top The Challenges and Threats of Automated Lip Reading
related dilemma: should we develop algorithms that can lip read? Of course we should, we should develop any tech. The real question is, will it be used for moral or immoral purposes?
Certain technology can be declared illegal. Like guns in certain countries. Radar detectors in some US states. Blue lights on non-police cars in most US states. Mechanisms for counterfeiting printed money. Cloning of human embryos. Et cetera. It's perfectly plausible for a society to declare some particular technology illegal.
Heck, even certain knowledge is illegal for the general public to own, let alone internalize, like plans to make nuclear bombs.
top The Challenges and Threats of Automated Lip Reading
"Dude, you punched a f-ii-sh."
The real point is, though, that although some of those redubbed conversations are like Jabberwoky, some exchanges are reasonable (and some are spot-on visual homonyms, like the fish interpretation above), demonstrating that lip reading is wildly underconstrained.
top Curiosity Rover Arrives At Long-Term Destination
"All extended missions were rated higher than "Good""
Excoriated --- here are a few choice excerpts (there was some positive language, but the panel really did come down hard in this report):
"The panel viewed this as a poor science return for such a large investment in a flagship mission."
"Despite identification of two EM1 science objectives, the proposal lacked specific scientific questions to be answered, testable hypotheses, and proposed measurements and assessment of uncertainties and limitations."
"It was unclear from both the proposal and presentation that the Prime Mission science goals had been met. In fact, it was unclear what exactly these were. "
"After the presentation and subsequent discussion within the panel during executive session, other questions were formulated and then presented to the Curiosity team. Unfortunately the lead Project Scientist was not present in person for the Senior Review presentation and was only available via phone. Additionally, he was not present for the second round of Curiosity questions from the panel. This left the panel with the impression that the team felt they were too big to fail and that simply having someone show up would suffice. The panel strongly urges NASA HQ to get the Curiosity team focused on maximizing high-quality science that justifies the capabilities of and capital investment in Curiosity."
"As Curiosity is a flagship mission, the panel was surprised by the lack of science in the EM1 proposal
"In summary, the Curiosity EM1 proposal lacked scientific focus and detail."
top SanDisk Releases 512GB SD Card
No, a "traditional" GB is the one that was defined way before computer scientists got their hands on it –1000.
Computer scientists? Did they just choose it at random? I thought it was because 2^10 = 1024, therefore 2^30 = 1073741824.
That would suggest, to me, that it was a mathematical definition and not chosen by computer scientists.
More than that, it would suggest to me that 1 GB = 1,000,000,000 was a redefinition of a known quantity by a third party.
Ah, let's get one thing straight here. The notion of a byte did not appear before computer science. Anything that measures bytes is ultimately CS-derived, even if marketing folks like to confuse people.
top How Scientific Consensus Has Gotten a Bad Reputation
As an experimental scientist, I can, with certainty, state that you are wrong when you claim "science is about provability."
It is extraordinarily difficult to prove something experimentally. Most advances come about because we (both individually as experimentors, and collectively as members of a given scientific field), think we've accounted for most potential confounds and artifacts, not because we've conducted perfect experiments. Biological sciences, especially, suffer from a huge number of uncontrolled variables that often we are not aware of, but impinge mightily upon our results. Biology, to continue, is noisy. Very, very noisy. In my lab, we measure phenomena related to visual perception, and I can tell you unequivocally that individual variation usually swamps any underlying phenomenon we examine (meaning, we need to measure with lots and lots of individuals to make sure we aren't being fooled, and even then, we can easily get fooled).
Rarely, if ever, do we prove something experimentally. It's only through the consensus of reproducibility that scientific facts get established.
Piltdown Man, to discuss your example, was due to observational error (ie, a hoax), not experimental evidence demonstrating provability. Observational science, as opposed to experimental science, is rife with missteps and re-interpretations. Look up the history of shooting stars, as one example -- they were considered purely terrestrial phenomena well after the establishment of the United States as a country. It took repeated observational events, not experiments, to establish that meteors are astronomical in origin.
