Wednesday, 29 January 2014

Some thoughts on Google selling Motorola to Lenovo: from a red ocean to blue oceans?

Google is selling Motorola. 

My first reaction of the news is shock and dismay. I really like new concepts of Motorola like its 8X platform with separate always-on processors for voice recognition and motion sensing. Not to mention, its revolutionary Project Ara.  
But it could be the best thing happens for Motorola since it, currently, is handicapped by Google itself. Motorola has more barriers to contact Android group than other companies like Samsung. Also, right now, Google is also in conflict of interest since the success of Motorola would piss off its current partners like Samsung, LG and ASUS. On the other hand, a loss making Motorola would disappoint investors. This could be the best deal for Google, stopping partners to turn to rivals like Microsoft. Who knows? This deal also could be a result of negotiation with Samsung for recent patent cross licensing deal. 
The only question is how could Google think Motorola worth less than NEST? 
Reading Larry Page's post, it seems Google is swapping industries , from smartphones to "wearable and home market".  Giving up Today's technology for tomorrow's, getting rid off saturated  Red ocean for  Blue oceans. That is really good news. And also, Google is keeping Motorola's Advanced technology group including Project Ara, making sure that the next revolution in smartphone is in its own hands. But who needs smartphone if they could improve Google Glass into a real augmented reality glass like Meta space glasses.
Well done Google.

Sunday, 19 January 2014

The current state of AI and posthumanism in popular culture Part 1: the review of the novel "The Fear Index" and the movie "Her"

It is the first month of the year 2014, but I regret to inform you that public perception of Artificial Intelligence(AI) and Transhumanism is yet to be improved. The general public is still mislead by artists with no theoretical understanding in the foundations of biological life and intelligence or their machine counterparts. I agree that the conversation and contemplation of AI's potential and consequences is imperative and I am thankful for creative concepts they bring to the table.
This article was started as a review of the science fictions I recently experienced "Her" and "The Fear Index" , but it took shape as how I wanted those stories to be.



The Fear Index

Short: A very good thriller with an intriguing concept. A bit let down by plot holes and an attempt toward the script of a Hollywood movie. All in all, the good overweight the bad points. Well worth reading.

Long: [spoilers alert]
  The best part I love about the novel is its concept of AI. The worst part? It is also the AI. 
Allow me to explain. I love the concept of  a self learning and propagating machine intelligence. But the real handicap of the story is the author's obvious lack of understanding of AI and underestimating its capabilities.
  Robert Harris is a proven mystery thriller writer with his previous work on "The Ghost". The Fear Index is certainly a gripping thriller you won't want to put down.
  The novel follows Alexander Hoffmann, a former Physicist of CERN who is working as the brain behind a hedge fund. He developed an AI algorithm of trading, which he termed VIXAL. Now, Algorithmic trading is nothing special. It has been applied so widely that most of the stock trading nowadays are done by algorithms. The algorithms trades based on its prediction of the stock values considering historical performance and current trends. In AI terms, it is pattern recognition and optimization problem. 
The book title and its AI got the name from VIX, stock market volatility index which is colloquially referred to as fear index. Hoffmann's algorithm,VIXAL, feeds on volatility or fear in the market. The book has excellent explanation of hedge trading. 
The idea is not novel,  it is a usual AI run amok story of science fictions that has become a cliché. But the beauty is in detail.
[Major spoilers ahead]

   The AI, VIXAL, has knowledge of the entire web, which it uses to predict certain events that will rattle financial world. Use this advantage to hedge tens of billions of dollars. At the same time, VIXAL is also secretly building and extending itself throughout the world, unknowingly to its creator. It was done through contractors via email and financial accounts of its creator. It is classic identity fraud but by someone who knows you just as yourself , someone through whom you usually did transaction, your computer. Consider your smartphone or computer is trying to assume your identity for its benefit. In the case of VIXAL, it is to make itself more powerful by building data processing centres around the world. That is the concept I love. Not only that, VIXAL ask the security contractor of the company to install hidden camera and microphone in all the rooms of the office and home so that it could monitor the human world. 
  Though highly unlikely, I would like to see Google's deep learning AI or its search algorithm coming alive and starts building itself. 
There is a part I could certainly agree with Alex Hoffmann on AI. In his own word,

