Tag: research

  • A horse designed by a European Committee

    I take offence to the expression ‘a camel is a horse designed by a committee’.  Camels are incredible animals, possibly ideally suited to the desert.   They have been instrumental in thriving civilizations and commercial breakthroughs which changed mankind forever.  A camel is in fact an animal no European committee could ever imagine.  Someone would argue it doesn’t look right, another would object to it’s saliva based on some european health guideline or other, for sure we would bicker for years about whether or not we can eat its meat; and with more than 100% certainty, the number of humps would become the bone of contention between heads of State for decades.   In total, we would probably spend millions of euro in meetings, public hearings, research and other pleasantries and end up with …a horse.

    An excellent recent example is that completely stupid button you have to “accept” when visiting websites.  It is of course completely useless.  A bit like making a sign reading “attention! If you get on this camel, a lot of people will see you because you will be higher up than before“.  A paper sign.  Which camel owners will have to put on the camel everytime a “new” rider comes along.

    Perhaps the best illustration of the futility of approaching technology is carrier neutrality.  To put it simply, this is the notion that Facebook has to “deal” with hate talk or sexism on its platform.  Or that it is Twitter’s “fault” that some people spread false rumours via tweets or bots.  In practical terms, this is like asking the telephone company to interfere if two or three of us start talking about building a bomb one day on our phones.  Completely and utterly ludicrous.  In fact, we could make telephone companies completely ban profanity on the telephone.  All calls could go through voice recognition systems and when a swear word was recognised it could cut off the line or send you a fine.

    There are two reasons we don’t do this and both are interesting.  The first is that despite spending billions on automatic translation research, Europe still lags far behind in terms of real time machine translation.  Things you can do for free on any Android phone, simply can’t be handled by any European infrastructure in technical terms to handle the task.  Much like no French company could serve videos as well as YouTube.  The second reason is of course that we could never all agree on what constitutes “profanity”.  A French man’s “merd” is not exactly the same as a British “oh, poo!” or even a German “scheisse!”

    In the same time Europeans would take to “initiate a working group to deliberate the need for a committee to address the issue”, Google staff would have solved it.  In fact they did.  Not for ethical reasons, but for commercial reasons, YouTube made video channels with profanity inelligible to take adverts.  Problem more or less solved.    Air BnB had a similar problem with users of their platform who refused tenants based on race or ethnicity.  This is no easy problem to solve.  It is virtually impossible to find a solution talking about it around a table.  Air BnB didn’t “initiate research”.  They tried, tested, improved and made it work.

    Essentially the problem is one of friction.  Technological networks operate on the premise that less friction is better.  You want your phone to serve instantly.  Search results at a the blink of an eye.  What Silicon Valley does when presented with a challenge is usually to actuall add friction.  That ludicrous european website button informing us about cookies is in essence an added step.  It is meant to ensure we all understand cookies.  Except it doesn’t.  Compare it to Facebook trying to teach us about privacy.  They constantly change the way messages pop up, the content of the messages, the way they try to make sure we are all on the same page concerning who sees what when we post on their platform.  Other platforms have online mini lessons about hate speech.  You start to post something and it pops up saying “hey!  Do you know that this word you are using is considered negative in some parts of the world?  Would you rather use one of the following suggestions:……”  They even give mini history lessons relating to words or uses of words to help make sure you say what you want in a way which will actually get the message across.

    Europe will never, ever catch up with Silicon Valley like this.  Artificial intelligence is not about installing a “kill” button.  I read through the blurb and it is a bit like bad French or Italian academic literature.  Too much theory and mostly outdated.  Impressive for headlines, useless in practice.  In this particular phase of technological development we need to be building infrastructure and platforms.   We need millions of experiments and we need to learn much faster.

    Sir Alec Issigonis, designer of the legendary Mini, is often credited with coining the expression about horses, camels and committees.  The question is whether today he would be enticed by a cushy university job, doing European research and enjoying European committees and funding, rather than building the iconic Mini car.  I think he would prefer to work for Google and just get things done.

