thallukrish

It shines new light on the mindset that promises bullet trains and ‘smart cities’ while neglecting public health and malnutrition; the pursuit of heartless policies calculated to keep agriculture economically unviable, driving the rural poor "wholesale into the cities to serve as cheap fodder for India’s ‘economic miracle’; and the permanent destruction of million-year-old mountains, forests and rivers to create a ‘nine-day’ industrial wonder."

I cannot agree with this more.

We need more middle class people who are educated to come into political stream to counter the existing political class which caters to the rich and the poor. Then things my change for the better.

Human race has evolved from that sub-atomic particle of star dust since eons and reached this point where we are typing this comment. Obviously this cannot happen without the necessary checks and balances throughout history and it will be this way going into the future. So no need to spell doom, though the cautious gene in us still keeps vigil.

Guess the key lies in grasping the meta data in images even if they are less somehow. May be this will come by clustering similar things. Like my brain may put a cat close to a dog than a human as they have something in common. But between a cat and a dog, I find some metadata that are dissimilar.

This is what I wrote years ago in this blog http://productionjava.blogspot.sg/2014/07/the-broken-web.htm... and I have been working since then to fix it.

I guess the key is our ability to extrapolate, imagine things from a single picture. When a child is shown a proper image of a cat and it sees a cartoon cat, it has done that extrapolation of cat's body contours. Or rather I would say it is some sort of meta data that is learnt out of each experience like the way we model in OO - class and object instances. We somehow are able to abstract the class out of a image even if it is just a single image and I feel it is the meta data that gets refined over time rather than the storing pixels of actual images.

because every data model is different, every process flow is different, every UI is different.

In short, it is better to identify and drop things that would hang in your mind from actions you did. Sleep is the natural way to get rid of hangings. If it is worse and you cannot sleep, then meditation is a good way, though not a easy way.

To err is human. We can punish erring drivers by law. But we cannot punish a robot. Unless there is law that applies to the owner of the autonomous vehicle companies on erring automatons unleashed by them :-) and they are implementable.

The issue with giving advise on medical information is there is no one "right" way. It is the same with a lot of domains like education or finance. In a way it is better for a search engine to put out information in a more meaningfully related way for a user to easily find stuff and be able to get a broader / deeper view of a subject by himself / herself.

Self-learning is brought about by Ego. Reward / penalty is self determined or determined by environment. Without self-learning capability which lets humans to learn on their own with the limitations imposed by environment, by their Ego, by their body, which is driven by survival, whatever the AI learns has to be taught in someway by training, by reward/penalty imposed by us. Then it will be always a poor limited imitation of humans. Not a intelligent being.

Humans have the drive to go about because of a body that has various demands. And with every doing, humans accumulate Ego which again helps their drive to learn, explore, do stuff or define their behaviour. Humans are self-learning systems. The rewards and penalty are also self-created.

Unless general AI systems start to model the above which creates self-learning, they will be just specific systems with a tiny aspect of human world ingrained.

Being able to "Infer" from what it learns and "Apply" it to new scenarios in a general way is all about intelligence. I do not see how making it to win one game or one million will move it towards achieving general intelligence of this sort.

I feel to make programming easier, we need a assistant (intelligent in whatever sense you call it :-) that can figure out the 'intent' of the code by reading it and be able to point out bugs by learning from millions of source code in Github.

I wrote in this article

https://medium.com/@thallukrish/debug-assistant-2aa37ac13ca5...

Dont try to answer this question when it comes up. I have spent a lot of time personally, developed software (prototypes) and figured out that this is a holy grail.

Looks detailed. But if you are someone doing an App all by yourself it looks scary. At the end of the day, a lot depends on if you have built something that the users cannot refuse I guess.

Being general or specific is just a thought process about an idea. You can go from one end to another and vice-versa while ideating and even when you build stuff. You can assume the whole world is going to use your product. But when you try to define that first set of faces who may use, everything starts to become specific and it questions the overtly general assumptions you made.

