Tag: NOTABUG and WONTFIX. Case aware filesystems so you can have normal names and not like AUTOEXEC.BAT and CONFIG.SYS from the DOS days is great, case sensitive file systems are a really bad idea. Is there any kind of sane situation where you'd like to have two files "Config" and "config" actually coexist that isn't just begging to be confused/abused/exploited? For a marginal performance optimization all POSIX systems have shitty usability. Why am I not surprised? I guess for a server it just doesn't matter, but for the desktop you should file this as a bug against Linux, not Windows and OS X.

Sony decided it wasn't economically worth going to war with North Korea over this. That's all. Even the very slight risk of a theater bombing could have cost millions, so they said no.

That said, I hope they carpet bomb North Korea with Korean dubbed DVDs of the movie, complete with DVD players. Also hoping Netflix will start streaming it tomorrow, or that they just release it for free.

I think it was fairly clear what was going on, the neural networks latch on to conditions that are necessary but not sufficient because they found common characteristics of real images but never got any negative feedback. Like in the peacock photo the colors and pattern are similar, but clearly not in the shape of a bird but if it's never seen any non-bird peacock colored items how's the algorithm supposed to know? At any rate, it seems like the neural network is excessively focusing on one thing, maybe it would perform better if you divided up the work so one factor didn't become dominant. For example you send outlines to one network, textures to a second network and colors to a third network then using a fourth network to try learning which of the other three to listen to. After all, the brain has very clear centers too, it's not just one big chunk of goo.

"First they ignore you, then they ridicule you, then they fight you. Then you lose and kill yourself."
- Hitler (well, not really)

I never understood what that Gandhi quote is so popular, sure that's what a victory looks like out the rear view mirror but most defeats start just the same.

I'd love to be able to publish these statistics for our organization, (I'd estimate we have close to a quarter million drives in the field) but there is a big hurdle in the way: legal liability. If I was to say something negative about Western-Sea-Tachi drives, their lawyers might call our lawyers, and we could easily spend a million in court fees.

The thing I think would be interesting is that we have a completely arbitrary mix of drives, based on drive availability over the last 6 years or so. We also have a mix of different service companies who replace the drives in our workstations. Our contract is such that we don't control the brands, or even the sizes, as long as they meet or exceed our specs. As a service organization, they're responsible for picking the cheapest option for themselves. If our spec says "40 GB minimum", and they can't get anything smaller than 500GB, they'll buy those. If 1TB drives are cheaper than 500GB drives, they'll buy those. And if we're paying them $X/machine/year for service, they can do the reliability decisions on their own, so if they think some premium drives will last two years longer than stock drives, they might be able to avoid an extra service call on each machine if they spend $Y extra per drive. I expect these service organizations all have their preferred drives, but that's not data they're likely to share with their competitors on the service-contract circuit.

I don't take pictures for "posterity", or for people who outlive me. I take pictures for me, and my family, for now. While I only have thousands of total pictures, (not 10,000 per month) I can still find the pictures I want on my hard drives. So when I die, if some future grandchild wants to trawl through those terabytes in the vain hopes of finding a good picture of a great-great-grandparent they never met, why should I care? What difference would that make to me, today, in how I choose to save or discard photos?

Somehow, I doubt that any country outside the USA is going to tolerate this nonsense for very long. Entertainment can be boycotted. Other networks can be created. If the MPAA is dumb enough to try this (which is likely, because, you know... entertainment industry), they will just hasten the creation of a new and better decentralized set of internets.

Seriously, unless for some magical reason all progress in AI stops, it will just get better, either incrementally or with some number of breakthrough techniques. More and more cognitively complex jobs will be handled by relatively cheap AI. Doctoring, lawyering, teaching, government policying, CEOing.....

Eventually, AI is good enough to design better AI and away we go. We're not far from that now. Genetic algorithms, a physical full sensory substrate, a set of built-in motivations plus an evolvable neural network system would probably get us there in the next 50 years, assuming technological civilization can still support such things (not a given).

So, the more interesting question is, "What do we do when there's nothing a machine can do better?" Effortless hedonism has its place, but I doubt humans could stomach this on a full time basis. We need purpose.

It was the 1960s. You were lucky to have a 300 baud modem, they wanted to save two bits by chopping the "19" off 1960 and encryption was regulated as munitions. Heck, even in the 1990s they wanted to restrict my browser to 40 bits so I didn't have "export grade" cryptography. I still hear cost for servers and battery life on clients as an argument for why sites don't move to HTTPS, The very idea to build the Internet with strong encryption by default was ridiculous on technical merits and I don't recall anyone even suggesting it so feel free to quote some sources.

Yes, MITM attacks are possible. But unlike wiretapping they're also detectable and I don't just mean in the theoretical sense. You could still use CAs to "boost" the credibility of an IP encryption key fingerprint (The CA signs my cert, I sign a message saying my IP uses fingerprint aa:bb:cc:dd:ee:ff), you can verify by proxy (connect to your server from friends/family/open wifi/proxy or ask a third party to what certificate fingerprint they see) or you can use in-band ad hoc verification. For example you're in a chat and it says at the top "finger print for this session is aa:bb:cc:dd:ee:ff" you might say "reverse it and you get ff:ee:dd:cc:bb:aa" or "third pair is a double c" or "last two are 255 in hex" as part of the conversation. Even better if it's voice communication, think they can MITM a buddy saying the fingerprint?