  • Study Of ‘Ouzo Effect’ May Lead To Design Of Improved Drugs, Cosmetics

    Scientists studying the cloudy emulsions produced by anise-flavored liquors such as Ouzo have discovered new molecular insights into their formation, findings that could lead to the design of better commercial emulsions used in making pharmaceuticals, food products, cosmetics and other materials.

    Although transparent when bottled, Ouzo, Pastis, Pernod, and other popular anise-flavored alcoholic beverages form milky-white emulsions when diluted with water prior to drinking, a phenomenon commonly known as the “Ouzo effect.” These emulsions occur spontaneously and are stable for weeks and even months, a feature that is attractive to industry. However, scientists are unclear how these mixtures form and stabilize.

    In the new study, Erik van der Linden and colleagues measured the stability of various emulsions prepared from commercial Pernod and compared the results to theoretical predictions of their formation. The scientists found that their experimental observations were often opposite the predicted behavior of the emulsions in the presence of various concentrations of oil, water, and alcohol components.

    “More knowledge of the parameters that determine the stability of these emulsions, besides interfacial tension, solubility, and density difference, might lead to better control of theemulsification process,” the study states.

    Note that when adding water and ice to ouzo, it is customary to add the water first.

    Click here for the source of the original article.

     

  • Open sourcing privacy: my master plan

    The advances in neurology are fascinating right now. MRI scans no longer limit themselves to one brain.  It is the interaction of people which ups the ante.  How do my mood changes affect you?  If you don’t want to get bogged down in mirror neurons and spindle neurons and the detailed science of it all, I thoroughly recommend “Social Intelligence” by Daniel Goleman.    Snap judgements on whether or not we like someone or a product are well worth analyzing because the same principles are even more important online.

    I famously wrote that I would sacrifice one of my hands for access to the raw data of Facebook or Google.  Well, I just thought of a good way of keeping all my hands and still getting those invaluable insights!  While the media suddenly get excited about privacy online, projects like Diaspora are a good idea.  Yes, I don’t like the idea of giving away my personal information, my thoughts, photos or anything else to someone else.  Of course they will try and make money out of it!  But what if we could establish a research cause worthy of being a part in?

    Remember when everyone installed SETI screensavers to help crunch data from alien-looking observatories?   Or the global appeal to help process human genome analysis?  (The idea was we would help discover solutions to global diseases, turns out we just helped pharmaceuticals get richer!)  But we can use the same principle, that same altruistic sensibility to get people’s personal data.  Heck, we can beat Facebook at it’s own game!

    Here’s how it would work: a set of totally trustworthy institutions throughout the world, something like the United Nations, runs the show.  OK, we don’t really trust anyone and we all think that once data is digitized it can and will go anywhere, but we will have to settle for the best available trust levels.  Then we get widgets, could be in the browser, could be anywhere else on your computer or mobile phone, that monitor what we do.  Facebook, Twitter, email, whatever we feel comfortable sharing.  Here’s the catch: the data is whitewashed of our names and other personal details from the beginning.  I may choose my data to be shared as “a guy in Greece”.  In fact I, and many others I suspect, would be more willing to give really private information to such efforts, stuff I would never put online otherwise.

    Open privacy policy from the beginning because the whole point of this tool is to help you understand how much information you are giving away with everything you do.  (Yes of course Symantec or some other security company could attach it to their antivirus but it wouldn’t be the same, read on.)   And here is where I get my data at last:  universities or other researchers from private or public institutions can apply for access to your data.  They write up a proposal, what they will do, what they will look for and what insights they will give us.  Maybe they will give whoever gives their data up more detailed information to make it worth participating.  So in fact, I won’t even have to do the research, I will just install the browser plug in and choose whichever scientist makes nice proposals!  Then they will give me their findings to mull over.

    We will effectively be breaking the monopoly that large institutions like Google or Facebook have over user data this way.  Sounds hard to sell but simply getting the academic community involved would be a huge leg up; in fact they would sell it for me as they would all need the platform to get their research done.   We could even make sneaky Facebook apps for it!