FaceKicker

I guess, in the same sense that the implementation of any local binary executable is "all there" because you could theoretically figure out what all the machine code does. (But granted, I'm sure that reverse-engineering obfuscated JS is generally much easier than reverse-engineering a binary.)

I'm not sure you read past the first paragraph ... the author is on your side; Netflix isn't.

Right, you can get more out of the assignments if you try, but to me the purpose of assignments (versus passive learning - lectures, reading, etc.) is to force your brain to synthesize rather than just comprehend. The ideal assignment, then, is one that forces you to synthesize as many concepts it intends to teach as possible.

Just like you could go back and implement for yourself the skeleton code they handed you, you could also go out and implement everything in the lectures without any assignments at all. It's just that, like you said, the assignments provide a useful starting point. And I'm only saying they could be even more useful by requiring you to implement more of the complete pipeline.

The fact that an incredibly self-motivated person could learn everything there is to know about machine learning with the course as a starting point doesn't mean that it's bad to make the course more useful for a somewhat lazier or less interested person.

From the headline: 33 Percent Of Google Users Will Use Bing More After ‘Bing It On’ Challenge

From the article: 33% said they would use Bing more often after taking the Challenge

Two very different statements IMO - how many of those 33% are ever actually going to think about it again rather than just reverting to their old habits?

> My one complaint is that the programming assignments weren't interesting at all. The results were interesting, but the setups were mostly given to us, and we just had to code an algorithm that was in our notes. For someone who understands the basics of linear algebra and programming, it was just a syntax challenge, and that got irritating after a bit so I stopped doing them.

I agree with this. The programming assignments I've done so far in the Machine Learning class are usually 5-7 matlab functions, many of which are about 2 lines of code (the longer ones might be ~10 lines of code). If you've ever done matlab/octave programming the assignments will take about 20-30 minutes and be completely unenlightening as you're literally just translating mathematical notation into matlab (which is, by design, already a lot like mathematical notation anyway). They provide entirely way too much skeleton code to learn anything from if you're not actively trying to learn. If I weren't already mostly familiar with most of the material presented in the class, I imagine I would never retain knowledge of how the machine learning "pipeline" worked or have any high-level understanding of the algorithms, because the assignments just require you to implement the mathematical pieces of each step, without ever asking you to, for example, actually call any optimization routines, or put the pipeline together.

The problem, I think, is that it would just be too difficult to do automatic grading in a way that is reasonably possible to pass if they don't turn most of the work into skeleton code. Since the automatic grading needs nearly exactly matching results, one minor implementation difference in a perfectly good implementation of the algorithm itself (e.g., picking a single parameter incorrectly, picking the optimization termination conditions incorrectly, choosing a different train/dev split, etc.) would make the entire solution completely wrong.

Regardless of the effectiveness of spam filters (unless they are literally perfect - 100% precision and 100% recall), doing something like the "Building with Javascript" method seems strictly better than doing nothing based on the linked study - it reduces spam by 99.3% and adds no work for humans who want to email you. Except for humans who have JS disabled by default, I guess...

Right, it only lets you view up to 30 days at a time. I want to see it all at once, and I wouldn't even mind if it's just a messy pile of incomprehensible lines. Something like this fog of war idea would be even better, though.

This is neat, I wish Google Latitude had it. For that matter, I wish Google Latitude had any useful visual representation of the 2+ years' worth of location data I've been feeding it. =|

Seems exactly wrong to me - Google uses its free customers as beta testers, whereas its paying customers (Google Apps users) don't get to see the products until a year later.

If you're defining "free market" that strictly, then I don't see how your original post has any relevance. Shortages might not exist under your definition of "free market", but if that's the case then we clearly don't have a free market and properties of a free market are irrelevant to discussions of any real economy in existence.

I've had way too many experiences driving in a city when I've just needed to run inside somewhere for 2 minutes and really wished someone was in the car with me to drive around while I'm inside rather than spending 20 minutes to find a parking spot. An "orbit" feature would be pretty useful for situations like that, and would even save gas occasionally, considering how long you can spend driving around looking for parking.