"Silly people, encouraged by journalists, thought the aim [of AI] was to replicate the human mind, and to produce a digitalised version of ourselves. But really, why would one bother to imitate anything so vulnerable and unreliable, or with such built-in obsolescence...?"

   I love it for the different purpose than the author's. For the singularitarians, the purpose of AI is not to imitate of human mind, but create a more perfect intelligence to compliment our puny brain.  Many science fictions (such as I, robot and Doctor Who) try to paint lack of emotions as a weakness of AI. But really, you can have Emotional Intelligence (EI) (understanding others' emotions) without having to feel any emotions. I am not a fan of giving AI human intelligence or emotions but of making humans logical and rational devoid of primal urges. From Evolutionary Psychology point of view, emotions (such as hunger, fear, love) are primitive positive and negative feedback for the organism as a motivation for a certain action. AI or robots will certainly have distinctive but analogous feedbacks. But will their purpose and actions will be like ours? This leads us to another part of the novel. 

   Robert Harris makes a compelling techno-thriller. However, there are some inconstancies and incomplete information of AI. Here are things that can be improved. 

   Firstly, the novel failed to make sense the actions of AI. Out of control trading is completely rational as it is its main directive. So is  taking control of itself and expending itself stealthily as it would be to any living beings. But why does VIXAL need to kill its inventor with elaborated operations or is the AI trying to fulfil his wishes? It was never make clear. The author's lack of understanding AI is obvious in first few chapter when he wrongly describes natural language processing problem as speech recognition.
In the novel, this super intelligent VIXAL send email with mostly numbers which author probably thought is machine-like. But even simple AI could do much more than sending account number to transfer. There have been a lot of simple AI that has already outwitted basic form Turing's Test. [A machine is said to pass Turing's test, if human operator can't differentiate machine's response from a human]
    Second, I am really tired of artists renaming conventionally accepted technology names. In the novel, the author call Artificial Intelligence as Automated Machine Reasoning(AMR). This trend, which you can also see in Transcendence movie (renaming the Singularity as transcendence), is becoming annoying. 
  Third, Professor Hoffmann was described as pure academic nerd who once had depression. But on the most of part of novel, he looks more of a rich obnoxious banker than an academic. Especially, his actions in every part of novel(when he tries to catch the intruder alone, when he went to meet the killer), he is more like an impulse-driven testosterone filled person than a rational academic. Especially, the joke is when the professor decided that the best way to shut down AI's computer cluster in the warehouse is to burn the building, instead of trying to turn off electricity  or cutting fiber optic cables. 
All these gone back to author's attempt to force the thriller. The novel would be great without those action sequences but it is not designed as literary piece but a Hollywood movie. The movie is already in development for 2014 release. I still love the novel and it gripped me until the end even though I have figured out the culprit in less than 80 pages. That is a really good novel.


Part 2 will be about the movie "Her".
Update: Well, I don't need to do that any more. Since Futurist Ray Kurzweil did an excellent review here. http://www.kurzweilai.net/a-review-of-her-by-ray-kurzweil

Just an addition, if Samantha has such high intelligence why don't she engineered a biological body in the shape of Scarlett Johansson or a robotic body with such proportion that would not be possible to achieve biologically, like the proportion of Barbie dolls. That would have solved human surrogate episode entirely. Advanced AI like Samantha would not be necessary to leave human or feel human sadness. She could well understand human emotion better than human themselves and empathise it. But she doesn't need to feel that herself.  She could handle that whole human love and relationship thing with like one percent of her processing power.