Seville (Naboo?)

When I go to big cities I am always a bit stressed, because there are always more things to do than I can fit into my schedule, and that makes traveling a bit unpleasant. It may seem like a good problem to have, but if I am stressed about having to see everything, doesn’t that defeat the whole point of traveling?

I am at a point in my life where I can really use some down time to reflect on recent happenings, so I intentionally picked a place that’s not quite as busy as other places I’ve been to in the past – Seville!

Not quite as popular or well-known as places like Barcelona or even Madrid, but it’s still a city with magnificent architecture, epic stories of cultural clashes and conquests, religions, and art. A lot of art.

Many famous operas were set in Seville, and they serve as testimonies to its beauty. To name a few among a dozen – Carmen, Marriage of Figaro, Barber of Seville (where the Bugs Bunny theme song came from), and Don Juan.

The city is also an enduring record of the clashes between the Moors (Muslims from North Africa) and Catholics in the 8th century, and as a result the architectural style is a unique blend of Jewish (came with the Moors), Gypsy (also came with the Moors), Muslim, and Catholic. Often in the same buildings as well, and that’s super cool!

Also, circumstances dictate that I travel by myself again, so I’m back to meeting strangers! Like last time, I’m going to number them in this post instead of giving their names – both in order to protect the innocent, and also because I am shit with names.

The hostel is nicely nestled in the middle of Santa Cruz, a labyrinth of narrow alleys that make up the medieval Jewish quarter of Seville.



It all looks very confusing, but for some reason, I can always manage to find my way back to the hostel without a map if I just follow my instinct. Maybe my brain decided to grow some grid cells?

The hostel is a bit anti-social with everyone doing their own things (like me on my laptop typing this post…), but the decorations are interesting, and there’s free dinner!

img_20161110_181607 img_20161110_193718

img_20161110_220904I love hostels with free dinners! Dinners are amazing for bringing people together and getting people to start talking. And talk we did.

Dutch girl #1 used to live in Brussels doing spreadsheets for an insurance company, but decided that’s too boring and went to live in Australia for a few months, and met Australian guy #1, and they started dating. They now travel all over the world as nomads. How cool is that?!

Belgian girl #1 also works in Brussels, and after a few moments of confusion, established that she actually works at the same insurance company that Dutch girl #1 worked at. They did not know each other.

She also speaks perfect English, despite English being her 4th language. Belgian people are amazing.

American girls #1 and #2 were from Chicago and Michigan, and they are on an exchange program, spending a semester at a university in Madrid, while teaching English for 4 euros/hour (that’s not survivable even by Spanish standards). They are doing some traveling before going back to the States, and they both speak Spanish really well! I really need to work on my Spanish. Being only bilingual is getting a bit old, when everyone I meet at hostels are at least trilingual.

Canadian guy #1 is a mechanical engineer from London, Ontario, working for General Dynamics. We mostly just geeked out and talked about mechanical engineering shit that I assume most of my readers aren’t terribly interested in, but he did tell me how a friend of his in the bio-med field mentioned to him that there’s this simple device many research labs pay thousands of $s for that he (and even I) can build in a few days for about $200, and he is going to launch a startup to take advantage of that. I don’t want to say too much about it since I’m not sure if he wants more people to know, but as an engineer, I’ve had lots and lots of people pitch their startup/project ideas to me, and I found most of them BS in some way. Not this one. This one is actually cool.

Norwegian guy #1 is a vet working in a research lab at a university in Madrid doing research in animal nutrition. I don’t remember what he works on exactly, but it’s impressive.

There are wars and power struggles in the history of every city, and while in northern/western parts of Europe the big war is usually WWII, here it happened centuries earlier – between Muslims and Catholics. This results in interesting architecture. For example, La Giralda, one of the most symbolic icons of the city used to be a minaret – a Muslim tower for prayers. Now it’s in the middle of Catedral de Sevilla – the largest Gothic cathedral in the world… which used to be a mosque.


The cathedral itself is huge, but not really impressive otherwise. Looks like someone had a few crates more gold than they knew what to do with, and just poured it all over this place.

img_20161111_145238 img_20161111_145346


The tower is amazing though!

The way up is a very very very long ramp. It’s a ramp instead of steps because the imam had to go up 5 times a day, and the imam was usually a very old guy. With a ramp he could ride a donkey up. DONKEY! UP A TOWER!!


I am not the proud owner of a donkey, so I walked. And walked. And walked. 20 minutes later… bird’s-eye view of Seville!

img_20161111_151527 img_20161111_151353 img_20161111_151252

How amazing would it be if I could jump off with a paraglider or something?

To be continued… (spoiler: I did not jump)


You are walking on a 100-metre long trail, and there is a pebble every metre. When you encounter a pebble, you can choose to pick it up or leave it. If you pick it up, the game is over, and you get to keep the pebble. If you leave it, you cannot go back to it later.

Your goal is to get the largest pebble. What’s your optimal strategy?

This is the famous secretary problem in decision theory.

Monogamy feels a lot like the secretary problem to me – you encounter a more or less fixed number of potential partners over your relationship forming year, and you have to pick one without being able to peek ahead, and you can’t go back to potential partners you have passed over.

Does that make much sense to you? It doesn’t really to me.

Does loving a person really turn off our capability to love another person? It doesn’t for me, and judging by the number of “monogamous” people who cheat, I don’t think it’s just me.

Love is just a more intense form of friendship. We can have multiple best friends and no one has trouble understanding that. Why do we have to be so selfish with love? Why can’t we just love everyone we want to love?

Why do romantic relationships have to be formed and broken with so much deliberate effort? Why can’t they evolve organically just like friendships?

Those are questions I’ve had for many many years (in fact, I believe I wrote a blog post about this back in 2010 or something). Questions that I never found satisfactory answers to, and as a result, I mostly just refrained from dating.

Until recently, that is. I came across the concept of polyamory a few months ago, and have been reading up on it since. Everything just makes sense!

If you are not familiar with polyamory – it’s a relationship model where each person can have multiple partners, but with the knowledge and blessing of all partners involved (this is a huge over-simplification – there at least as many different kinds of poly relationships as there are mono relationships).

Is it for me? I don’t know. It makes perfect sense on paper, but I don’t know how my brain will actually react.

There’s the practical aspect as well – my friend who did give it a try didn’t find it worthwhile, due to the effort required to maintain multiple relationships. As a famous poly saying goes: “Love is infinite, but time and energy aren’t.”

In any case, if I do actually go into poly, it will probably end up being a mono-poly sort of arrangement with me on the mono side… I barely have enough time for one person!

So, what’s the solution to the secretary problem?

As it turned out, there is an elegant mathematical solution – the optimal strategy is to skip over the first n/e pebbles, and then take the first one that is bigger than all you have seen.

n is the number of candidates, and e is the base of natural logarithm (e ~= 2.72).