(Too bad it'll probably be a while before robot cars can legally operate without a human in the driver's seat and paying attention, though)

I think the point is that Spotify is so convenient that you essentially own any music on it, especially if you have premium, potentially making it pointless to buy the album for many customers. Prior to Spotify and similar services, if you wanted to listen to music legally (which piracy doesn't give you) and conveniently/on-demand (which the radio doesn't give you), you had to buy the album.

(I'm assuming your point was that Spotify probably pays at least as much per listener as the radio.)

Is it possible that there's in some sense a unique "global optimum" of UI and/or aesthetic design (as opposed to just a bunch of equally good but very different designs), at least as far as is possible under the constraints of available technology and resources at any given time? I would guess probably not for aesthetics (too subjective), but it's reasonable IMO in at least some very minimal cases of UI design - e.g., the idea of using some kind of touchpad gesture to scroll on a laptop. If so, maybe that optimum should be considered a "public good" that no single entity should be allowed to monopolize (or, they could be allowed to monopolize it only for a certain period of time like in the pharmaceutical industry and/or forced to license it for a reasonable fee) - otherwise for all of eternity it's illegal for any other company to use the "best possible UI" in their competing product, and in the meantime the company that filed a patent for it could have stopped developing their product entirely.

Less abstractly, I think the author's point about the standardized "UI" for cars is a good one - society is better off, not worse off, as a result of companies copying each other in the auto industry. If you can drive one car you can drive them all (well...except for the automatic/manual thing). You don't need to spend three days relearning how to drive every time you rent a car. In the same way, society benefits a LOT from familiar UI design across different software companies - e.g., anyone who has ever used any mainstream desktop OS in the past 15 years or so can pretty quickly figure out the basics of any other mainstream desktop OS - they all have programs partitioned into "windows", have a "desktop" with "icons", etc. These UI elements all seem trivially obvious to us precisely because of copying, but there are certainly alternatives to all of them.

> free food

Am I being dumb or are you saying Facebook pays for their employees' groceries if they choose to work from home instead of working at the office and eating at the free cafe(s) there? I'm impressed if so (even though it probably wouldn't really add up to that much money for a company that can afford to pay $1B for Instagram).

I think I understand the reasoning behind open floor plans, but what's the point of having everyone in a single mega-room? It's not like the two people in opposite corners of the hangar will be able to communicate seamlessly just because they're in the same "room"; they'll have to walk over to one another just the same as they would if they had walls separating them.

(Of course being in the same room doesn't hurt in that respect either, but partitions have other advantages as well, e.g., aesthetics and noise.)

Assuming the article is completely accurate, the doctor's phrasing and mannerisms (e.g., shaking his head, emphasizing not only that the child is in the bottom 1% but that 99% are better than him, "seemingly perplexed"), don't seem consistent with your explanation. His comments and tone don't strike me at all as being encouraging or reassuring, he just sounds completely dumbfounded by the parents' situation, almost as if he pities them for having such awful lives.

> "You know cognitively, he's functioning at the bottom 1 percent of children his age," [our neuropsychologist] said. [...] "That means 99 percent of children are doing better than he is." [...] He waited, seemingly perplexed. "Having seen what I saw, and of course you have to be with your son all the time -- I have to ask you, how do you have the patience?"

I'm confused, can anyone offer any insight as to how a neuropsychologist would be genuinely surprised by this case? I would have expected that a rather large percentage of a general neuropsychologist's (or at least one that specializes in children) patients would be people who "function at the bottom 1 percent".

He may have thought so because the phrase "income inequality" is usually used in the context of describing it negatively (in the US at least). I guarantee you that the phrase "income inequality" is used by Democrats in the Senate/House 10x as much as Republicans. No Republican is ever going to say their tax cuts for the wealthy "increase income inequality", they'll say it "helps the job creators create jobs".

Edits: Wow, I need to learn to make sure I reread sentences in full when I go back and reword something.

There is recent work along these lines for birds specifically, e.g., [1]. It's not anywhere near human accuracy and I think it's only a couple hundred species but it does kind of work.

[1] http://www1.icsi.berkeley.edu/~farrell/birdlets/iccv11-camer...

