Science and Technology Links – Holy Cow It’s June Already! Edition

Oscar Gordon

A Navy Turbine Tech who learned to spin wrenches on old cars, Oscar has since been trained as an Engineer & Software Developer & now writes tools for other engineers. When not in his shop or at work, he can be found spending time with his family, gardening, hiking, kayaking, gaming, or whatever strikes his fancy & fits in the budget.

Related Post Roulette

38 Responses

  1. Kolohe says:

    ” wonder how long before someone turns out GMO variants that can produce other colors?”

    a quick google search comes up with this – not a different color than something currently in use, but a GMO substitute for existing plant and chemical feedstock dye sources.

    edit – I misread where the story was going, I thought it was about food/plants used for dyes, not the dyes used for food.Report

    • Oscar Gordon in reply to Kolohe says:

      My question was more to the idea that if we have corn variants for blue & purple, how long before we have GMOs for green, red, etc.Report

      • dragonfrog in reply to Oscar Gordon says:

        No need to genetically modify corn – there’s already all the colours you could possibly want, just by selective breeding. (Though the “red corn chips” in the grocery stores are made with regular white corn and beet juice, as it’s cheaper than red corn…)

        Do a google image search for “glass gem corn” – it’s a pretty spectacular variety.Report

        • Oscar Gordon in reply to dragonfrog says:

          But can they extract the various color from the corn and A) have it be a viable food dye, and B) still use the bulk of the corn for other food purposes?Report

          • dragonfrog in reply to Oscar Gordon says:

            Ah, I see. At a guess, it seems the fact that red corn chips are usually made with beet juice colour suggests that white corn + beets (since by the sound of it the beets aren’t usable for much after dye extraction, so their entire value is sunk into the corn chips) is cheaper than red corn on its own.

            So, it might be possible, but not economical.Report

            • Troublesome Frog in reply to dragonfrog says:

              That’s where the genetic modification comes in. If you can get your highly productive bulk corn varietals to have strong red pigment, then it might become economical enough to extract dye from them.

              I find genetic modification pretty interesting in part because pretty much any problem that we can transform into “growing lots of corn” becomes economical, given how the US is tooled up. Say what you will about the US, but we can grow us some corn.Report

  2. Kolohe says:

    The quantum computer that IBM had built (is building?) looks ‘functional’ but not yet ‘practical’, am I reading that right? The programing still seems dedicated to programs that experiment with quantum computing itself, but it’s not yet running any ‘applications’ per se.Report

    • veronica d in reply to Kolohe says:

      Well, 17 qubits isn’t enough to factor any prime numbers we haven’t already factored, and given that factoring large primes is the only practical application of quantum computing (yet discovered), no, this will have no application.

      (Note, I’m saying “practical application of quantum computing,” not “practical application of quantum communication.” The latter is a rather different beast.)Report

      • DensityDuck in reply to veronica d says:

        It seems to me like it will have some applications to matrix mathematics, which (like factoring) is one of those “proven by the laws of mathematics that there’s no way to do it other than to do it” things.Report

        • Oscar Gordon in reply to DensityDuck says:

          Oh, man, imagine a tri-diagonal or penta-diagonal solver algorithm running in a quantum computer! Although I don’t know that they could run as they exist on such a machine, but you get my drift.Report

          • DensityDuck in reply to Oscar Gordon says:

            One of the biggest realizations about electronic computers was that we could quit trying to invent closed-form solutions for everything and just do dumb-but-simple brute-force numerical solutions.

            Maybe there will be a similar realization for quantum computers; that there’s no need for all the tricks we’ve come up with to do matrix math faster because “size of input” is no longer a driver in the solution process.Report

          • veronica d in reply to Oscar Gordon says:

            The thing is, you get something like a sqrt(N) speedup for polynomial algorithms, which fine, but that is against your base clock rate. What’s the effective clock rate versus claasical computers? How many qubits can you entangle? At what scale? Distributed?

            Keep in mind, quantum states are fragile. Classical computers are limited by bus/network contention, heat dissipation, and ultimately the speed of light. These are well-solved problems.