Reproducibility is the cornerstone of modern science. Everything else is consensus. We think we know things, and mostly, we've been correct with a high degree of probability, since we've been able to take given conclusions and build, predictably, upon them. But, every now and then, even firmly-held beliefs with eons of structural experimental integrity are demonstrated to have been mistaken. There is very little scientific truth, merely scientific certainty. If you want absolute truth, look to mathematics instead.
top How Astrophysicists Hope To Turn the Entire Moon Into a Cosmic Ray Detector
If the baseball analogy is accurate, the impact of such a ray should cause something more than just a burst of radio waves. Why don't we see evidence of inexplicable pockmarks on the earth's surface? Or do we? 1 per km2 per centry is a lot when you have such a large surface area like the Earth. Heck, we should have reports of people being stricken down in broad daylight from time to time.
top White House Names Google's Megan Smith As CTO
Control Theory is part of Mechanical Engineering.
And part of Aero-Astro, and Applied Computer Science, and Theory of Computation, and Applied Mathematics, and
Personally, I'd put it in Signals and Systems, smack dab in the heart of EE.
top Buenos Aires Issues a 'Netflix Tax' For All Digital Entertainment
The standard sales tax (VAT) in Greece is currently 23% for most things. (It varies, but that's the most common.) That's on top of the punishing property taxes, income taxes, taxes because you left your money sitting in a bank, taxes because it's Monday, etc. I jest, but only a little.
For those of you living in the US, can you imagine 23% states sales tax on essentially everything?
Argentina has instituted what amounts to a 35% import duty. Yes, that's a lot, but most things are purchased domestically.
top The Data Dome: A Server Farm In a Geodesic Dome
The article lists the requirements for the structure, which include things like massive air flow, high heat density, high electrical power density, etc. Constraints like that tend to point toward structures with high surface area to volume ratios. A sphere (or section of a sphere in this case) has the MINIMUM surface area to volume ratio. So why would you want to put this structure into a dome rather than a long, low building?
(And if you really insisted on getting all fancy, architecturally, you could still make the long low building into a ring and retain most of the advantages.)
top NASA Releases Footage of "Flying Saucer" Braking Test, Declares Success
I saw the live press release on nasa.tv (highly recommended). The principle scientists involved recognized the parachute failure, but emphasized that this is unknown territory, and the mission objectives -- which were to make an attempt and gather as much data as possible about that attempt -- were fully realized.
Yes, the parachute failed. The vehicle was going something like Mach 2 at the time, having successfully aerobraked from Mach 4.7. They got excellent video of the entire process, and only four days (or something like that) after the mission, already had revisions on the parachute in mind to prevent such failure.
This was the first of THREE planned tests. Was the mission successful this time? Absolutely not, if you expected to have a first time test succeed. But if you were looking to gather data on potential failure mechanisms, it was an overwhelming success.
And, it should be noted, the deceleration inflatable ring (which has some kitchy acronym) worked very well, and importantly, they got good data on the design and how much it deviated from perfection (1/8 of an inch deflection at Mach 4.7
... I dare anyone to do that with rigid materials, let alone inflatables). And the blute (the droge which pulls out the main parachute) worked entirely as intended. The downside? The shape of the parachute apparently needs to be more rounded.
They are exploring entirely new territory. Who here really, really, thinks that every such testing and development mission is going to be successful? Anyone? Raise your hands, I want to see, because NASA would love to hire engineers (hell, screw NASA, *I'd* hire engineers) who have that level of talent. They're called experimental missions because the outcome is not known.
about a month and a half ago
top MIT Considers Whether Courses Are Outdated
Yes, it is two subway stops. And about 30 minutes of transit time each way, once you factor in the time to walk to and from the subway stations, the unpredictability of the Red Line frequency (although I must admit it has gotten heapsload better in the last few years; and major kudos to that skunk works project that brought the T administration kicking and screaming into the 20th -- yes 20th -- century by implementing time-to-next-train displays). While not an insurmountable impediment, it does mean that any given inter-campus class requires an empty slot before and after in your schedule. That too is not insurmountable, but now you're talking about two big impediments, so the motivation to attend physically has to be really high.