I think I may have been slightly exaggerating about trading 5K a year in income for free full service laundry, but I was mostly serious :)

It really doesn't need to cost that much though. I wouldn't be surprised if every state in the US has a public school that costs less than a quarter of that. (and obviously college is much cheaper in every other country). To clarify, I wasn't saying I thought that was the sole benefit of college - just one that never seems to get mentioned.

> "And in addition to your salary, we're also giving you a roll of quarters every month in case you need to pay tolls, do laundry at one of those barbaric laundromats not in the Googleplex, or play arcade games."

... is it just me or would this be an incredible perk? If a roll of quarters and an envelope with $250 worth of $1, $5, and $20 bills were delivered to my desk every month I'd save a lot of time going to the ATM or the laundromat for tips/laundry/cash-only restaurants/etc. I think I'd willingly give up a lot more of my salary than it's actually worth for the convenience of perks like that. Free laundry (like, full wash and fold and dry cleaning)? Shit, for that I'd probably take a $95K job over a $100K job (all other things being equal) even though it's really worth ~$1000 at most (comparing against the cost of laundry services, not machines, of course). Does Facebook still have that?

Not that this is a particularly insightful point, but I keep wondering, why does it never come up in these discussions that college is FUN? Regardless of its financial value or lack thereof, it really is an experience like no other.

And isn't it also possible they gave Starbucks a special deal considering the marketability and value of a deal like this?

Right. Did I say something contradicting that? Google could have decided not to offer application-specific passwords at all, but from any individual user's perspective, that's exactly equivalent to just not using them. At least having application-specific passwords gives you the option, and is at least as secure as giving your master password away to every client application.

I suppose there is one possible negative consequence to users who opt not to use app-specific passwords: their existence alone removes some of the incentive for client applications to implement 2-factor themselves (which I don't know if Google even has an API for). And sure, it would be nice to have features like access control on a per password basis (e.g., so I could allow Pidgin to access only gchat, but no other part of my account). But the implication that the mere existence of application-specific passwords somehow makes Google's 2 factor auth useless is just wrong.

The app-specific passwords are a feature and if you prefer the extra security over being able to use apps that don't support 2-factor, then you can choose not to use them, and get the full security benefits of 2-factor. It's just that, short of expecting every single third-party client app to implement 2-factor authentication or not allowing access to any that don't, there's no alternative to the app-specific passwords.

They are strictly better than using a single password for everything though, in that they are unique and strong (due to being automatically generated and 16 characters long), and easily revocable.

To be honest I was mostly using the comments section for this link as a soapbox, I realize the idea isn't too relevant to this particular case with Dropbox, as no passwords are known to have been revealed. Sorry.

Though, I do think it would often mitigate the damage from this type of security breach that it seems like we've seen so much of from big name tech companies lately. I'd guess that a majority of accounts created on the internet are pretty unimportant to the account creator, and with how often passwords are reused indiscriminately, the worst effect of these password leaks is often not the unauthorized access to all those accounts on the hacked site but rather the usernames and passwords themselves - which are very often reused for bank, email, etc. accounts. With my proposal, anyone who opted not to have a password wouldn't be vulnerable to that.

I wish websites with user accounts would offer the option to "login via email" - as in you'd type in your username (or preferably your email) and maybe a captcha and then you'd login by clicking a link it sends via email afterwards. Ideally having a password associated with the account at all would be optional.

I have a Gmail tab opened just about 100% of the time I'm on the computer, so this would be very convenient for me as an alternative to having to remember passwords for sites that I visit once a month or less (and end up having to get a "password reset" link via email every time I log in anyway), and then I'd only have to keep my Gmail account secure (which I do via 2 factor).

> There's no way for LaunchServices to know, a priori, whether a given executable file's default app is going to perform a safe editing operation, or execute its contents.

But in that sense an executable file isn't any different from a non-executable file, is it? The default app for the file could do equally dangerous things with a non-executable.

> Kind of an odd article, the writer spends the first part of it trying to come up with compliments, which usually fail.

Personally I thought that part was hilarious. I'm not sure whether it was true or not (due to how naive it made the author seem) but I definitely enjoyed the article more because of it.