2014-08-26

Consumer grade health monitoring equipment.

Right now a somewhat new category of health/medical equipment is beginning to emerge.  These devices are not your traditional medical devices, with their bureaucratic, long-winded approval processes that guarantee that whatever comes out the other side is ten times expensive and several years behind the technological status quo.  These devices are for everyday use of consumers to have some idea of how their health is doing.

When I talk about measurement devices to track your health, you probably think of glucose meters and blood pressure measurement cuffs.  Expensive stuff that needs to be carefully calibrated, tested, rated and approved.  Well, those are critical and very nice to have.  But there are lots of other parameters that you can look at and analyze to gain useful information.

For instance body weight.  For people with certain medical conditions, keeping a close eye on developments in their body weight is critical.  And this is much harder than you would think using old-fashioned bathroom scales.  Since your weight will vary throughout the day and even between days, trends can be hard to spot if you do this by eye and by memory.  But if you store the measurements and you have software that can smooth out the rather noisy data, you will be able to spot trends much easier.  For instance I can easily spot a 200 gram gain or loss over a week using my internet-connected bathroom scales with data visualization software -- I seriously doubt that even if you record your weight carefully with pen and paper you'll spot that change unless you do a bit of work.

And that's an important point.  If you want people to measure something related to their health, you have to automate it and preferably, not have it interfere with their lives.

Taking a weight measurement with a WiFi-connected bathroom scale takes zero effort.  You just step on the thing for a few seconds while you are brushing your teeth and you are done.  If you brush your teeth 2-3 times per day, that's 3 datapoints right there.  (Compare to one datapoint at your doctor's office during the yearly checkup.  An almost worthless measurement).

Taking a BP reading, on the other hand, is a pain in the neck.  The traditional way to measure BP is with a cuff -- a device called a sphygmamanometer.   The measurement procedure calls for sitting down, putting the cuff on, and performing a series of measurements.  To the greatest degree possible you have to try to eliminate sources of error and to make the measurement the same way every time in order to produce data points that can be compared reasonably well.  This takes real effort.   So you won't get people to gather lots of BP data points every day.  This is a parameter in dire need of a proper measurement technology.

What will be an important focus is how we can instrument ourselves and our surroundings in order to measure health-related parameters without effort.  And there are a lot of opportunities. Just take your bed.  You can monitor movement to determine how well you sleep.  It should be possible to pick out your pulse either by sound, or by making sheets out of fabrics that can utilize contact with your skin and make use of conductive materials embedded in the sheets.  If you have enough leads, proper signal processing and some software to figure out how the sensor grid is aligning with your body, there is no reason you wouldn't be able to produce many hours of EKG-data every night. Meaning that not only will you know more about your heart than was previously possible -- it is going to be possible to detect heart problems much, much, earlier than before.
Ditto for breathing.  Combine audio analysis with the minute movements of your bed as you breathe and it should be possible to detect a number of anomalies.

Recently I built a prototype for monitoring body temperature while in bed.  My theory was that it should be possible to automatically detect if someone is running a fever.  Without having to get up and use a traditional thermometer every now and then.  The sensors are just embedded in your bed (no pun intended) and you log the measurement values over the network.  It took me just a few afternoons to build and even after the first full night of testing it was obvious that non-intrusive fever detection is feasible.  I'm still working on it, but I've already reached the cost goal:  you can do it for under $100 in materials cost.  And that is for a WiFi-enabled device with a big, bright LED display.

So to sum up:  here are some of my beliefs around consumer health monitoring:

  1. We have to develop cheap, automated technologies for health monitoring to free up the time of medical personnel so they can spend less time on menial measurement tasks which they are not very good at anyway since they are humans -- and humans are terrible at observing and recording physical phenomena.
  2. It has to be non-intrusive.  If it takes effort, users won't stick to it.  The goal has to be to hide these measurement devices in everyday objects and to opportunistically gather data whenever data is available.
  3. New high quality sensors available for very little money combined with cheap connectivity and high capacity computing means that observations that would have been pretty hard to come by a few years ago are now easily attainable.
  4. Privacy and security will be a huge problem.  On one hand you want to keep this data private. On the other hand, you want it to be easily available to you and your healthcare professionals.
  5. Regulatory bodies and traditional makers of medical technologies are slow and incremental -- meaning that there will be an unresolved tension here for years.  Ultimately this means that public healthcare organizations will forego incredible cost savings opportunities because they lack the kind of people needed to figure out how to expedite this.

(When I say that humans are terrible at measuring things, I mostly base this on papers I've read that quantify the error rates of defined measurement tasks where the accuracy of the measurement devices are known.  For instance for a given urine collection device used in many hospitals the theoretical accuracy would lead you to believe that errors would typically be in the 2-5% range.  The observed error is as high as 25%+.   Measuring well is hard.  Leave it to the machines if you can.

