Page 308 of 418 FirstFirst ... 208258298306307308309310318358408 ... LastLast
Results 3,071 to 3,080 of 4172

Thread: The OT thread V1

  1. #3071
    Insider
    Join Date
    May 2012
    Posts
    1,182
    "So you've done this before?"
    "Oh, hell no. But I think it's gonna work."

  2. #3072
    Insider
    Join Date
    May 2012
    Posts
    1,182
    "So you've done this before?"
    "Oh, hell no. But I think it's gonna work."

  3. #3073
    :mixed emotions intensify:

    i was talking with the ASML guys at the last conference i was at (one was also a Formula Ford racer) and they assured me that moore's law has at least 20 more years in it. they are developing technology to write 7.5nm and smaller tracks right now. they are basically targeting single photons at this point.

    on one hand moores law continuing is unquestionably a good thing for civilization.

    on the other, based on my very real slytherin tendencies, moores law ending will finally make computer and software engineering like actual engineering.
    social conservatism: the mortal fear that someone, somewhere, might be having fun.

  4. #3074
    Insider
    Join Date
    May 2012
    Posts
    1,182
    Can you elaborate on what you mean by that? I assume you mean in that different schema will need to be brought to bear to continually improve performance, as opposed to turning the lithography dial down. Perhaps you mean something a bit subtler, for instance the relationship of the discipline to bounded uncertainty, or imprecision, but unclear.

    I relish an era of divergent computation. At the highest level of abstraction, most of the new paradigms increase computational efficiency by co-localizing memory and processing. It will be very interesting to see what's scalable, and indeed what even matters, to 21st century computing.

    Macro trends in this space (cloud, mobile, SaaS) tend to coalesce both data and processing in the hands of a few gargantuan players. It's interesting that the Silicon Valley ethos is largely libertarian, as the previous computational epoch had a relatively flat distribution of resources, as evinced by the classic John Carmack quote:

    "In the information age, the barriers just aren't there. The barriers are self imposed. If you want to set off and go develop some grand new thing, you don't need millions of dollars of capitalization. You need enough pizza and Diet Coke to stick in your refrigerator, a cheap PC to work on, and the dedication to go through with it. We slept on floors. We waded across rivers"

    With deep learning, which essentially builds algorithmic classifiers from data, rather than from insight, we've largely undone the very underpinnings that allowed the titans of technology to arise in the first place. In some ways, I feel like the only solution to the the privacy concerns the currently plague us is to make all information available. A kind of digital radical honesty. It's unclear to me whether this would drive the value of the data down or not, but certainly it would drive the cost of processing up. The fact that the data is used for nefarious and manipulative purposes (advertising, hijacking elections) is of far greater concern than it's mere existence, but in a world where everything is quite obviously open perhaps we can develop our informational immune systems.
    "So you've done this before?"
    "Oh, hell no. But I think it's gonna work."

  5. #3075
    Quote Originally Posted by Lurker27 View Post
    Cannot believe that worked. Wood unbelievable in the right application
    Largest problem with wood is the imposed limitations you have to design in because the characteristics of each board varies so much.

    Your allowable bending stress is between 800-1500 psi usually which is insane when you watch something like that

    (From the guy who is mad at wood while doing some support of excavation designs)

  6. #3076
    Insider
    Join Date
    Apr 2012
    Location
    Knoxville, TN
    Posts
    2,303
    Part of Moore's law was self regulating - in trying to keep up with Moore's law computing limited growth when they could have expanded faster, instead of just trying to plan for the expected growth. This helped software keep up and related growth was kind of planned, so it was nice, but only aiming for an expected bit of growth means they did just aim a little higher each year. Now they are at some limits of existing, amortized technology, where they might have pushed some of the tech like 3D board design or photonic processors decades ago as moonshot units that could have resulted in leapfrogging our existing tech.

