"Real Artists Ship"

Colin Johnson’s blog

Archive for July, 2014

Continuous Improvement

Wednesday, July 30th, 2014

Is there anything that we can get continuously and consistently better at by extensive and sustained work longer than a decade or so?

A question that has been asked a number of times on discussion boards is “could someone of decent fitness reach Olympic standard at some sport if they started at the age of 25?”. The usual response is that there are some examples—equestrian sports, sailing, archery, shooting—where there are serious international competitors aged in their 40s and 50s, and so it doesn’t seem unreasonable that someone could get there starting at 25. A further strand to this argument is that there aren’t only competitors in that age range there. There are perfectly competitive people in, say, their twenties competing against the older competitors. It isn’t as if you need to start at 5 years old and put in 50 years of practice before you stand a chance of being up there with the best. You plateau out—whether at local club level or Olympic gold medal level—after a number of years of sustained effort. You don’t just continue, Duracell-bunny-like, to get better and better as you put in the effort over the years.

So, experience might not be a disadvantage in these activities, but beyond a certain (rather advanced!) point it isn’t an actual advantage either. Are there any areas where it is almost necessary to have put in the years to be any good? I struggle to think of anything. Let us consider some other areas of human endeavour.

In science and maths, there doesn’t seem to be anything like this. The rather addleheaded idea that “mathematicians are burned out at 25/30/35” is on the wane. Nonetheless, it seems that with the right combination of study and focus and talent you can get to a research-frontier understanding of most areas of science and maths in about ten years of hard study, from a fairly standing start. Some topics have gotten pretty complex, but not so much that you need to spend ten years learning the basics, then another ten years learning how to use those basics, then another ten years learning about the real frontier stuff.

Craft skills similarly seem to need a number of years to reach professional standards, after which there isn’t really a notable advance in skill. There might be more diversity of practice, richer application of skill, etc., but isn’t as if we only regard as world-class the craft-work of makers in their 60s, say. We would probably make a distinction between the work of a one-year-experience potter and that of a ten-year-experience one in terms of basic skill. But, we wouldn’t make the same distinction between one of twenty-one-years-experience and one of thirty-year-experience; we would talk instead of the ideas that they use their skills to execute, not that the thirty-year one was better at handling the materials.

The arts are more complicated. It is possible to be a child-genius performer. Less so a creator. With the exception of the occasional high-concept work, the number of writers/composers/painters who gain recognition equal to that of the established practitioners at the age of 15 are nugatory. Novelists in particular are generally older. This is presumably something to do with the sheer length of novels. To bash through a few hundred mediocre poems, songs or drawings is just part of the process of becoming a practitioner in those areas; to bash out a few hundred novels whilst getting to grips with the medium is impossible. In music the 10,000 hour “rule” seems to hold sway, overall. A top-ten band might seem to be full of fresh-faced youths, but probably fresh-faced youths who have been practicing guitar in the garage every spare hour since they were 11. Again, the high-concept exception applies, with punk as a clear example. But, again, once we are past the 10,000 hour mark, we aren’t really into “improvement” any more, we are into depth and diversity. Orchestral conductors are usually older, but that is probably a “second job” phenomenon, you probably don’t become a conductor until you have spent a good number of years studying an instrument and being a player. A similar argument applies to football managers, another wunderkind-free zone.

Talking about second jobs, there are the areas in which a certain amount of relevant lived experience is appropriate. There aren’t going to be any whizzy 12 year old marriage guidance wunderkinder. But the relevant experience isn’t in the job as such; it is that the job builds on reflection on life experience.

Perhaps parenting? Parents are said to be much more relaxed with their second an subsequent children, and I’ve met the occasional parent of four or five kids who basically seems to have “a system” after the first two or three, but it doesn’t seem like the tenth would be any better parented than the fourth (indeed, sheer weight of numbers might make it harder). Similarly, the advice of well-intentioned grandparents doesn’t seem obviously better than that of the parents.

So, are there any examples? Any area where the second decade or more of work gets you to a different level of achievement, such that people at the end of that first decade are regarded as amateurs/students? I struggle to think of one.

“In Anger”

Wednesday, July 30th, 2014

I’m surprised at the number of people who don’t know the phrase “used in anger”. Also, if you use it at the person you are talking to doesn’t know what you are taking about, it sounds like quite a nasty accusation: “no, of course I’ve never got angry about this, what sort of person do you think I am?”.

You don’t want to do this (1)

Thursday, July 17th, 2014

It isn’t good when you see something like this:

It is recommended that you use a single window for this system. More than one may cause unexpected behaviour.

The Extensional Revolution

Wednesday, July 16th, 2014

We are on the threshold of an extensional revolution.