For example, in the case of 100 pebbles, that means you should skip over the first 100/2.72 = 37 pebbles, and then take the first one that’s bigger than all of the first 37 pebbles.

See the Wikipedia page for proof.


2015 in Review

I have looked in the mirror every morning and asked myself: “If today were the last day of my life, would I want to do what I am about to do today?” And whenever the answer has been “No” for too many days in a row, I know I need to change something. – Steve Jobs

2015 was about answering “yes” to that question unhealthily often. It was a good year.

I quitted my job last year to do an MSc in London, but it was mostly just an excuse to spend a year living in a new country, and ticking a few more things off my list of things to do before I die. That turned out to be a very good decision.

I am finally at a point in life where if you tell me I am going to die today, and ask me what I wish I could have done but never did, I would have to think long and hard about it.


In the first half of the year, I finally got into gliding with the Imperial College Gliding Club. It’s something I have always wanted to do being a powered airplane pilot, and I’m happy to report that it did not disappoint. We went on quite a few trips to the local airfield (Lasham), as well as a week-long trip to Long Mynd over the Christmas holidays, where we flew an unhealthy amount, and got to try out a really cool launch method. See this post for videos and photos.


I didn’t end up getting licensed in gliding because gliding is very time-consuming (whole days at the airfield), and unfortunately time is one of the things I don’t have a lot of… it was still amazing, though!

SCUBA Diving


Thanks to all the fine people in the Imperial College Underwater Club, I was also able to get certified in scuba diving! It was a bit cold, but everyone in the club, especially the instructors, were absolutely amazing! It’s something I definitely plan to keep doing. Somewhere warm?



I got orange! I probably should have gotten it a long time ago, if only I didn’t switch dojo every few months.

What’s more important though, is what I learned. I felt like I really made quite a bit of progress in my understanding of Judo this year. I am much more relaxed now when I go into a fight, and can more accurately use my opponents’ force against themselves, which is really what Judo should be about. My fighting style is now much more reactionary (that sounds like a bad thing, but it really isn’t). Instead of always initiating and trying attacks from stable positions, I am starting to be able to let the opponent attack first, recognize weaknesses in their movements or balances, before moving in to attack. That made Judo a whole lot more fun.

Still have lots and lots to learn. Looking forward to another year of Judo!


Imperial College subsidizes musical tickets through the Art Society, and we were able to get tickets to most West End shows for less than £30.

Living in Vancouver, I am used to only be able to watch musicals when they go on tours, and it’s often necessary to book half a year in advance. Living close to West End was really nice. I could watch pretty much anything I wanted, and only had to book a few days in advance.


It was amazing.


I was in desparate need of an excuse to travel outside of my room during the thesis writing period, so I decided to join a week-long full time course on woodworking and furniture making.

It’s something I have always wanted to learn. I am into DIY and I would say I am pretty good at DIYing electronics stuff (maybe having a degree in that helped?), but I have always sucked at making mechanical stuff, so this was an attempt to fix that.


We each built a coffee table! I was surprised how much we learned in just one week.

Can’t wait to start building more stuff!

Also, sawing wood is therapeutic.


There was supposed to be a graduation trip, then everyone else bailed. So I ended up going backpacking on my own. It ended up being a whole lot more fun than I ever imagined!

I took about a month to visit Lisbon, Prague, Vienna, and Budapest, staying in hostels and meeting new people every day. I was a bit apprehensive at first, and wasn’t sure if travelling by myself would be a good idea, but it turned out to be probably my best decision of the year!

I wouldn’t go as far as calling it self-discovery, but I did discover a very fun way to travel, and met tons of really cool people and heard tons of really cool stories.

My travel log, if anyone is interested:

Solo Eurotrip, Prelude

Solo Eurotrip, Lisbon Part 1

Solo Eurotrip, Lisbon Part 2

Solo Eurotrip, Prague Part 1

Solo Eurotrip, Prague Part 2

Solo Eurotrip, Vienna Part 1

Solo Eurotrip, Vienna Part 2

Solo Eurotrip, Budapest

Degree, Piece of Paper, and Being (Briefly) Famous

Oh yeah, and I worked on an MSc degree on the side, and got another piece of paper.


I also discovered my love for machine learning after taking the introductory course, and ended up doing a machine learning project for my thesis… and that’s where the fun started.

I posted my thesis on arXiv, and it was miraculously picked up by MIT Technology Review, which started my approximately 3 weeks of fame. It was reported on by quite a few tech and even mainstream media, and I started getting 10-20 emails from random strangers every day about the project, or their new hot startup. That was unexpected.

It was pretty fun in the beginning, because I am passionate about machine learning, and loved talking about the project. But over time, it became pretty consuming, when I realized I was spending 2 hours a day responding to emails.

I always thought ignoring emails is rude and that I would never do it. Then I started doing it, because there’s just no other way. I couldn’t really afford to spend hours replying to emails everyday. Wouldn’t be able to get anything else done! On the bright side, I am much faster at writing emails now. I’m sure it will be a useful skill later.

The fame left just as quickly as it had come. I was back to my normal level of emails within a few weeks. That’s good. No more feeling guilty about ignoring emails!

One good thing did come out of this whole ordeal, though – a job!

Google DeepMind

Having read a few of their papers during my research, I was really impressed by what they were doing. I didn’t apply there, though, because I thought it was way out of my league, being one of the most famous machine learning labs in the world.

One of their research scientists saw my paper on arXiv, and got me in contact with their recruiter!

We then had a series of Google Hangouts interviews with a bunch of people, including a research scientist, a PM, their research engineering lead, an engineering lead, and one of the founders!

It was a pretty gruelling process totalling to about 10 hours of Hangouts, but although they were all heavy weights in the industry, they were also all very nice and down to earth people, and I really enjoyed talking to them.

It was also a bit scary how smart they are. We talked about my project in one of the interviews with a research scientist. He hadn’t read about it before, and as soon as I am done describing it (it wasn’t a trivial project), he was able to immediately offer a few very helpful insights and things to try, and things to think about. Those are things I never thought of, and I spent a few hours a day working on the project for a few months. That was cool.

I also really enjoyed talking to the founder about more high level ideas in artificial intelligence and machine learning. Really cool insights on how they believe ML will play out in the future, about the nature of intelligence, and possible routes to achieve artificial general intelligence. I don’t think I’m allowed to disclose the details, but it was cool. How often do you get to talk to a founder of a company like DeepMind about AI for half an hour? Not very often for me!

I got an offer in the end!


I didn’t originally plan on returning to the UK, but hey, I am used to making last minute decisions on where to live, and Google isn’t half bad :).

Looks like I’ll be flying back to London next week, and also shipping all my stuff over… most of which I just shipped out of London a few months ago… at least someone else is paying for the shipping this time!