There cannot be one path to software development craftsmanship. It is a large space and there are several paths. It is continuous evolution and experience. You need to be a practitioner all your life. You have to be learning in the process of solving a problem and not learning for the sake of it.

I feel if you have something better to offer, say a better way to look at information, not just a better page rank algo, but a different way to look at information that is scattered across, one that enables a better insight into existing web pages and less overhead of constant 'searching' by typing a query, something that automatically helps to "see" relevant information with less effort, then it can be an alternative to what a Google search does today and can position itself as a competitor in the long run even if it does not exist today.

Unless there is a compelling reason I never felt comfortable using frameworks. They tend to make a small problem much bigger. You end up spending time understanding, fixing and adjusting to the framework than making that little adjustment to your code to extend the functionality or relook at your needs.

I guess initially when I test it, I need it everyday for few hours at least when it will have bursts. Storage per se may be fine with 120 or 240 GB as I deal with text only and not videos/images in phase 1.

Thanks. Will check out this paper.

While in general I felt something similar that it has to be not just reams of data but also some sort of model (meta data) that when combined can produce innumerable combinations more easily.

If I need to show 1 million cat images to train a neural net to see cats, I wonder how a human brain can figure out cats of any kind just by seeing one or two.

Is there something fundamental we are missing in going about building these deep learning stuff ?

I bought a MBP in 2012. The motherboard went bad after a year. And after a year the cable to the hard disk had to be replaced. Also I feel the upgrades we do on OSX causes the MBP to slow down. It is sleek, elegant, battery life is great (actually I need to replace it now, its dying), great design etc. But I feel the more perfect it is designed to be, the more vulnerable it is.

I say this when I compare this to my HP pavilion sleekbook which did not have any trouble so far in the last 3 years.

I have seen the following happen to me while reading books.

Event driven - you go back to a book based on thinking around a problem and look for information around it. I have found this very effective to understand the content, especially something not so trivial and those that you may not get by passive reading. However, this requires more time and effort from you in terms of self-thinking around problems.

The second way is passive reading - just read because you got interested in the topic. You don't understand a lot of it immediately, and so either your interest may carry on and you may get more than where you began or you may give up.

But if you give up, in the interim you may encounter some things related to what you read and you go into the first mode. Or you find that what you started is not your cup of tea at least for the moment. You iterate over other topics of interest.

More than programming, often a lot of time goes in discussions and meetings. I feel the best thing that can happen is, if you hire a programmer who does end-to-end and releases a product all by himself/herself. This whole job of parallelizing, scheduling and managing looks so artificial and thrust upon unnecessarily just because we were not able to slice and dice the problems in a way they can be addressed completely individually.

A human child learns much more easily by seeing only a handful of images of a cat and then almost being able to say any type of cat image as it grows (without ever seeing 1 million or billion images). So, there seem to be something that shows that more than the amount of data, the "reality" of seeing a real cat probably includes all possible aspects of a Cat ? There seem to be something missing with this whole deep learning stuff and the way it is trying to simulate the human cognition.

Nice, you drove home the point of simplicity in building software which largely occurs over time in the hands of a good software engineer. Do more with less code.

Coincidentally, I had similar views in this piece some time back.. https://medium.com/@thallukrish/on-coding-e3c03771a369#.bb5k...

I feel Email has evolved organically since eons and what we have today is like the way humans have evolved over millions of years. While the standard look and feel remains the same pretty much, it has become one's identity for receiving messages on the web of all forms and types, of all needs from commerce to friends to business.

It is also a database of all our past communications. Apart that mobile adds further complexity. It is likely that any one answer for it can miss out things in the enthusiasm to bring out a simpler version or in the over-enthusiasm to make it complicated than what it should.

It is hard to quantify the "needs" as it caters to a wide audience as a one stop answer. Slack like app can take away the professional side, or a Facebook can take away the social side, but no matter what, you still need a email as it serves as the identity to reach out for the Web applications.

So whoever is building a new email App needs to keep all of this in mind and come out with something nifty yet powerful.

Arriving at the most meaningful words eliminating the noise is the hardest when you try to get these working for some real world scenarios.