Philosophers draw a distinction between two ways of describing collections of objects. Intensional descriptions give some abstract definition, whereas extensional descriptions list all examples. For example, consider the difference between “the set of all polyhedra that can be made by joining together a number of identical, regular, convex polygons with the same number of polygons meeting at each vertex” (intensional), and “the set {tetrahedron, cube, octahedron, dodecahedron, icosahedron}” (extensional).

Despite its claims to be (amongst other things) the science of data, computer science has been very intensional in its thinking. Programs are treated as realisations of descriptive specifications, satisfying certain mathematically-described properties.

As more data becomes available, we can start to think about doing things in an extensional way. The combination of approximate matching + the availability of large numbers of examples is a very powerful paradigm for doing computing. We have started to see this already in some areas. Machine translation of natural language is a great example. For years, translation was dominated by attempts to produce even more complex models of language, with the idea that eventually these models would be able to represent the translation process. More recently, the dominant model has been “statistical language translation”, where correlations between large scale translated corpora are used to make decisions about how a particular phrase is to be translated. Instead of feeding the phrase to be translated through some engine that breaks it down and translates it via some complex human-built model, a large number of approximations and analogies are found in a corpus and the most dominant comparison used. (I oversimplify, of course).

More simply, we can see how a task like spellchecking can be carried out by sheer force of data. If I am prevaricating between two possible spellings of a word, I just put them both into Google and see which comes out with the most hits.

Once you start thinking extensionally, different approaches to complex problems start appearing. Could visual recognition problems be solved not by trying to find the features within the image that are relevant, but by finding the all the images from a vast collection (like Flickr) that approximately match the target, and then processing the metadata? Could a problem like robot navigation or the self-driving car be solved by taking a vast collection of human-guided trajectories and just picking the closest one second-by-second (perhaps this corpus could be gained from a game, or from monitoring a lot of car journeys)? Can we turn mathematical problems from manipulations of definitions into investigations driven by artificially created data (at least for a first cut)?

The possibilities appear endless.

Dilemma (1)

Wednesday, July 16th, 2014

Here’s an interesting situation. Several times a year, I take part in university open days, where I sit behind a desk answering questions about courses from prospective students. Typically, at the undergraduate open days, the punters consist of a shy 16/17 year old and one or two rather more confident parents.

Here’s my problem. I don’t want to make the assumption that the older person is the accompanying parent and the younger person the prospective student. I’d be mortified if I made that assumption on the day that a parent, bringing their child with them for moral support or lack of childcare, was the prospective student. But, this happens so rarely that the parents and student just sit down assuming that I am going to read the situation as the obvious stereotype.

How should I react in this situation? Asking “which of you is the prospective student?” is treated as a joke or, more troublingly, as evidence of density or weirdness on my behalf. But I still feel uncomfortable making the assumption. I’ve taken to starting with a broad, noncommittal statement like “So, what can I do for you?” or “What’s the background here then?” and hoping that it will become obvious. That isn’t too bad, but there might be a better way.

More abstractly: we try to avoid stereotypes and making assumptions about people and situations based on initial appearance. But, what do you do when the stereotype is so commonplacely true that even the people being stereotypical are expecting that you will react using the stereotype as context?


Wednesday, July 2nd, 2014

I didn’t realise that they were so desperate for members (from my spam folder):

"Illuminati Membership Application" repeated many times


Wednesday, July 2nd, 2014

The new “Methodis Bottom” church?
methodis&#x22A5 church

A Plan for Antibiotic Resistance

Wednesday, July 2nd, 2014

Antibiotic resistance is in the news again today. The focus for a solution is primarily an economic one. The economic benefit to pharma companies to discover antibiotics is minimal, because the market for new antibiotics should, rationally, be small; unlike new drugs in most areas of medicine, which will rapidly displace older drugs through a mixture of greater efficacy, fewer side-effects and canny marketing, there is a motivation for new antibiotics to be held back as a “drug of last resort”, and only used when existing antibiotics fail. This offers the most sensible way to avoid bacteria evolving to become resistant to the new antibiotics. As a result, the focus is on the creation of artificial market incentives to support pharma companies in the development of those antibiotics, despite the small market that they will have.

Here is an alternative suggestion. Instead of routinely using all currently available antibiotics, have a schedule where every few years, a subset of antibiotics is authorised for use. Then, after a few years, these are proscribed from use, and another subset used. The ides is that during the “fallow years” when a particular antibiotic is not being used, the bacteria will have no evolutionary pressure to maintain that resistance, and an energetic cost to keeping it, and so over time it would lose that resistance. The timescales would need to be worked out based on the timescale during which antibiotic resistance is gained/lost.

Critical to whether this would work is the question of whether antibiotic resistance would be lost by the bacteria on a short enough timescale. There is an informal discussion of this here, which suggests that is can be retained for a long time. But, it would be interesting to know if this has been verified experimentally.