I am not going to write about what I plan to do this year, because looking at my past predictions, it seems like I am terrible at making plans. I do hope it will be as exciting as this year, though!

Love of a Woman?

Who’s got time for that?! 🙂

Sidenote: I decided to try OkCupid recently at a dear friend’s suggestion. After all, I am a machine learning guy. Why not trust machine learning to solve this?

Did not work very well. I don’t like the shotgun approach, so after going through tons of profiles of potential matches, I only sent messages to 2 women. No reply. I do understand that women get a lot of messages on OkCupid, but if I have to write 10 thoughtful messages to get 1 reply… who’s got time for that?!

I am not sure why it worked well for my friend but not me. My guess is it’s because in addition to being very intelligent and humorous, she also happens to be highly conventionally attractive. I did not work as hard and was not as gifted :).

I guess that means it’s back to real world dating for me. Hopefully I’ll have more time this year? Apparently a lot of people (especially men) choose online dating because they are too shy to talk to women. I am not. I have no trouble talking to anyone. I just don’t have the time. Arghhh.

In any case, if you are interested in what a Unsuccessful Application for the Affection of Women looks like, here it is.

By the way, the OkCupid blog is very cool. Lot’s of data analysis and making inferences from data.

On Quora

How do you waste your time online?

I used to blog quite a bit, and spend a lot of time on Facebook. As you may have noticed, I am now much less active on Facebook, and my blog posts are fewer and further between.

Of course, that doesn’t mean I don’t waste time online anymore (ha!), but I’ve found quite possibly the most productive way to waste time – on Quora!

If you haven’t already heard, Quora is essentially Yahoo Answers done right. It’s a site where people post questions, and others answer them.

The main difference between Quora and all other Q&A sites is that, Quora, for some reason, seems to attract very high quality content, unlike Yahoo Answers and Reddit (when used for Q&A). It’s either that low quality content don’t get posted, or that their machine learning systems are better than other sites’ at filtering out low quality content.

I spend about an hour on Quora everyday now for my daily doses of random knowledge. I used to read Wikipedia for that, but Quora is a little more social, and the content tend to be more casually-readable.

On Quora, your feed is personalized by “following” either people or topics. Following people means their content will show up in your feed more often (usually answers), and following topics means you get mostly unanswered questions in those topics. If you follow the right people, after a while, more or less everything on your feed will be interesting.

IMHO, the best thing about Quora and what sets it apart is the fact that many active users are experts in what they write answers on. This could mean professors (there are many), seasoned industry veterans, lawyers, doctors, or people who have their own restaurants (making them experts at running restaurants).

If you are getting started on Quora and have similar interests to mine, here are some of the people I follow. Maybe you’ll find some of them interesting, too? By the way, if you want to see their answers instead of all activities, click the “answers” link on the left.

Eva Kor – “Holocaust survivor and forgiveness advocate”. She was a child during the Holocaust, and was subject to medical experiments by Josef Mengele. Lot’s of very touching content on the Holocaust, Nazies, and forgiveness. Why did Eva Kor shake hands with a former Auschwitz guard?

Jimmy Wales – Founder of Wikipedia. He is actually a very active Quora user, and answers many question on the philosophies as well as day to day operations of Wikipedia (and Wikimedia Foundation).

Clayton Anderson – ex-NASA ISS Astronaut. Lots of answers on how ISS works, daily life on ISS, orbital mechanics, etc. Also, cool pictures :). What would happen to astronauts if they got detached from the ISS during EVA? Would they fall back to Earth or drift away into space?

Adriana Heguy – Professor of Pathology and genomics researcher at NYU. Answers on genomics and evolutionary biology, and biology in general. Given that eyes appear to have evolved multiple times independently through evolution, why has human-level intelligence not evolved more than once?

Robert Frost – NASA instructor. He trains astronauts! More space and ISS stuff.

Brian Bi – Competitive programmer and software engineer. And physicist. Lots of answers on C++.

Viola Yee – Generally awesome person :). I have no idea what she does for a living, but she writes a lot of good answers on a lot of different things. Mostly things to do with animals and plants. Is extracting wool harmful for sheep?

Emma Homes – Australian flight instructor. Answers on aviation, parenting, and pregnancies.

Yoshua Bengio – If you do any machine learning, he probably doesn’t need any introduction. He is one of the pioneers in deep learning. Answers on deep learning, big data, life in academia, etc.

Sergey Zubkov – Living and breathing C++ standard. He knows just about everything about C++.

You can, of course, follow me as well, and I’d be incredibly honoured :). Most of my answers are in machine learning (especially deep learning and neural nets), electronics design, CS, and aviation. I also occasionally answer questions on martial arts, chess, viola/violin, scuba diving, and a few other things.

Happy Quora-ing!

University Review: Imperial College London


(Image courtesy of Wikipedia… does it actually snow in London?!)

So far on this blog I have reviewed books (1, 2, 3, 4, 5, 6, 7, 8), musicals (1, 2, 3, 4), movies (1, 2, 3, 4, 5, 6), computer games (1, 2, 3, 4, 5), 3D printers (1), and an airplane (1).

Since it’s pretty clear that my blog is going to be reviews of all things under the sun, and that I don’t want to study for my exams, I am going to try reviewing a university – Imperial College London, where I have been spending the bulk of my time in the past few months.

This post is written from the perspective of a postgrad (MSc Advanced Computing), but because our course (*1) is incredibly small at 12 people, I’m trying to make it non-course-specific.

Campus and Surrounding Area

To be honest, the campus is quite underwhelming.

The campus consists of a few tightly packed mostly-ugly buildings in a rectangular block in the middle of South Kensington, probably the most expensive part of London (London Part II – South Kensington, Imperial College London).

The surrounding consists of museums, art galleries, consulates, and embassies. They probably also contributes to the perceived ugliness of Imperial, but I guess it’s better to be in an ugly building in a beautiful area rather than in the same building in an even uglier area that makes the buildings seem beautiful in comparison.

There is obviously no student parking, but that’s not too big of an issue since London’s full time congestion and good public transit means no one drives anyways.

There is also no subsidized accommodation at all for postgrads, which is more of a problem. We either have to pay an arm and a leg to live in a leaking hole somewhere nearby (among all the diplomats and billionaires), or pay only an arm to live somewhere 40 minutes away by tube. Fortunately my degree is only 1 year. I really feel sorry for people doing their PhD here.

Coming from a campus the size of South Kensington (yay for UBC!), it definitely takes some getting used to.

It’s nice to not have to walk between buildings in the rain, and being 20 minutes from the West End (*2) by tube almost makes up for it, but it’s still nice to have a bit of space.

The lack of space also affects non-academic facilities on campus. For example, food.

There are 3 main food places on campus – JCR, h-bar (restricted to masters and up), and SCR (PhDs and staff). The food is reasonably good and reasonably priced, but selection is very limited, and they are always crowded. We always go to h-bar because JCR food is absolutely terrible. If you are going as an undergrad… be prepared for terrible food (that you have to queue for 20 minutes to get).

There are small cafeterias in a few buildings that all sell the same things (sandwiches mostly).

Academic Facilities

This section is specific to Department of Computing for the most part.

The facilities are amazing. Imperial is a very well funded university, and it really shows in their hardware investments.

We have our very own 13000 cores cluster to play with. How awesome is that?! There is also a 32 core 512 GB RAM system we have SSH access to. It may not sound very important, but they are very convenient if your work requires high computation power. It’s very nice to be able to do a parameter optimization run in a few minutes rather than overnight. Or a large scale Monte Carlo analysis. Or something.

There is also an internal Amazon EC2-type service, for people trying to host servers, etc.

We are also very rarely out of lab machines, unlike at UBC where we have to wait 20 minutes for lab machines at peak times. The lab machines are all very well equipped with minimum 4 cores, crapload of RAM, and some have GTX 780 and Quadro/FirePro for people doing graphics or GPGPU work. All lab machines run Linux, and are all accessible over SSH, and there is a distributed task system that allows us to distribute arbitrary parallel workload to be run on spare cycles in all those machines. That’s about 1000 IvyBridge and Haswell i7 and Xeon cores, across 250 quad core nodes.

There is also a hackspace in electrical engineering building with many 3D printers, a laser cutter, and a bunch of other tools. I haven’t had time to check it out, yet.

No complaints in this department.


Imperial is one of the top universities in the world, so we do get a fair share of famous professors. It’s nice to be taught by leaders in their respective fields, but the downside is that they are all incredibly busy, and don’t really have much time to talk to us.

You can tell that they are really making an effort to talk to students, but they just don’t have time for the most part.

Most of them are pretty good at teaching, though I have only taken 4 courses, so there is high sampling error.

Class sizes aren’t usually too bad, but there are a few very big ones (like the intro to machine learning course).

TAs are generally helpful and know what they are talking about.


There is a saying that if you are not the worst programmer in an office, you are in the wrong office. It’s something I always strive for – to work with people more capable than I am, and I think I’ve found just the right place for that.

This is perhaps the biggest advantage of going to a competitive university – there are no stupid people (sorry :P).

At UBC, if you are randomly put in a group of 6, chances are there will be 2 people who have no interest at all in the subject matter and are only doing the degree because they heard it pays well (or their parents told them to), 1 person really struggling with the course and requires explaining everything 5 times, 1 person who is failing anyways and doesn’t care, and 1 other useful member of the group.

At Imperial, that just doesn’t happen. Everyone is competent enough to get into this university (though even I managed to get in, so maybe the standard is not very high after all), and most importantly, passionate about the subject matter. It’s nice to work with smart people (like our machine learning group, which is pure amazingness), where when you explain ideas to each other, you never really have to say anything twice, and everyone will actually get it. We can all brainstorm at a very fast pace, and don’t need to spend the bulk of the time keeping everyone on the same page.

Imperial has the reputation that it’s full of nerdy people. While that may be true to some extent, Imperial is also known to have very active and well funded student societies.

I am part of 6 different societies, and I find that the distinguishing quality of Imperial people is dedication. Not just dedication to academic work, but also other things in life.

People in the SCUBA diving club are dedicated to diving (they are amazing – volunteer instructors to spend hours with us in the pool, when they have their studies/research to worry about as well); people in the gliding club are dedicated to gliding – the captain wakes up at 6 on both Saturday and Sunday every weekend to take people to the airfield, and spend the whole day there; people in the archery club are dedicated to archery, and go to competitions all the time, in addition to teaching all us newbies.

No one does anything half-heartedly, and it’s mutually infectious. This is probably what I love the most about Imperial.

People definitely don’t go clubbing and such as much. I don’t care. I have never and will probably never go anyways. We usually have better things to spend our time on.

Demographics-wise, Imperial has a reputation of being full of Chinese students. Well, whoever said that have obviously never been to Vancouver.

The student population is probably about 10% Chinese, and the rest is a pretty even mix of all European countries. There seem to be quite a few people from Germany and Spain, but there are some British as well, Singaporeans, and eastern European/Russian. North America is definitely under-represented. But that’s fine, I know way too many North Americans already on the other side of the pond :). It just seems like North Americans don’t generally come to Europe for university, and British people don’t generally go to North America. North Americans are so under-represented here that we actually have a North American society.

Other Asian countries also seem to be under-represented. I haven’t met anyone from Japan or Korea, and I’ve met people from probably 15+ countries already.

It’s a much more culturally-diverse mix compared to even Vancouver, and it’s awesome! All the cool stories!


Imperial does have a high reputation (at least outside of the US), and I’m not sure if it’s a blessing or a curse.

It has been ranked anywhere from 2nd in the world to 20th in the world, depending on who you ask, and curiously, some people are quite caught up on that.

Did you know that people prefer to read magazines that rank their universities higher? Is it really surprising that British magazines all rank British universities higher, and American magazines all rank American universities higher?

They provide a few different measures, and add them together using weights drawn out of thin air, tweak them so the scores of the top universities come very close, and compare universities to second decimal places. You can tweak those things to say anything you want.

There is often an air of snobbery. People feel they are somehow superior just because they go to a university some magazine editor decides is better than another. It’s pretty annoying, but I suppose it’s unavoidable at any good university.

That’s much more prevalent among undergrads (at least the ones I talked to at society meetings and events).

I still remember one of my very first engineering lectures at UBC (that’s almost 7 years ago!). I don’t remember what course it was or which professor it was, but he said something to the effect of –

Be proud of what you do, because engineers are doing something valuable for mankind and making the world a better place… but not with the arrogance of those in the medical profession.

It’s easy to become arrogant when success seems almost too easy, and I’m sure I am arrogant at times, but it’s something I’m working to fix.

We must not fall victim to that. The day we start becoming arrogant is the day we stop learning, and engineering requires life-long learning.

*1: For my North American friends – in the UK, they use the term “course” also to refer to a degree program(me)

*2: One of the 2 biggest centres of performing arts in the world, the other being Broadway in New York. Dream for musical-goers!

London Part II – South Kensington, Imperial College London


My first week in London was spent buying the necessities, and I was busy being deathly sick for most of the second week, so it wasn’t until now that I had a chance to actually explore London a bit.

Sorry for the lack of blogging!

London is a very old city with some very interesting stories – Knights Templar, Freemasons, kings, queens, dragons, William Shakespeare, churches, etc, but that’s for another time.

This post is about a fairly new area called South Kensington. It’s fairly new by English standards – most of the iconic buildings in this region were built in the latter half of the 1800s.

Why South Kensington? Because that’s where Imperial College is, and I had a chance to walk around it this afternoon and snap a few pictures, thanks to the generosity of the people in Department of Computing responsible for arranging schedules.