            After all, we just used one to beat Go.

            You can’t just “scale up” QC that way. It looks like a very hard problem.

            Is there a sweet spot where you can entangle enough qubits to invert a larger matrix than we can easily handle with large distributed classical systems, dividing out operaring costs?

            I’m pessimistic.Report

            • veronica d in reply to veronica d says:

              For a more thorough review, see this list: http://math.nist.gov/quantum/zoo/

              You’re looking for the “superpolynomial” speedups, for algorithms where we might expect QC to be worth the enormous extra effort of getting the needed 10k+ qubits running. Most of the (in theory) successful algorithms to date are based on a single “neat trick,” which is related to the abelian subgroup problem, combined with a convenient form of the quantum Fourier transform.

              Anyhow, it’s been a while since I dug into the math. From a brief review of the “zoo,” they’ve been making many advances. So good. But still, these are very narrowly described problems. Right now we can invert pretty large sparse matrices without much trouble. At my employer, we will network 10k+ cores together, with gigabits available memory, to solve “cool trick” puzzles, like winning board games. We continue to scale out, do more with less, push things toward the light-speed/”how the hell do we cool this” limits. 17 qubits is laughable compared to this. I don’t need to invest millions to discover that 12034 equals 2 * 11 * 547.Report

            • veronica d in reply to veronica d says:

              Specifically on linear solvers (which I guess HHL is new since I last delved into QC), I’m reading this: http://www.scottaaronson.com/papers/qml.pdf

              I’m not all the way through, but it is a familiar story for anyone researching QC: great speedup is possible, but only for a limited class of problems that happen to have very nice mathematical properties.

              This is cool. We will probably find application for this stuff, as we learn its contours and limits. But to outperform a 10k+ classical distributed cluster, crunching away at a general problem — don’t hold your breath.Report

      • Kimmi in reply to veronica d says:

        Plenty of practical applications of quantum computing. It’s just proving them to the mathematicians that’s difficult.Report

    • DensityDuck in reply to Kolohe says:

      The first non-quantum computers couldn’t do a whole lot beyond basic mathematics, either.

      That said, we had to invent semiconductors before we had computers that were both big enough to do useful things quickly and small enough to be portable.Report

    • Oscar Gordon in reply to Kolohe says:

      I’m just kinda stoked we got one running, and kudos to IBM for doing it.Report

  3. Michael Cain says:

    Re the FENG material (took me a bit to figure out that SHUI remark)… My first reaction wasn’t “foldable loudspeakers” or “talking newspapers”, but rather the possibility that we might finally have flat loudspeakers that sound good and are affordable. Decent sound in small rooms…Report

  4. John Wetzel Jersey.Professional Mike Reilly Jersey
    Outlet Sale, 83 Vincent Jackson Jersey From China Store for NFL Fans.

    Buy Discount NFL Football Jerseys, Hockey jerseys here.Report

  5. Michael Cain says:

    The Braga/Goodenough cell uses an exotic solid glass electrolyte to suppress dendrite growth rather than an exotic anode material. Me, I want someone to solve the expansion problems that plague sulfur-based rechargeable batteries, since we have mountains of the stuff piling up at places around the world.Report

  6. DensityDuck says:

    These were some really neat links. Thank you!Report

    • Oscar Gordon in reply to DensityDuck says:

      Thanks. I am trying to focus the links down to discoveries, breakthroughs, and novel applications, rather than just the latest whizbang. It helps that my time is tight right now, which made tossing the whizbangs in the slush pile easy.Report

  7. Jaybird says:

    When I was reading the article about folk contraceptives, I was wondering if they’d bring up Silphium: a plant driven to extinction by the Romans by the 2nd or 3rd century BC because, tah-dah, it was a folk contraceptive. (The seed, apparently, looked like a Valentine’s Day heart.)