Of course, I am beating myself up over being really bad at referencing these papers when I discuss things or when I write.  This leads to misremembering the precise figures as well as not being able to provide helpful references for people who wish to understand these things.  I'll try to get better at this. Sorry.  Of course, it also doesn't help that some of these papers are behind paywalls -- which is just the disgracefully obsolete way scientific publishing works)

2014-07-09

Teachable moment: Watch Polar die.

If you want an opportunity to watch an established market leader falter and eventually die, you should have a look at Polar Electro.  I think it is very likely that they are going to be the next Nokia.  Grab some popcorn and watch as they are going to get their lunch taken away by newcomers.

Years ago Polar would be your go-to brand if you wanted a heart monitor.  By virtue of the competition being either non-existent or offering significantly worse products.  I remember buying one of their high end models which would log my pulse at 5-second intervals or so.  It was incredibly fiddly to use and it didn't offer much in terms of usable software or meaningful interoperability (which in my book means: easy access to the raw data).

Today I think Polar is poised to go the same way Nokia, Blackberry and a lot of other companies run by people who are mere businesspeople -- people who are out of touch with the rest of the industry and who are not properly attuned to how the world is changing.  Here are some reasons why:


Products are not improving.


Polar's products have not improved significantly over the past 15 or so years.  In fact, the last time I bought one of their pulse monitors it turned out to be a huge step back from my old monitor from the early 2000s.   But I digress:  their biggest accomplishment over the past 15 years is that last year they managed to put a GPS into a heart monitor.

I am not making this up.  This was their accomplishment in 2013.

Their heart monitors haven't grown any significant new features and they still rely on really clunky technology:  belts around your chest.   We do have the technology to avoid this crude method of measuring pulse,  so you would think that Polar would be really interested in developing it before anyone else does and aggressively push it into the marketplace.  They aren't.   They think their unpleasant and clumsy old solution is just dandy.

They also have not developed new categories of sensors.  As we speak there are lots of small startups making wearable technology to analyze any number of things from pulse, O2-saturation, muscle activity to what your brain is doing.  And everything inbetween.  And we are not talking about high end stuff that only professional athletes can afford;  we are talking about consumer priced stuff.


Their product strategy is obsolete.


If you have not visited the Polar website yet, please do.

Okay?

What did you see?  See anything familiar?  If you are perceptive you will no doubt have noticed that Polar is clinging to the old ideas of aggressive segmentation.  Without going through the tediuos exercise of detailing what segments Polar thinks exists, the main upshot of this is that it is not IMMEDIATELY clear to you what product you want.  This lack of clarity translates into hesitation.

Years ago, this is exactly what faithful Nokia customers experienced.  They'd need a new phone and it wasn't immediately clear to them which phone they wanted.  They'd have to browse and compare. Sometimes for days and weeks.  This is not a good way to approach the market.  In fact, it is exactly how you do NOT approach the market.

When product lines are properly put together, prospective customers do NOT need to visit a website or look through a catalog which product they want.  They know.  As an experiment: ask anyone which model iPhone or which Tesla they want.

If you think this comparison is unfair, you are absolutely right:  it is unfair because both Tesla and Apple know exactly how much product line complexity consumers can take.

The Polar product range is old fashioned in that they do not have the confidence to offer a single product line where only a single product line should exist.   They have multiple product lines and none of the marketing material gives you a very good reason why they would need lots of different devices to do more or less the same thing.


The competition will make far better products.


Right now everyone is waiting for Apple's watch.  When Apple enter a new product category they usually redefine what that category is all about.  They did so with computers, music players, phones and tablets.   Despite Steve Jobs no longer being with us,  there is a chance that they will do the same with watches.

Their newly unveiled APIs reveal that Apple are taking aim at two areas:  home automation and health & fitness.  Combine this information with whom they have been hiring lately and it is quite obvious that we can expect Apple to do some new stuff.  Like heart monitoring directly on your arm without the need for any extraneous sensing nonsense.  No more awkward belts around your chest.  No more pausing whatever activity you are doing to press your finger against a sensor.  None of that nonsense.

Expect a bloodbath.  Polar isn't going to be the only manufacturer which will find itself with obsolete products.   Products like the Pulse 02, Fitbit and a raft of other products will all be obsolete in a few months.  Most likely because they can't measure pulse properly -- which is to say: continously on the wrist.

I think that Apple are aiming for the ultimate body monitoring device that will start off by doing fitness stuff and which will then extend into health and possibly even healthcare.  This will accomplish two things:  it will move the goalposts and it will result in more specialty sensing hardware falling in price to consumer-tolerable price points.

(It is also going to lead to an interesting conversation on automating healthcare and how the world's most technophobic professionals will have to adapt to a new reality, but that's the topic of a future blog post).