It’s quite an interesting area, with various museums, one the most famous concert halls in the world, various embassies, various colleges, and quite a few billionaires.


Beit Quadrangle, our student union building that’s apparently also a hall for undergrads?


Right across the street from the college (and Royal College of Music) is Royal Albert Hall, one of the most famous concert halls in the world, where the Queen goes to watch performances.


Royal College of Music, sharing the same block with Imperial College


Imperial College geoscience building (I think?)


National History Museum, just south of the college.


The other side of Royal Albert Hall, from Hyde Park. Not pictured, but to the right is Royal College of Art – according to Wikipedia “the world’s most influential art and design institution”.


Albert Memorial in Hyde Park, across the street from Royal Albert Hall.



Hyde Park.

Hyde Park is the site of The Great Exhibition in 1851, where inventors from many nations were brought together to showcase their inventions. Surplus from the exhibition funded all 3 museums in this area – Victoria and Albert Museum, the Science Museum, and the National History Museum (according to Wikipedia). The Exhibition essentially kickstarted the development of this area.

Nowadays it’s a lot more humble. It’s just a very big open space with a big pond for the most part, which is actually quite nice, in the most expensive area of London where every square metre of land probably cost more than any house I have lived in.

It’s pretty cool that Imperial College is in such a cultured area, and my daily commutes take me through all these sights people come all the way to London to see, though the school kids in long lines do get annoying sometimes… and it’s a little inconvenient that we don’t really have any off-campus dining options, since anything in this neighbourhood is probably £20+ per meal.  Maybe I just haven’t looked hard enough?

London Part I – You Are Poorer Than You Think


That’s the view out of my room in London. Isn’t it amazing?!


People say London weather is bad, but I’ve actually found it quite agreeable, compared to other places I’ve lived in.

OK, that’s not saying much. I’ve lived in Taiwan and Vancouver, and after living in those 2 places, one would probably find the weather in northern Russia agreeable, too.

It’s end of September, and we still get quite a bit of sunlight. Temperature has been fairly steady at 15-20C, which is quite a bit warmer than I had expected. I still go out in a tshirt, though I do seem to be a little out of place at times. I blame that on being Canadian :).

A new Welsh friend told me London has its own little weather system, and is usually a few degrees warmer than the surrounding areas. That’s pretty cool. I haven’t had a chance to verify it, yet. It’s actually pretty difficult to get out of London from the centre, just because how big it is.

It has been 1 week since I moved here, and I haven’t had a chance to see London much, since I have mostly just been shopping for the essentials, and food.


Like this £9.15 (~$15) meal from IKEA. I am pretty sure in the US this would have been less than $9.15.



I found that to be the general pattern here – a $5 thing in the US would be £5 here. Even after taking out the 20% VAT that’s usually included in prices, that’s still a good 40% more. I guess that’s the price to pay for living in the most expensive area of London!

After a while, you just learn to mentally replace the £ sign with $ instead of actually doing a conversion to evaluate the prices. Because if you do the conversion, you’ll starve.

The place I’m living in is just north of Shepard’s Bush, which is reasonably close to Kensington (where Royal Albert Hall, Hyde Park, and all the interesting museums and all the billionaires are), hence £££.

That’s pretty much by necessity, because some genius decided to build Imperial College in the middle of South Kensington, right across from Royal Albert Hall – quite possibly the most expensive land in all of the UK.

Well, it’s nice that I can go check out a museum or watch a musical between lectures at the most famous concert hall in the UK, but I think I’d be willing to give up that privilege for £300 per month lower rent.


It did take a bit of adaptation to get used to not having a car.

While I did walk a fair bit in Canada, I’ve always had access to a car, and always drove if I had to buy a lot of things, etc.

Here, it’s all buses and tube. I had to make 3 separate trips to IKEA just because I can’t physically carry that much stuff in 1 or 2 trips. I have never had to do that before.

It’s also slightly mind-bogging how big London is. In Vancouver, I can run from one end of the city to the other in about 2 hours. In London, 2 hours would move you by a few pixels on Google Maps, if you zoomed out to see all of London.

I have visited quite a few huge cities, but this is the first one I have lived in, and having to tube and bus everywhere really puts the size of the city into perspective.

If you are ever bored in London, you can just hop on the tube in some random direction and  sleep for 2 hours. When you wake up, you would still be in London. You’d be in a middle-of-nowhere borough you’ve never heard of, but still London.

For comparison, the population of London is about 8 million. Vancouver has about 500K, San Francisco has about 800K, and LA has about 4 million.

Compared to a similar sized city like New York, I quite like it.

It does have the usual problems with all metropoles – congestion, high prices, crime, pollution, etc, but overall, I like it a lot better than New York.

The streets are not nearly as clean as Vancouver, but also not nearly as bad as New York. Public transportation is much better – this one probably doesn’t really need an explanation if you have tried public transit in New York (subway drivers can randomly skip stations, take random detours, etc, and some lines still run at about 20km/h). London’s tube system has its quirks, but it’s pretty good on the whole, and very extensive. It is a little confusing at first with something like 8 lines and 250 stations, but that’s not really a problem if you have Google Maps. If you just follow their directions, the tube will take you to pretty much anywhere in London, relatively fast, especially in times of congestion. I like it over Vancouver’s system because Vancouver relies on busses too much, and busses do get stuck in traffic. Obviously also like it over Bay Area’s lack of a system at all.

Trains are reasonably fast. They aren’t as new and high tech as in Vancouver or Washington DC (their stations are absolutely gorgeous btw), but they do work, and are reasonably clean.


Also, their rhinovirus strands are amazing! At least the one I am sampling at the moment. Give it a try?

Technology Progression in Harddrives

I recently had to destroy a few hard drives before throwing them away, since they contain confidential information.

I could have used dd (or any secure erase program), but some of the drives have the older PATA (IDE) interface, and I no longer have a computer with a PATA interface. Which means, it’s time to get…


20140820 001734

In this picture I have taken out all the PCBs from the drives, as well as all the platters (the metal disks that actually store data).

  • 80 GB Seagate Barracuda ATA IV, date code says it’s from November, 2002.
  • 200 GB Maxtor DiamondMax 21, PATA, from October, 2006.
  • 2x1TB Western Digital Green SATA, from 2009.

The oldest of the drives are about 12 years old, and the newest about 4, and I thought it would be interesting to see exactly how technology has changed over the years, through these drives.

Obviously the biggest difference is in the platters. All 4 drives happen to have 2 platters – 40GB per platter on the 2002 drive, and 500GB per platter on the 2009 drive.

However, they looked exactly the same under naked eyes (exactly the same diameter!), and I don’t have a microscope, so I am going to focus on the PCBs, and look at how electronics design has changed, and how we are now doing the exact same thing differently from how we were doing it back in 2002.