    Here is a plan for a photonic Thz chip for example: https://www.nextbigfuture.com/2018/0...es-faster.html

    And 3D at 1000 times current speeds?: https://www.livescience.com/52207-fa...uter-chip.html
    Josh Coray
    J4 Paintball
    Lead Design
    www.j4paintball.com

  7. #3077
    Insider
    Join Date
    Apr 2012
    Location
    Knoxville, TN
    Posts
    2,303
    Oh wood....:

    https://www.smithsonianmag.com/smart...ood-180968117/

    *Soft woods like pine or balsa, which grow fast and are more environmentally friendly, could replace slower-growing but denser woods like teak, in furniture or buildings,* says Liangbing Hu, a materials scientist at the University of Maryland, College Park and one of the study authors, according to a press release.

    Other methods of *densifying* wood produce material that tends to expand back toward its original size and shape when humidity strikes, tells Perkins. The new approach enhances wood in a way that is moisture resistant by creating a more uniformly dense end result.

    *This kind of wood could be used in cars, airplanes, buildings * any application where steel is used,* Hu tells David Grossman, reporting for Popular Mechanics. Skyscrapers of the future could even be built from wood.*

    The researchers then subjected their densified wood to a battery of tests to measure its mechanical properties. They tested its capacity to bear a load. They tested its stiffness, it*s tensile strength (how well it holds up to pulling forces), and how well it could flex. Overall, the result was one rough and tumble piece of wood. In a humidity test, the wood only swelled about 8.4 percent, while still remaining tough.

    They also fired a steel projectile at the wood with a ballistic air gun. Five layers of the material together halted a slug moving at the speed that a car might be before a collision, reports Mark Zastrow for Nature.
    Josh Coray
    J4 Paintball
    Lead Design
    www.j4paintball.com

  8. #3078
    Quote Originally Posted by Lurker27 View Post
    Can you elaborate on what you mean by that? I assume you mean in that different schema will need to be brought to bear to continually improve performance, as opposed to turning the lithography dial down. Perhaps you mean something a bit subtler, for instance the relationship of the discipline to bounded uncertainty, or imprecision, but unclear.
    no, i mean literally, they are making transistors modeling the masks to control basically single photons. 7.5nm and below traces is 4 or 5x smaller than current high end machines. ASML is the leading company that makes the lithography machines that write the chips themselves. so there is still 10-20 years left in the hardware alone to support moore's law. remember, people have been saying since the 1980s that moores law is ending, and it hasn't yet.

    chip makers literally cannot make them fast enough, they also told me that literally every single lithography machine they have ever made, is still being used. and when one fails, it is expected to have the machine back up and running in less than 24 hours, because its so expensive to have the machine down, and it fucks with there supply chain so much. they are building them as fast as they can, and every one is still in service.

    its an amazing industry, that puts the level of precision machine control we run every day look like child's play. they can do this day in day out compared to us for many reasons, but its impressive none-the-less.
    social conservatism: the mortal fear that someone, somewhere, might be having fun.

  9. #3079
    Insider
    Join Date
    May 2012
    Posts
    1,182
    Yes, we sell filters into that industry, I am very familiar with it and the drive to EUV at smaller nodes (and associated challenges). Because of the cleanliness (parts per quadrillion of nm particle foulants) and uptime requirement, our premium products command premium prices without blinking.

    Die shrink has perhaps an additional 10 years left - this is good reading for an introduction to the topic:
    https://en.wikipedia.org/wiki/5_nanometer

    Moore's Law, however, has already been violated, because it is an improvement per time metric. To respond to Josh's comment, die shrink was always the best expectation value research project, and it's only now that exotic schemes will get a good look. These are the downstream effects on computer and software engineering, I wanted to discuss. which should have been obvious since I wrote 5 paragraphs about it.
    "So you've done this before?"
    "Oh, hell no. But I think it's gonna work."

  10. #3080
    Insider PBSteve's Avatar
    Join Date
    May 2012
    Location
    Los Angeles
    Posts
    3,084
    Does anyone have any recommendations for UUID/serial number generation software? I need to be able to generate a unique, 4-8 character alphanumeric string on-demand, preferably in the format of reputable software. I could hack together something in excel but I'm looking for something more professional due to management.
    Ever so many citizens of this republic think they ought to believe that the Universe is a monarchy, and therefore they are always at odds with the republic. -Alan Watts

    I work for the company building the Paragon

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •