Phil's blog

As so many of great influence, it suffered a troubled birth. Will it - won't it, we all wondered. All of us who care any about voice recognition. Will Apple throw a tiz, or are they just milking it. (the latter of course).

Is it any good? Yes.

Is it perfect? No.

Is it world changing? Well...

1. Crowd sourcing is exactly what speech reognition has been needing for sometime. If Google is doing what it normally does, and making good use of the user metrics to fine tune the service, their recognition will be as good as human in 2 years or less. 
2. Chat bots are not so much fun to talk to. Reason? Typeing... Not talkling.
3. Detecting affirming or otherwise tones in a voice is simple. Chat bots will soon be winning the Loebner Prize

That will change the world.

As I watch Terminator: The Sarah Connor Chronicles (what?), mirroring DARPA's lunacy, I wonder when the 'resistance' will start emerging...

I'm serious - sort of.

There are life-like, swarming, self repairing, self replicating, self governing, 1 million rounds per minute, killer robots out there and some that can pass for human to more people than most politicians. If you said that 20 years ago, people would have said you were mad, because it would / could not ever really happen. Now they'd say you're mad for bothering to worry about it, and stop watching too much YouTube. Right?

Well it won't be long before we see more stories of what happens when they do what they do best, and make Turkey Twizzler meat out of some unsuspecting humans.

Even I think Kurzweil is a little bonkers (in a great way), but he does raise some interesting points about where tech is going. And fast.  

So when do the humans figure out that they should be building robo-cidal androids? Otherwise, if they can kill, and learn faster than we, how do we stop them when they go wrong? Or a little too right (depending on which side you're on at the time)?

We divide and conquer, numb nuts. Its what we always do.

But of course will we even have too? Complexity theory suggests that such behavior would simply emerge. Making us mere onlookers at the grizzly mecha battle of the fittest. Ironically, as once we removed all the beasts of the earth that threatened us, we go build our own, more fearsome monsters.

What next?...

I often wonder what the difference is between noise, and creativity. I would argue the two are intrinsically linked. Chaos theory suggests that creativity is essentially an effect of noise. And experience suggests vice versa. Its hard to know how one could exist without the other. Symbiote constructs, if you will.

A theory I've long pondered and played with is that noise cancellation, a huge part of our brain's daily business, is a  large part of our mental creative machinery.

You filter out an enormous amount of information from what you observe, to acquire that which you perceive. I'm not wittering about LSD and the doors of perception. I'm talking about the quantifiable (but not by me) squizillions of petta bytes of Peta bytes of data your nerve endings observe each day, leading to the relatively tiny volumes of data you are mentally aware of though out your waking hours.

 But we know that a lot of it gets stored in our subconscious mind. Stuff that's interesting, but not required for immediate processing. I personally believe that most of out inventions, ideas, and intuition comes from these fuzzy data sets. The stuff that seemed interesting, seems recognisable, seems like it should be in a category, but didn't make it into the 'reality' categories.

Now if you're still with me, what is the effect of digital noise on are creativity? We hear and see petazillions of digital artifacts, poluting our observations of reality.

Pixelation, lowfi, compression data distortion and loss, small colour pallets, poor anti-aliasing, jitter... the list goes on. We are being bombarded by noise of a very different nature. But our subconscious is likely stacking it all away."Interesting, but not reality" it reads on the mental data tables' ID.

So will this shape the inventions of our near future, just as "organic" designs have throughout human history. Just as we've emulated the interesting noise of nature, to solve previously insurmountable problems, will we start doing the same with anomalies in the e-chaos, of our modern living environments. 

This is almost certainly already happening. How much will it increase? Where will it lead us?

How many times a day do you barely notice a CGI or a digitally enhanced photo, or a heavily compressed movie? But if you want to, you can tell which are and which aren't right? Most of the time. Which means your mind is storing the patterns in the digital static.

So when we dream, those patterns are floating in our subconscious, waiting to be played with. Manipulated, considered, envisaged in alternative scenarios.

Wheather androids will or not, we are already dreaming of electric sheep.

The great and mighty interweb will continue to evolve come-what-may, by all accounts, and one of its rising stars in its current gene pool of technology, is, like it or not, javascript...

Problem: this whole mishmash is not threaded like we're used to. Programming is becoming non linear. No real solution found.

Meanwhile parallelism is surging in the offline as well as the online thechnoverses.

Problem: threading parallelism is a big pain in the arse. e.g. Efficiently scheduling Map-Reduce systems is far more difficult in many situations, than anyone realy wants to have to deal with... Programming is becoming non linear. No real solution found.

And over in the world of accessibility screen readers are trying to deal with the fact that while javascript may be unthreaded, Ajax wouldn't even know what a thread was if one came and sat in its chair. Programming is becoming non linear. The solution is politeness.

 Noted for a long time  to be a virtue, politeness is becoming an important part of assistive tech's arsnel for dealing with the more dynamic, structurally unpredictable, and chronologically scatological applications, sites and services the web is currently seithing with.

ARIA's labeling is much needed, but not exactly a work of genius. People ain't using elements for the purpose they were designed, got to relabel them then. Not a huge leap.

 But screwing with causality, as the Assistive Technology perceives it, is a much trickier problem, and social graces appear to be an admirable solution.

ARIA's standard dictates that events, such as content being dynamically altered, without or out of synch with, a user's interaction, should be announced IF appropriate. Assessment of appropriateness involves looking at  the politeness weighting (attribute) of the region in question.

So can't we learn from this in other areas having similar problems with asynchronicity? Scheduling of parallel processes, multi core scheduling, AJAX logic flow design, ...? Couldn't we all do with watching our manners when schedualling processes?

Just a thought.

So that was good!

I missed a lot of it :( And the internet connection diing on me just as I was bout to demo Masher Nations was frustrating. But a good turn out and a some great talks.

Always fun to see Christian on stage.

Learnt about SATOGO.com a great intaller for a useful tool, and free! Nice.

Finally met AC and Heni  !

Great to see you. And Jonathan, DotJay, Jim, Ian, Lucy,...

Some seriously interesting thougts being had, centreing around the notion of an OpenID keyed data store of JS mods. Either pulled off appengine from head script links in a page or via mashernations proxied pages. So making an API for Masher Nations that allows this might be the next port of call...

So now MasherNations = Python based proxie service + JS API + JSON + Dynodes + GData + Crowd sourced js mods

Or at least it will...

Nice.