Turned out we can learn quite a bit just looking at the PCBs!

All these PCBs do pretty much the same thing – power supply conversion, a main controller that interfaces PATA/SATA to the platter sensors, and a spindle controller to drive the motor, and some kind of memory for buffer. But their designs are quite different.

The first board we are looking at is the oldest one – the Seagate 80GB from 2002 –

20140820 004314

The central processor is ST branded, but I cannot find any information on it, probably because it’s a custom OEM chip.

– On the right is a Hynix HY57V161610DTC-6 – 3.3V 166 MHz, 16Mbits synchronous DRAM.

– On top of that is a ST M29W102BB – 1 MBit parallel flash memory with a boot block and 2 main blocks.

– Above that is a MSC LX8815-33 – dual channel 3.3V 1A low dropout linear regulator… with pretty crappy specs (1.1V dropout, up to 5mA quiescent current).

– On top of the main controller to the left is a chip labeled SH6950. I can’t find a data sheet for it, but according to some random site, it’s a TI spindle motor + voice coil motor driver.

– The chip on the left of the main controller is an agree MS455. No idea what it is.


Next up is the PCB from the 200GB Maxtor drive –

20140820 004319

– The main chip this time is an agere chip. They seem to be making a lot of hard drive stuff. No information can be found for this chip unfortunately.

– On the right is a Hynix HY57V641620ETP-6 – 3.3V 166 MHz, 64 MBits synchronous DRAM.

– The tiny chip on the bottom left is an Atmel chip, but the part number doesn’t return any results. I am guessing it’s EEPROM for the main controller.

– On the top left, the big TQFP chip is a Smooth 100369972, which I believe is a spindle motor driver

– The small chip above the main controller is a Fairchild chip. I am also not sure what it is, but since it’s connected to a big inductor and a big resistive divider, I am going to assume it’s some kind of switching regulator.

– On the top right there is another switching regulator, in QFN package. The printing on the chip is too small to read.


The last board we have here is from the 2009 Western Digital Green 1TB –

20140820 004324

– The main chip this time is a Marvell 88i9045-TFJ2 micro controller with external EEPROM (the Winbond chip just below)

– On the left is Hynix HY50U121622DTP-D43 – 2.5V 133 MHz 512 MBit DDR synchronous SDRAM (note that even though the clock frequency here is lower, data is transferred on both edges of the clock, so it’s equivalent to 266 MHz data rate SDR).

– The only other chip is a Smooth L7251 spindle motor controller.


It’s interesting to see how many electronics trends can be seen in just these 3 PCBs.

First of all, there is a clear trend in reducing number of ICs. The first board had 8 chips each doing very specific things, and the 1TB drive only has 4 chips, with most things integrated into the bigger chips. This is mostly to save cost. Many more chips are custom designed for single products now than how it used to be – IC manufacturers making small chips to do very specific things, and circuit designers putting them together to do what they want (like in the first hard drive).

Looking at the PCBs, I am surprised most ICs are still TQFP instead of BGA. Most electronic designs have moved on to BGAs now for space saving and better signal integrity at higher speeds. Interestingly, the only BGA chip here is the main controller on the oldest board. However, looking from the side, it seems to be about 1.2mm pitch (distance between balls), which is huge by today’s standards. Most modern chips are going 0.5mm or lower, and require more precise placement using optical alignment.

The oldest PCB has almost all 0805 (0.08” by 0.05”) and 1206 passive components. That’s pretty huge. On the 200GB board, we are mostly down to 0805 and 0604. On the 2009 board it’s almost all 0402.

The routing on the oldest PCB definitely looks less dense than the other 2. It looks like it’s roughly 8 mil traces and 8 mil spacing, whereas the newer boards use 6/6.

The old board also seems to have a HASL (hot air solder levelling) finish (or some kind of organic finish? though I don’t think they existed back then), whereas both the new boards use immersion gold (ENIG). Immersion gold offers better planarity (good for BGA chips that require pads to be perfectly flat) and oxidation resistance (so boards have longer shelf life before soldering).

Since the new board also doesn’t have any BGA chip, there is really no need for ENIG. It’s possible that the cost is not much higher in recent years, so they are just doing it with all boards now. Maybe ENIG wasn’t popular back then? Maybe it didn’t exist back then? It was already popular when I started doing electronics about 5 years ago.

Cool eh?

Technology in the (Small Airplane) Cockpit

When I learned to fly 2 summers ago in San Jose, California, I flew this beauty –


A 1960s Bellanca Citabria 7ECA.

It’s about as simple as airplanes get – no flaps, constant pitch propeller, tailwheel, center stick, front and back seating, and a 115 HP Lycoming engine with a gravity fed fuel system.

And the cockpit looks something like this (I don’t have a photo of the instrument panel for that plane, so this is a random Google image for a plane of the same type) –


There is no attitude indicator. No heading indicator. No any kind of navigation equipment. Basic 4-dials Mode-C transponder (required because we were under SFO Class B). 90 kt cruise.

Simulated instrument flying was very interesting. I did all my PPL hood flying with a partial panel!

On the plus side, on the PPL written exam there are always a few questions on minimum equipment lists – is this instrument legally required for day VFR? Is that instrument required for night VFR? etc. Those questions were very easy for me, since all I had to do was to think about the cockpit – if it’s on my airplane, it’s legally required to be there, otherwise it’s not.

I did OK. Flew some cross-countries using my Android tablet for navigation, and passed my checkride.

1 day after I passed the checkride, before I had a chance to exercise the privileges of my FAA license even once, I had to move back to Vancouver… and with that a 1 month wait for the highly efficient organization Transport Canada to issue me a license validation so I can fly in Canada.

As soon as I got the validation (and medical), I took a virtual Google-stroll to the airport to find an airplane to rent… there were a lot of 30 years old Cessna 172s as I have expected, but then I also found out that I can also fly an almost fully loaded Diamond DA40 with G1000 avionics for only about $20/hr more… sounds like a deal to me.

I had never flown a technologically advanced aircraft before, but I decided to give it a try, and I was immediately hooked.

This is the airplane I have been flying –



2005 Diamond DA40. All composite. G1000 avionics. Way more navigation instruments than I need or know how to use. 2 axis autopilot with altitude hold and preselect. A database of the entire world’s airports and airspaces. Live weather and TFRs from satellite link. Secondary surveillance radar with active interrogation. Fuel-injected 180hp Lycoming engine with variable pitch propeller. 145 kt cruise.

I pretty much just time-traveled about 40 years forward, from a fabric Citabria with barebone instruments to a Diamond Star.

And I don’t think I will ever turn back, if I have a choice. I really like all the additional safety technology provides.

There is a lot of debate in the aviation community about the merit of these glass airplanes, which is what prompted me to write this post.

How has technology made flying safer?


It is more tiring to fly the DA40 by hand in cruise than other aircraft of the same class, just because it cruises at 145 kt.

