Excellent post! A couple of examples for some very odd and potentially ruinous experiences which I believe are fairly common.

• An insured person who is admitted to a hospital that is in-network may be treated by a doctor who is out-of-network and then faces enormous bills. Such a person thinks that he or she has insurance and is in an in-network hospital.
• A person with traveling insurance is faced with a huge bill from the hospital and when the hospital was told that the insurance is from a different country and the bill is very high to pay out of pocket, the hospital offers a 75% "discount". So the initial bill was hugely inflated.

You have to compare it to what existed before, and when you do that there is a significant progress. The number of image-captions used and the way it is trained with loose supervision shows significant progress. I work in the field and as far as I know, there is no other "technique" that comes close - and it is not for want of trying. In my personal experience, there are academics and researchers who would love to come up with some other technique to beat DNNs and if they do, that is progress as well.

IMO, this is a step in that direction (i.e. universal, although I wouldn't call it AI). Progress is often incremental. Also, it is not programmers and white collar workers who will be the first casualties. From where I am, we are still far from the singularity (however you define it).

I do think that AI in media is portrayed to be more advanced than it is - I'm not sure if this is a problem with the media or the field. I mean, this is a published paper and you can read it to understand the claims and problems. The presented work is state of the art and is a big step forward, but there is a gap between that and deploying it in applications. In the end, this is *just* a neural network trained to map input images and text into a vector. However, the CLIP model is an impressive step forward in terms of how it is trained: with hundreds of millions of image-caption pairs and very little supervision. The P in CLIP stands for "Pre-training". Is this perfect? No. Does it have problems? Yes. You don't have to speculate - the authors talk about it in the paper. But to ignore what it can do and focus on a bug (which actually may be a feature) is lame. To elaborate on that, the model is doing what it has been trained to do: i.e. match the image and the caption - reading text in an image (without being explicitly trained to do so) is cool in itself. However, the correct label often scores high as well.

I am not a fan of Hillary, but pretending that she is as bad as Donald Trump is just insane. Bernie Sanders is trying to influence the party platform and trying to prevent the disaster that would be Trump. I don't think that is unprincipled. If you think that is, the rest of the house and senate are crooked as hell.

The US system doesn't have the possibility of coalition governments and you can't really do a protest vote unless you want to risk Trump. Maybe one decides to bide the time until the next election or move to a different front in the senate / house races.

This post is insightful?

Firstly, where is your evidencce that mathematicians don't wear shoes, bathe or keep a tidy house?

Secondly, the quote in question says that "Anyone who cannot cope with mathematics is not fully human. At best he is a tolerable subhuman who has learned to wear shoes, bathe, and not make messes in the house." Not that a person who can cope with mathematics needs to wear shoes, bathe, and not make messes in the house.

Give India a break, it *is* a democracy and it is only 60 years since it has gained independence from Britain which left it in not-the-greatest shape. It is true that there are problems of the lower castes, etc, but something is being done, albeit slowly, in terms of reservation in education and government jobs for the lower-castes. Why it was only 50 years ago that afro-americans in the US were second-class citizens.

iFad - 'nuff said :)

This one has ratings and comments enabled. (http://www.youtube.com/watch?v=9oWWt_L-qeo&feature=related)

I'm not an APPLE fanboy though I own an iPhone. While Apple is to blame for giving in to AT&T's alleged pressure, I don't see how it benefits Apple. Can you elaborate?

Well I don't like Microsoft not because they make bad software, but because they make it difficult for my OS of choice (Linux) to function well and compete on an even footing with their anti-competitive practices, sabotaging of truly open standards, etc. If they just make bad software, I don't have to use it. Unfortunately, if I need to edit an MS Office document cleanly I need MS Windows.

I read somewhere that it meant "I can't configure Debian". :)

I think there is much confusion about the term agnostic. Since one cannot disprove the non-existence of any god, agnosticism is the most logical stance one can take. However, the fact that one is agnostic should mean that one is equally agnostic about Zeus, the christian or any other mainstream god, and (not to ignore the pastafarians) the FSM as there is little evidence to support any of them. I am, strictly speaking, agnostic, but for practical purposes, I'm an atheist.

A comment with the heading "I find most Indians incompetent" is "interesting"? It doesn't matter if you replace "Indians" with any other country, it is a patently stupid generalization. I am an Indian and I didn't go to IIT, but I can tell you that most of the guys who get into IIT every year are some of the smartest 2000 guys in a country of 1 billion people. True, some of them have an attitude but they are rather raw because they are undergraduates, but they are definitely very talented.

Really? I hold a Phd in Electrical Engineering from Maryland and I have several good friends who got their PhD's in the US (all of them had their tuition paid and stipends) and are now in India. Their average age is around 30 as well. Some of them worked here for a couple of years. I still think that the US attracts a lot of students who want to excel in engineering and science. Getting a suitable job in India for these highly qualified people is difficult because of the competition and the scarcity of good jobs, but that is fast changing.

They should just call themselves "Association for Dealing with Competitive Technology."