Before long you will have cheap-ass heart-monitors that beat the shit out of anything Polar has ever made at a fraction of the cost.  Made by Chinese companies you have never heard of.  Designed by people in their 20s.

But first, someone has to move the goalposts.  And it isn't going to be Polar.

Watch and learn.


I think there is a significant chance Polar may be the next company to do a Nokia or a Blackberry.  This makes it worth keeping an eye on Polar over the next years to learn how companies shrivel up and die.

You might wonder why all of a sudden I am picking on Polar.  There's a good reason for this:  we are on the verge of an explosion in wearable computing.  In a few years wearable computers will be everywhere.

Polar has occupied a niche of wearable computing for decades -- a niche that is no longer a niche but is rapidly turning into a huge mainstream market.  You would think that years of expertise and experience would give Polar an unique opportunity to capitalize on their potential market growing by orders of magnitude.  My guess is that the opposite will happen:  the wearable market will develop in ways that will eventually make Polar irrelevant.

Hey, if everything else fails, Finland will be the place to learn about how to doom your company by allowing yourself to be disrupted to death.

2014-07-08

Ignorance and uncertainty.

Years ago I was involved in building a web scale search engine.  At the time this was the sort of undertaking that wasn't really well understood by a lot of people.  Information retrieval had existed for many years and was well enough understood, but almost nobody had any practical experience applying it at a massive scale.

And back in those days: 100 million documents was massive.  Computers were dog slow, we didn't have a lot of them and there wasn't a lot of open source software to help you do things at scale.  Not like today, where searching 100 million documents on a device that fits in your pocket is easily within practical reach.

What was hard back then is easy now.  Well, easier, at least.

But when I think back at those times it isn't really the technology that interests me.  By today's standards what we did was pretty crude,  but it did represent some of the best work that was done up until that point. Ever.  Anywhere in the world.  This is increasingly rare today:  the opportunity to work on something very few people have figured out how to do really well.

What interests me are two things.

The first is: what it takes to do something much better than anyone else.  I think the biggest advantage we had was that we had no idea what we were doing.  Which meant that we were not anchored by whatever other people did.  We truly had to explore and invent ourselves.

If you spend too much time trying to solve a given problem using whatever approach other people are following, you will be held back by the same limitations that they are going up against.  Worse yet, you will be tempted to not spend a sufficient amount of time trying to understand the problem you are solving.  Understanding the problem or problems you need to solve is always where you should start. This is why you should always try to spend a few weeks thinking about a given problem yourself.

This is why I avoid paying too much attention to what people who came before me did to approach a problem.  I might inform myself of their general approach, but I try to develop my own before devoting time to understanding theirs.  This doesn't always yield results, but if inspiration or deep understanding does strike, I will have an advantage.  Once I feel that I am starting to understand the problem and I know how I want to attack it I start to peek at what other people have done.

This is applicable to a surprising breadth of fields.

The other thing that interests me is the continuum from ignorance and uncertainty to cocksure certainty. While we were building a web scale search engine there was the engineering effort on one side and the sales people on the other side.

On the engineering side the basic question "can we do it?" wasn't a given.  Whether it was solving some problem of a deeply theoretical nature, turning ideas into working implementation or just delivering a product within an acceptable set of parameters (time, cost, quality etc.).   The sales side, as seen from the engineering side, seemed to take for granted that we had an infinite supply of magic hats from which we could pull rabbits at the last moment.  Theirs was a world that was about promises and at least the pretense of certainty.  Ours was one of sleep deprivation and raw panic.

I used to say that the salespeople sold our customers technology that we hadn't yet imagined we didn't know how to realize.

I liked to compare much of what we were doing to when Grumman built the Apollo Lunar Module. Nobody had built a lunar lander before.  Which meant that there was no known right way to do it.  Nor was there any meaningful way to estimate how long it would take or how much it would cost -- or indeed if it was at all possible.

In particular I got a facefull of this when we indexed one of our very first indices and performed searches against it.  The results were atrocious.  The results were so full of duplicates that for some searches you would get just pages and pages of identical results.  This was a wednesday.  I was asked to look into the problem.  The initial guess was that we should have a workable solution within a week. On sunday I had to call my boss and say "listen, the duplication problem is much harder than we had feared -- it'll take a bit longer".  Turns out there are several classes of duplication on the web.  And sorting them out is properly hard.

Note that it wasn't just about being able to do it -- it was about being able to do it in a practical manner. And most published efforts up to that point assumed small document sets and lots of time within which to do the deduplication.   Which meant that a "solution" that has quadratic (or worse) complexity just doesn't work.

We eventually came up with a battery of solutions.  All of them much faster than the feared quadratic solutions.  Many of them completely novel.