On a C172 if your attitude is 1-2 degrees off in cruise, you’ll probably end up climbing or descending at 100 fpm or so.

On the DA40, if your attitude is 1-2 degrees off, you’ll be off the assigned altitude by couple hundred feet in no time (though autopilot will say “leaving altitude” in that case to warn the pilot, if properly set up, even when not engaged).

Yes, hand flying the DA40 is definitely doable, and I have done it a few times, but nowadays I almost always use autopilot in cruise (when I don’t have a passenger to use as autopilot).

That way I don’t have to spend 50% of my mental capacity on flying the airplane, and can focus on other things like navigation, traffic scans, and keeping up with building a mental map of where everyone else is by listening in on the radio, etc.

Computers are good at mundane tasks like keeping an airplane straight and level. Why have a human do it?

Has my piloting skill gotten rusty due to my use of autopilot? I don’t think so.

I still do most of climbs and descents by hand, and those are more difficult than cruising anyways.

Crew fatigue is a common contributing cause of accidents, and autopilots make flying much less tiring.

As an aside, unintentional flight into IMC is another major VFR-pilot-killer, and having autopilot (and knowing how to use it) makes those situations much more survivable. In fact, the new FAA recommendation for people getting into those sticky situations is to use autopilot to get out, if one is available and the pilot is familiar with it.

Situational Awareness

If I have a gradual loss of oil pressure or alternator failure in flight, how long would it take me to notice it?

I try to glance at engine instruments every 10-20 seconds, but I know I have gone for couple minutes without glancing at them in very high stress situations – eg. flying a long final into busy airspace, being 5th in sequence and looking for 4 planes in front, while trying to maintain 60 kt (stall warning constantly blaring) in strong crosswind because the airplane in front is a C152 flying a 50 kt approach, taking care to not stall, all while configuring the airplane for landing and looking for a VFR checkpoint I am not familiar with.

I am a creature of probabilities, and in those situations, my chance of dying from a mid-air collision is much higher than my chance of dying from a loss of oil pressure, so I prioritize accordingly.

But wouldn’t it be nice if there is a computer that is constantly checking those things for you, and will give you a nice audible warning if anything is wrong?

G1000 does that.

It also monitors the integrity of input sensor data (presumably by cross-checking), and will put big red crosses over instruments if there is any doubt, just so the pilot won’t use wrong information.

Sure, it’s slightly annoying that it always warns about loss of oil pressure while idling the engine on the ground, but I would still rather have it than without.

Could something like that have saved Air France Flight 447?

There is also real time weather available (METARs, TAFs, satellite overlay, position of recent lightning strikes, etc), with obvious benefits.


Every pilot in Vancouver has a story or two.

Not fun stories. Stories of close calls.

Every year, 1 or 2 pilots die in Vancouver from mid-air collisions.

It doesn’t matter how good the traffic scan is – every pilot will get a few close calls if they fly enough (a few hours is usually enough).

It’s a scary place to fly, especially since unlike in the states, Vancouver Terminal doesn’t offer flight following, and everyone is on their own.

A series of mid-air collisions and close calls starting from the 1956 accident over the Grand Canyon led to regulations that made collision avoidance equipment (TCAS) mandatory on commercial flights, but for another few decades those equipment were too expensive to be installed on small airplanes.

That started to change couple years ago, as more and more small airplanes are now equipped with traffic avoidance systems, including the DA40 I fly, and I am very thankful of that, as a Vancouver-based pilot.

It’s an active secondary surveillance radar, which basically means it can see all transponder-equipped aircraft. Integrated into G1000, it can display all those traffic on the moving map tagged by their relative altitude, and issue audible warnings (in ATC traffic advisory format) when another airplane gets too close.

Obviously it doesn’t relieve the pilot of the duty to see and avoid, but it’s still very nice to have.

It has already helped me see a few airplanes that I probably would have never seen, including one time when another faster airplane was coming up from behind, and ATC didn’t give us a warning until they were way too close.

There is no way we would have seen that guy.

You could say my traffic scan is not good enough, but I’d rather be alive than dead even if that’s true.

Mid-air collisions are one of the leading ways to die in an airplane, so I always use everything I have at my disposal to stop it – that includes my eyes, my passengers’ eyes, ATC, and now TAS.

In the future, hopefully TASs won’t be necessary once US NextGen ATC services become widely available (traffic maps are much cheaper to implement using ADS-B). Hopefully we will see much wider adoption then.

But until then (and in Canada, which apparently has no plans to implement ADS-B), TASs are very nice to have.


Virtually all airplanes being made today have glass cockpits, so technology is taking over the GA world whether we like it or not, and rent on glass airplanes is already getting pretty close to rent on older airplanes.

In places like SF bay area, there are so many G1000 (and Avidyne) airplanes that they don’t really cost much more than 1970 spam cans to rent (though they are still quite a bit more expensive to buy).

You could say we don’t need technology as long as we keep our eyes open at all times, but that’s much easier said than done, and flying for couple hours in one of the practice areas in Vancouver on a sunny Saturday afternoon may change your mind.

Everyone makes mistakes sometimes, and technology makes mistakes less deadly, on average.

New small airplanes now have almost as much technology as Boeing 747s from just a few years ago, and for some reason some people are saying that’s a bad thing and is totally unnecessary. I beg to differ.

Sure, they tempt some people into doing things they would not otherwise have done, like flying into IMC without proper training, but the same can be said for almost all safety features, and history has proven again and again that we cannot stop Darwin.

If cars don’t have seat belts, I probably would drive much more conservatively than I do now. Why isn’t anyone arguing for the removal of seat belts?

Solar “Freakin” Bullshit

If you are on Facebook, Twitter, or any other social media network, chances are you have already heard of this project – (I obviously do not endorse it). 20140429030846-LEDs_-_white If you haven’t heard of it – it’s a project that wants to replace the surface of all roadways, sidewalks, parking lots, airport aprons, runways, etc, all with these 7″ hexagons that contain about 50 high powered RGB LEDs to be able to make configurable patterns, photovoltaic cells (solar panels) to power them, heating elements to melt snow, and about 1″ thick of specially textured glass on top. The panels can also communicate with each other wirelessly, presumably as well as with a base station of some sort, through the mesh network.

Sounds cool. Looks cool (*1). Is definitely technically plausible. What’s not to like about it?  

*1: Though if you look at illustrations on that page, you’ll find something strange – those tiles seem to have much higher resolutions than would be possible using just 50 LEDs per panel.

If you actually read through the Indiegogo page, you’ll find that it has no mentioning of cost at all, except that it supposedly “pays for itself” by generating electricity using those solar panels.

Yeah well, if that’s the case, why aren’t all our roofs covered by solar panels already? Solar panels do not pay for themselves at the present time, in most cases (only in places with extremely high sunlight, and very high electricity prices). That’s why.