Here's an example from personal experience. MIT students are also allowed to cross-register at Wellesley College. As a male student at MIT, the motivations for doing so were really high when I was in school. I registered for an astronomy class at Wellesley, with the additional chance of getting some telescope time that I wanted almost as badly as to be around college gir-- I mean women. Even with all those attractions, I dropped the class because it was such a huge time sink when you factored in travel time (that and the class I had registered for was teaching stuff I had learned on my own as a kid by reading books).
So, two subway stops? Not quite close enough unless you have really motivated students. Internet attendance of lectures with once per week recitations that required physical attendance? That would work better.
about a month and a half ago
top Two South African Cancer Patients Receive 3D Printed Titanium Jaw Implants
At the prices medical-grade titanium goes for, it is most certainly not wasted. The machined Ti is reclaimed (or at least it would be if I were in charge). Stating that there is 80% waste is marketing hyperbole. A fairer comparison would count the unsintered powder in the 3D build machine, and would end up being unfavorable to the 3D process.
But if you're in the business of making replacement body parts, you might well be starting with a generic titanium casting (or one of a series of different sizes) and machining it down to fit. Artificial hip joints are sometimes made that way.
Don't get me wrong, 3D printing makes a lot of sense for highly-custom items... although one needs to worry about the potential infection and reaction issues given the inherent porosity of sintered material that give purchase for pathogens, and lots of surface area for irritants that will slowly leech out.
top A Warm-Feeling Wooden Keyboard (Video)
1. "All the Keyboards" didn't apparently include a Kinesis. At least there isn't one visible amongst the few photos linked.
2. The new keyboard looks an awful lot like a Kinesis.
3. I stopped watching the video after the first 10 seconds because it was too awful.
4. The web site shows a keyboard with what appears to be a metal case, and the text references aluminum, as does the blog. Wood isn't part of the equation here. Maybe in the early prototypes, but not in the production models, apparently.
5. Any decent keyboard driver (and there are lots of aftermarket add-ons) support macro definitions. Nice that this new keyboard supports it, but certainly not a defining characteristic.
6. Just go buy a Kinesis. It's been in production for a long time, and they work great.
top Favorite "Go!" Phrase?
top Comcast Customer Service Rep Just Won't Take No For an Answer
"The little yellow rabbits with the red eyes told me to cancel. There are so many of them here now."
"I need to tie up some loose ends before heading over to the Comcast office with my AK-47." Switch to German. If the CSR speaks German, switch to French. Then Portuguese. Russian. Greek. Etc. "Oh, that's really interesting, please tell me more [wait for reply] REALLY? That's so interesting. Please tell me more. [wait for reply] I didn't quite understand that, could you repeat it? ..." "Zorg said it must be done." "The CIA keeps sneaking into my house through the Internets to read my mind while I sleep." "What did you say your name was? [type, type, type] OK, then, I've updated your Facebook page for you. Hope you like the new photos -- boy that sex toy sure is big!" "Oh, can you wait a sec? There's someone at the door." [leave phone on table by loud, badly tuned radio, and walk away for a good 40 minutes] Go to the bathroom and make appropriate noises as if defecating with some difficulty and repeatedly flushing, but continue to hold the conversation. Hand the phone to your 5 year old. Aim an air horn at the microphone. "Each time I try to move the cable box, the electrons keep spilling out." "When I use the Internet at my friend's house, it's OK, but at my house, all the photos are upside down." "I need to spend more special time with my hamster." "I'm dead." "I've just had enough of the sparks." "It interferes with my hair dryer." "This chick said she'd sleep with me if I cancel my cable." "The Internet is too big."
top Ikea Sends IkeaHackers Blog a C&D Order
Tax avoidance makes more sense than any other speculation in this discussion.
Were IKEA organized as a normal for-profit venture, then anyone with half a brain at IKEA would see the utility of IkeaHackers and do one of a handful of things:
1. Buy them outright.
2. Obtain favorable advertising terms in exchange for licensing (eg, ikeahackers gets to keep running ads, but 50% of the spots must be filled with IKEA adverts, and sales-active links to the official items mentioned in each article must be included). 3. Think they're cute, and provide gratis branding coaching (including direction on proper use of their logo, precise color usage, etc) in exchange for disclaimers, and big, obvious links back to IKEA. 4. Sign an exclusive advertising deal: ikeahackers.com gets to keep operating, but must only carry IKEA adverts, with some affiliate payment structure for completed sales that originate on ikeahackers.com so that the site can continue to live on. 5. Similar mutually-beneficial arrangements . . .
IKEA have an enthusiastic fan base who, inspired by reading the blog, will likely go out to buy more IKEA product. This is not just a good demographic, but a great big juicy one. These are the people they *want* shopping in their stores, the people they *want* to reach through media campaigns.
So why turn them away? The only conditions that come to mind are when bringing the web site into the fold has larger, more threatening implications to the corporate structure, as the parent post suggests.
top TechCrunch and Others On the Microsoft Surface Pro 3
Paper. I have lab notebooks from my undergraduate years (through the present) that I refer to, because it's easy to do so. If I need to find something, and don't recall exactly where it is, a simple flipping of the pages, and I've found it in a matter of seconds.
Call me a luddite, but electronic lab books don't have sufficient usability yet. My post-doc uses one, and he's far less efficient with it (and writes much less as a result) than I am with pen and paper. Importantly, human memory is at least partially (some would argue primarily, and I wouldn't dispute that) visually based, and not having that aspect of where you wrote something on the page, or in what color ink, or with what size handwriting, etc., makes a big difference to usability. The closest I've seen to viable electronic notebooks are the specialized pens that have small internal cameras that require special paper to use. Last I checked, though, they were prohibitively expensive compared to a traditional pen and lab book.
Pen and paper have had thousands of years of development. It's going to take a while longer before the electronic version is usable.
top Brain Injury Turns Man Into Math Genius
I don't buy the thermodynamic argument. That's an epiphenomenon (i.e., correlation not causation).
Basic mathematics does not consider time. Nor does it really consider sequential ordering properly until it deals with the notion of state, State begets the field of Computational Theory. Before a proper wrangling of the ideas at the core of Computational Theory (as embodied in Turing Machines, for example), there was a horrific thrashing-about that was particularly inelegant, such as the attempts in First Order Logic to capture the meaning of state.
Now, when you talk about infinity, you are effectively saying that results from Computational Theory are being computed by a machine that takes zero time to get from one state to the next. Nothing in the physical universe takes zero time, and so infinity is considered to be ill-supported by Nature. That's the crux of the Constructionist objections.
Now why should mathematics be beyond time? It describes space pretty well. Why should time be exempt?
top Brain Injury Turns Man Into Math Genius
There are a fair handful of constructionists (aka finitists, or number theorists who do not like infinities) who would care to disagree.
Although I'm not a constructionist, I am related to one by birth, and nearly always find something off-putting about mathematical arguments that rely on infinities. Sure, they're fun to play with, but reasoning about them means you're essentially being fast-and-loose with time, and I've not been convinced that's OK.
top One-a-Day-Compiles: Good Enough For Government Work In 1983
And I bet that because the costs of running a program were so high -- in time and monetary terms -- that your programs were correct, nearly 100% of the time.
We get intellectually lazy when a debug loop is only a few seconds long.