    We, seriously, need to put a hell of a lot more money into drug R&D.Report

  8. dragonfrog says:

    Interesting that one of the folk contraceptive compounds is found in mangoes. I wonder if it would be effective at ordinary mango consumption levels – could someone sabotage their attempts to get pregnant by eating a few mangoes a day, for example?Report

  9. Gordon Chu says:

    This is the right site for anybody who wants to find out about this topic. You know so much its almost hard to argue with you (not that I actually would want laugh out loud). You certainly put a brand new spin on a subject which has been written about for years. Wonderful stuff, just excellent!Report

  10. DavidTC says:

    The ‘everyone is cheating on diesel emissions’ is what my brother (an auto mechanic) was saying back then.

    And some of the ‘cheating’ isn’t even illegal. Volkswagon did something that was clearly *on purpose*, and was illegal.

    But at this point, as people start doing real world testing of diesel cars, it’s becoming clear that *none* of them meet the emissions requirements. Like, absolutely none of them.

    A few more companies will be discovered to have done something on purpose, and get fined, but I suspect that most of them just spent every single second tuning the car to met the emissions standard while under that specific testing, and did not bother to even test it anywhere else.Report

    • Damon in reply to DavidTC says:

      “did not bother to even test it anywhere else.”

      Why WOULD you test it otherwise. It has to pass this test. Are there requirements that it pass other tests, say a real world mixed driving course? I doubt that. It’s the same as the EPA fuel economy numbers. Everyone knows that’s aren’t accurate, but the car is tested that way.Report

      • DavidTC in reply to Damon says:

        It’s the same as the EPA fuel economy numbers. Everyone knows that’s aren’t accurate, but the car is tested that way.

        I get where you’re coming from, but the MPG standards at least *try* to be accurate to the real world, they’re just a ‘perfect driver’. And you can usually remove 15% or so MPG and be accurate for most drivers. (Or, weird idea, we could actually teach people to drive like that and keep their tires inflated and all the other stuff that car manufactures use to squeeze that extra 4 MPG out of a car. Sorta weird that we’re complaining that car companies are better drivers than most people, instead of saying ‘Hey, why don’t we all drive like that!’)

        Whereas the emission standards aren’t even *close* to accurate…and they’re inaccurate for an *incredibly stupid* reason, namely, that collecting and testing emissions while driving *used* to be very difficult…30 years ago, or whenever these standards came into affect. (Whereas ‘testing’ MPGs while driving has always been easy….you put in X gallons of gas, drive the car until it stops, then measure distance and divide by gallons. You need no testing equipment at all.)

        Now, of course, and really for a decade, we’ve had the ability to put the emissions testing equipment *in* every car, in the trunk or even the passenger seat. The equipment is smaller than a person, and most of the problem is getting the exhaust *back out*. (1) Hell, it could be done *at the same time* as the MPG test, or at least on the same courses.

        We just…never bothered to actually set up any standard for that. We just stick them on treadmills and run them.

        1) Dumbest emissions testing result ever: Our study has determined that this car’s emissions, when piped into the car for testing, causes the driver to pass out, rendering us unable to test for more than a few minute at a time.

        More seriously, we should have car companies just cut two holes in the rear window of the car they’re testing (I.e., we explicitly say it’s okay to ‘alter’ the car before the test in that way.)…one to pipe the exhaust into the testing machine, and one to dump it out, in a place where it wouldn’t affect airflow much and thus not alter the car’s performance. As opposed to going through the existing windows or propping the trunk partially open.Report

        • Oscar Gordon in reply to DavidTC says:

          Cut a hole in a window? No, you just roll down the window and mount a plastic fairing with the appropriate hole in it. You can even do a little CFD on the whole thing, make sure it won’t impact the vehicle aerodynamics enough to matter.

          But holes in auto glass are a PITA.Report

          • DavidTC in reply to Oscar Gordon says:

            I don’t know what cars you’re driving but you cannot roll down the rear window of any car I’ve ever seen!

            I’m pretty sure you’re talking about a rear side window.

            If that doesn’t screw around with aerodynamics too much, that would be fine. I’m just saying if they object, let them cut a hole (or two) in the big pane of glass at the *back* of the car, whatever you want to call it, which is not going to affect aerodynamics to any extent. Or, hell, cut a hole up into the trunk. (As long as they have to carry around any metal they cut out of the car.)Report