If you look at the FAQ, there is a question “How much will your panels cost?”, and the answer essentially boils down to “we are not going to tell you”.

They already have working prototypes, and they can’t even give us an estimate of production price? That’s something all engineers are trained to do from very early on. How can you go as far as having working prototypes, and not having even an estimate of production cost? Or maybe they are hiding something?

But fear not, there is enough details on that page that, with the power of back-of-the-envelope calculations and Wolfram Alpha, we can come up with an estimate that should at least be within an order of magnitude of the actual cost –

How much does one panel cost?

A 7″ hexagon has an area of 0.082 m^2


Judging by the picture, and by the fact that they need to be able to support large trucks, the glass should be at least 3/4″ thick. How much does 3/4″ thick glass cost? I checked a few places, like this one, and they are pretty consistent. $21 per sqft, or $226 per sqm. I’m assuming margins are not very high in the glass material business, so let’s say the cost is $150 per sqm? The glass on a panel would cost $12.30, assuming post-processing is free.


Judging by the picture above, there are about 50 LEDs on one panel. Since they need to be able to change colors, they need to be RGB LEDs. Since they need to be visible under direct sunlight, through 3/4″ thick of textured glass, they need to be at least 1W, possibly much higher.

1W RGB LEDs cost about $8 each at 5k qty. Let’s say $5 at practically infinite qty? That’s $250 per panel.

Solar Panel

How much do solar panels cost?

The best solar panels available right now are about 20% efficient, and 1 m^2 gets about 1KW of sunlight on average, when the sun is directly above, with no clouds.

So with a 7″ hexagonal tile, we’d be looking at about 16.4W per tile.

Actual average power output will be 5W or so since it won’t be noon 24/7, but we still need to get 16.4W panels to get 16.4W when it is noon.

Before we go into price… did you notice something doesn’t add up? How the @#$@#% are we going to power 50W worth of LEDs with just 5W average?

Ignoring that for a second, the current cost of solar panels is about $0.70/W. That’s $11.48 per tile. Not quite as bad as I had imagined!


This is harder to estimate, but the wireless module will cost at least $5 at qty, and the CPU $2, and there will also need to be at least 3 MOSFETs per LED at maybe 3 cents each, for a total of ~$5.

I’m going to estimate $20 in electronics per tile. Unfortunately I cannot really justify this estimate further, but I believe it’s a very conservative estimate.

I am not going to include the cost of a 50W power supply… because we still don’t know where that power is coming from.

Total cost

Taking into account those components above, we are looking at a total of $294 per tile, most of which is from the high powered LEDs. The cost of the solar panels turned out to be almost negligible.

How many panels do we need?

First of all, what’s the total length of roads in the US?

6.5 million km in 2007.

Or just over 2 million km if we only count motorways (freeways), highways, and secondary roads (secondary roads are main roads in cities that feed into the highways). Let’s use this number, since it’s smaller.

How wide are the roads? According to the US Standards for Interstate Highways, the minimum lane width for a highway is 3.7m. Or 7.4m both ways. That’s the absolute minimum, and doesn’t include sidewalks. So let’s say the average width of roads is 10m, as a conservative estimate.

How much road surface is that?

2e10 FREAKIN m^2

How many tiles do we need?

2.44e11 FREAKIN tiles


But how much will that cost?

71 trillion dollars.

Note that the cost is almost entirely proportional to area, so changing the tile size won’t change that number significantly.

Since we usually have trouble taking astronomical numbers like that in perspective, let’s compare it to a few other astronomical numbers –

The entire annual US military budget is 0.683 trillion (it’s already the highest of the entire world), which is about 4% of the US annual GDP of 16 trillion.

That means, if we increase taxes to 100%, and shut down education, medicare, the military, and all the other government services, we will be able to make all our roads (well, just the big roads) shiny in 4.5 years… assuming we can still maintain a $16T GDP in those conditions.

Cost of labour, transportation, and disposal not included. They say they are aiming for 20 years life time for those tiles. That’s funding-seeking-speak, so let’s say 15 years as a more realistic estimate.

That’s still $4.7T/year, or almost 7x the military budget.

Even if I am off by an order of magnitude, that’s still 470B/year.

What else is wrong about this project?

Many things, in fact.

I’m not going to get into all of them, since I don’t have all day, but here are few of the best –

Charging EVs with solar panels

Inductive charging is 60% efficient over a gap of 12 cm, and decreases rapidly as the gap increases. Most EVs sit more than 12 cm from the ground. Even if one EV can draw power from 50 tiles, that’s 250W total, and about 125W after inductive loss.

A Tesla Model S base model has a 60kWh battery. That’s almost 21 days to charge at 125W. In those 21 days you can drive 230 miles.

More realistically, if you use it as an extra/bonus thing, if you drive for 4 hours on those tiles, you would have charged your battery 0.7%.

It will cost the same as paving


Sure, you can sell the electricity, if you don’t even need 5W to run a tile. How much can you sell it for?

Assuming a 17c/kWh buyback price (that’s the current solar buyback price in Georgia, and will likely decrease significantly if supply increases significantly), over 15 years, you would be able to sell 5W for $111, likely much much less if all the roads suddenly start selling back power, and government can no longer afford to heavily subsidize selling back solar power (or if the government own those tiles).

Why include solar panels on those tiles?

Why not just put them aside, or get rid of them completely since most of the power is going to have to come from the grid anyways?

The problem with solar power is not space, otherwise all our roofs will be covered by solar panels already.

Putting them in road tiles is solving the wrong problem.

Why not just build solar power plants elsewhere (in a desert), centrally? That would be much more efficient, and with HV power transmission, there is actually not much losses.

We don’t have many solar power plants in use today, mostly because they are still cost prohibitive.

Putting them in road tiles below 3/4″ of glass won’t magically make them any more efficient… quite the opposite in fact.

Of course, the real reason they are doing this is because people have warm and fuzzy feeling about solar power, and we WANT to be convinced that there is a way to use it cost effectively, and we are willing to turn off our common sense for a little while to get that feeling.

How do I identify nonsense like this in the future?

When you see a Kickstarter or Indiegogo campaign that is as technical as this one, yet doesn’t mention any numbers, and gives you everything in qualitative terms, you should be wary.

Just because the tiles cost money and you can sell electricity for money, doesn’t mean the amounts are equal, or even comparable in order of magnitude.

Just because you can theoretically charge EVs through induction, doesn’t mean you’ll actually be able to charge them at a non-negligible rate, and be worth the added cost.

It’s all in the numbers.

Just because 500 media outlets reported on it, doesn’t mean they have actually gone through it with common sense and a little bit of scientific literacy.

I suspect even the scientifically literate among the journalists actually intentionally turned a blind eye to the obvious infeasibility of this idea just so they can write an article and get viewers excited.