However, in retrospect, the interesting bit is the dynamic where people who like to deal with certainty and hard promises need to deal with uncertainty and ignorance.  For most complex technology projects, delays, disappointments and setbacks are inevitable because there are huge unknowns.  And more so for the projects that are really worth doing.

On one hand, hiding uncertainty behind a wall is dangerous because it makes our approach to hard problems fragile.  On the other hand, the pressure it creates can be valuable in spurring on discovery and progress -- by assuming "you will be able to figure this out".

2014-05-02

Symmetry.

In the 1950s the United States automotive industry was deeply inspired by aeronautical progress.  In particular rockets and spaceships.  Designers would adorn their cars with all manner of wings and protrusions that looked like rocket nozzles, flames and whatnot.  Of course, with the exception of air conditioning and electrical windows, the americans didn't really innovate all that much in the automotive realm.  An american car from the golden age of automotive extravaganza has no technology more advanced than what you will find on an 16'th century ox cart.  It was mostly about portion size, and not so much about flavor.

Fast forward 60 years and most american cars still can't corner, can't brake, have interiors made of creaky, hard, cheap plastic and exteriors that look like they were fitted by someone in the middle of kicking a lifelong habit of injecting refined opiates.  The engines are still mostly great big lumps of untapped potential with a side-order of disappointment.

Worse yet:  since nobody in the United States appears to know the physics behind springs and dampers, and thus all american cars suffer from unhelpful suspension tuning, consumers have adapted their taste. They prefer cars with defective suspensions.  So when european cars are sold in the US, their suspension setups are deliberately compromised to appeal to land-yachting enthusiasts.  And I wouldn't put it past them to fiddle with the brakes, rendering them suitably ineffective so as not to startle the american consumer(*).

But I digress.

What I find entertaining is that the 1950s were all about mechanically suspect cars whose looks were inspired by rockets -- whereas the 2010s is about the car industry (now getting most of their aesthetic inspiration from used bars of soap) catching up to a guy who started out building rockets and then figured he'd do cars while at it.  Mercifully sparing us of any attempts at letting his rocketry day-job be reflected in the exterior design of the car.

This has a neat sort of symmetry to it, I think.

(*) Top tip: when land-yachting in the colonies: do test the brakes before driving faster than brisk walking pace.  The car weighs about the same as a family of hippos and it has the stopping power of said hippo family on ice.

2014-03-06

Expansion

Back when I was in school there was this class we had to take called "society, environment and corporation" or something along those lines.  Quite obviously this class was a clever prank played on the rest of the departments by the industrial economics department in which they made a good effort to round up all of their most un-gifted morons, cobble together a pathetic syllabus and then proceed to waste everyone's time for an entire semester actually teaching it.

Initially my group did take it seriously.  We labored long hours and handed in our paper on time and on spec.  The week after we were called into the TA's office where we were told that our paper barely achieved a passing grade.  I can't remember precisely what made me suspect that the teaching assistant wasn't the sharpest tool in the shed, but I remember that from that point on, nobody in my group took him, or the class, very seriously.

I distinctly remember my friend Finn Arne having that vague smile and tone of voice that in later years was mostly reserved for dealing with telemarketers.

The subsequent works that were handed in were all produced in the most slipshod manner imaginable: we'd skim the problem text, pick out terms with elevated TF-IDF, find the paragraphs in the textbook where these terms had similarly high TF-IDF and simply rewrite these paragraphs.  Without regard for what the problem really was.

The product was then "bullshit-expanded".

Bullshit expansion is the subtle art of expressing something with more words, yet without adding more information to it.  This is not as easy as it sounds.  At the peak of my ability I could turn a single sentence into two solid paragraphs of text.  And before you poo-poo this, people like John LeCarré have built entire careers on their ability to pump out pages of featureless fluff.

This was necessary because the halfwits who taught the class thought that a given problem solution required a certain number of words.

I'm not sure if we were called into the TA's office after handing in the second paper, but not only did we pass, the TA was impressed with us taking his class so seriously, or something to that effect.

In fact, the only negative remark we got the rest of the semester was when we noted that we had made use of IRC to communicate while solving one of the problem sets.  According to the TA "no serious business would ever use such an unprofessional mode of communication in the future".  They were truly educating the visionaries of the future here.

But I digress.

The reason I mention this is because I think Rachel Maddow has now innovated in this field.  The field of bullshit expansion.  Although her method does differ a bit.  Her technique is to take the entire body of what is known in the Chris Christie bridge scandal, add a near homeopathic amount of new information to this body of knowledge and then excrete another permutation of the discovered facts with shrill enthusiasm.

Yes, this news story is vitally important.  Which is why it would be nice if it took more than the merest speck of a comma to reiterate the entire bloody scandal with a high-pitched, shouty torrent of words. It is killing the story.