MITM only works if there's a protocol you can use to automatically block/filter any information about the key. For example imagine you take a photo, overlay the fingerprint semi-transparently and display it on your website. Now they have to create a very custom solution for your site to create an identical photo to replace it with. Transparent MITM in an interactive process - not just your cell phone checking your mail - is going to be really tough to do on a mass scale. It won't have the perfect theoretical characteristics, but it sure will work for most people most of the time.

An English major is NOT getting into a STEM Ph.D. program, no matter what.

Even if they were, job prospects are worse for STEM Ph.D. holders than for MS/BS holders—there are far fewer jobs that require Ph.D. level qualifications outside of the professoriate and academics, and for Ph.D. holders in particular, employers are absolutely loathe to hire overqualified people.

Inside the professoriate and academics, the job market is historically bad right now. It's not "get a Ph.D., then become a lab head or professor," it's "get a Ph.D., then do a postdoc, then do another postdoc, then do another postdoc, then do another postdoc, really do at least 6-7 postdocs, moving around the world every year the entire time, and at the end of all of that if you've managed to stay employed at poverty wages using highly competitive postdocs that you may not even get, while not flying apart at the emotional seams, you may finally be competitive enough to be amongst the minority of 40-year-old Ph.D. holders that gets a lab or a tenure-track position, at which point the fun REALLY begins as you are forced onto the grantwriting treadmill and feel little job security, since universities increasingly require junior faculty to 'pay their own way' with external grants or be budgeted out."

And that's INSIDE STEM, which this person is almost certainly likely to be uncompetitive for as a B.A. holder trying to get into graduate programs.

Much more likely is that with great grades and GRE scores they'll be admitted to a humanities or social sciences Ph.D. program, with many of the same problems but with CATASTROPHICALLY worse job prospects due to the accelerating collapse of humanities budgets and support on most campuses.

Ph.D. is absolutely not the way to go unless you are independently wealthy and are looking for a way to "contribute to the world" since you don't actually have to draw a salary.

For anyone with student loans, it's a disastrous decision right now, and I wouldn't recommend it.

I say this as someone with a Ph.D. who is on a faculty and routinely is approached by starry-eyed top students looking to "make the world a better place" and "do research." Given the competition out there right now, only the superstars should even attempt it, and then only if they're not strapped for cash. Hint: If you don't know whether or not you're a superstar, you're not.

I think in a decade I've strongly recommended that someone enter a Ph.D. program once, and greeted the suggestion favorably maybe three times total, out of thousands of students, many of them with the classic "4.0 GPA" and tons of "books smarts."

In short, I disagree strongly with the suggestion. Unless you absolutely know that you're competitive already on the academic market, DO NOT GO. Don't listen to the marketing from the schools; it's designed to drive (a) your enrollment and tuition, and/or (b) your cheap labor as a teaching assistant/research assistant forever once you're in the program. It's a win for the institution, not for you.

The easiest sanity checks: Do you know exactly what your dissertation will be about and what you'll need to do, in broad strokes to conduct your research, as well as what resources you'll need? Do you already have personal contact with faculty on a well-matched campus in a well-matched department that are championing you and that want to bring you in as one of their own students/assistants?

If you answers to either one of these questions is "no," then while you may be offered a position somewhere, you will be on the losing end of the deal and would be naive to take it.

It comes to understanding the methodologies. What makes each effective? What are their weaknesses? Do you have enough good people who can execute them?

Waterfall is often appropriate, especially when it comes to physical world engineering, or for software products that cannot and will not be changed. Agile is great when you are committed to fully automated testing, have a committed stakeholder who is an active participant, and can deploy on demand for low cost.

But many clients now expect instant updates like they experience with their iPhone apps, and it's very difficult to deliver like that with waterfall. Agile is the answer, but for legacy projects that lack adequate testing, it's a big challenge to migrate to agile, and requires the business be put on hold while the developers clean up their technical debt. Most businesses can't afford such a shift.

That way is also not a guarantee of success. If management is implementing their imagined-perfect new organization structure, they are often blind to the problems they are creating, believing the problem lies with the underlings who "aren't trying hard enough", or "don't believe in the vision."

Not a damn thing. As a matter of fact, the original HBR story referenced in the TFA is not about IT at all. And TFA could have been written by Captain Obvious, except it's not nearly as clear.

That depends. For example many countries have laws regarding historical artifacts, you can own them but you can't take them out of the country. Or you can legally buy cryptography chips in the US that needs an export license. Just because Microsoft Ireland can legally possess the customer data in Ireland, doesn't mean they're free to ship it around the world or provide access to it in violation of Irish law.

So either you comply "voluntarily" or your lack of compliance is used as a reason do to if forcefully, either way the cops get to do whatever they want. Maybe they should start at home and repeal the 4th amendment first?