I don't know whether there are clear guidelines of how much psychology should be used when trying to 'convince' someone of something, or for scientific things, how much 'convincing' should be done. Certainly, it's better than not knowing it's going on - but is it a perfectly fine thing to do? The closest equivalent I could come up with is political lobbying, which on average seems to be quite disliked - and recently, increasingly so. The idea there is that people with access to money can buy influence, which is bad: their 'voice' is louder than those without money, which seems to weaken a democracy in an arbitrary way. Turning this around, the conclusion here seems to be that better access to social psychology findings can (/should?) also be used to increase your influence, which seems an even more arbitrary way to bias a democracy. Plus, just as with money, they theory is that if you don't do it, your 'opponents' will, so it's kind of a prisoner's dilemma race to the bottom.

For those Australian's reading, this came up a little while ago in the news there too - our former PM, Julia Gillard, was accused of 'leveraging gender wars' (paywall). While there was a bit extra history which complicates this, part of the anger seemed to be that she was pursuing a 'secret strategy' which seems to be just another way of saying 'figure out how to get the message across most efficiently to a larger number of people'. It's for things like this where I think the lobbying comparison still holds, but is a reminder that not all lobbying is bad too. If there was one finding to take away from the behavioural psych course (and one of the economics ones), it was that humans aren't perfectly rational - so even with the best argument supporting a cause that really would benefit people, you may still have to tweak it so that it will actually get through to people.

So in summary...like lobbying, 'applied psychology' can be good and can be bad. Unlike lobbying, however, there's no real discussion of it, so it continues to happen without any oversight. It would be interesting to see however what the reaction would be from the general public if it does become more of an issue, and my guess is that depends highly on whose application of it gets derided first.

The value of a dollar

Image from the Australian ABC

After my North Korea trip, it was pretty easy to realise that there are some problems with the current form of capitalism, and it's worth trying to identify and borrow good parts of non-capitalist systems; unfortunately, I'm not really experienced in that area (by that I mean, I know effectively nothing), so trying to improve that by reading a few recommendations on the Marxist/Leninist side of things - Das Kapital and The State and Revolution (I know, Amazon links, oh the irony).

In any case, I'm only part way through the first and making slow progress, but all the talk of commodities and the gold standard made me think about how we define currencies today. You see, way back when, coins were used to represent actual metal value - i.e. the british pound used to be the Anglo Saxon pound which was one pound weight of silver in value. This was nice in that it fixes the price of silver (i.e. a pound buys you a pound) but means that the cost of everything else varies as the metal price changes - this is essentially the idea behind the gold standard (and why I chose the picture up the top).

Stock photo interjection

The problem is, noone uses the gold standard any more. Instead, we have fiat money - essentially, it is just used for transferring value, and has no relation to anything physical. Yet there's still this concept of $1 (or whatever currency you like) which is defined over time, but seems to bear no resemblance to anything - you cannot define it by anything really that is not circular, or time dependant (e.g. $1 isn't X grams of gold, except at a given time), so really the only thing I can come up with is: "$1 used to be a certain amount of gold. Now, it's value is very close to what it was yesterday, but probably slightly less".

So this led me to: Is there something else that fulfills this property (of small variations which mean that people don't have to change prices a lot all the time) that would actually be more useful than this completely abstract amount? My first thought was along the lines of what Marx has covered a lot, and that is labour. If I denote the new currency as H, then define 1H as the wage you need (per hour) to get by ok - the closest I could approximate this to would be the poverty line. i.e. if your wage is 1H per hour, then you're right on the 'adequate income' level.

What is nice about this? A few things:

• It (hopefully) fits the criteria of being stable enough on a day-to-day basis to use as currency.
• Comparing the minimum wage to the poverty line is trivial. If the wage falls below 1H/hr, you have a problem.
• Exchange rates are a measure of comparative cost of living (I'm not sure what they are now, maybe this is more useful?)
• Relating the cost of items to work that goes into them is much more salient - i.e. if I buy something that's 8H, that's one day's work for someone on the poverty line - maybe we can go all british and non-metric and have 8H = 1D, 40H = 5D = 1W, 50W = 1Y = 2000H, nicely. So if I buy something online that costs 1000H.cn, that's equivalent to half a year's wage for someone just getting by in china.

Just a thought. I'm sure there are downsides to having 1 currency unit actually correspond to something. After all, the gold standard was dropped - plus, this is something that is non-trivial to calculate. But maybe there are upsides too. It is quite possible that it is also not something that can actually be done, or is theoreticallay possible but in reality not (or vice-versa).