Relive the magic of IoT, Data, and the New Last Mile; members may now access session videos on our website! (Login to access)
The NSA has open-sourced Ghidra, a tool it developed to reverse-engineer compiled code back into program instructions. (Armando Solar-Lezama, Washington, D.C., Sep 2018)
Ars Technica has an interesting article on the topological and metamaterial properties of knit fabrics. (Connie Huffa, San Francisco, Dec 2018; Julia Greer, San Francisco, Dec 2015)
Speaking of metamaterials, who hasn’t dreamed of Terminator-like liquid metals (Don Sadoway, Charlotte, Dec 2010)? When Chinese researchers added a bit of iron to a droplet of gallium–tin–indium alloy immersed in hydrochloric acid, a Ga2O3 layer formed on the droplet’s surface, lowering the surface tension of the liquid metal and making it possible to intentionally deform the droplet with magnets to stretch it to four times its native length or to connect to electrodes to complete a circuit.
For all the Nicola Tesla fans, a 10-minute film that beautifully depicts the genius and madness of the man. (Joseph Paradiso, Montreal, Apr 2004)
If you were on our field trip to NYU’s Tandon School of Engineering (Brooklyn, Jun 2018), you’ll remember the Sounds of New York City project to collect citywide data on noise pollution. (Greg Dobler’s Feb 2016 talk in Austin also briefly discussed SONYC.) But how would an AI know that one sound is a jackhammer and another is car horn? SONYC’s answer: citizen science. Volunteers are asked to identify and label 10-second sound clips, leading to a machine-learning understanding of the various noises assaulting urban dwellers.
A mere 17 years after the need for it was identified, someone is finally going to design the definitive secure, open-source voting system. DARPA has issued a $10 million contract to Galois, an Oregon-based longtime government contractor. (Thorvaldur Gylfason, Washington, D.C., May 2012; Peter Neumann, Avi Rubin, Austin, Feb 2004)
Computer scientists at UC Davis, Ireland’s Maynooth University, and CalTech have teamed up in the wet lab to compute with DNA (Ehud Shapiro, Madrid, Jul 2003; Len Adelman, San Diego, Feb 2002). Each short (42-base) strand consists of two six-bit input domains and two six-bit output domains; domains on separate strands selectively pair with complementary segments. They’ve already demonstrated the ability to run 21 distinct algorithms.
Splitting water molecules to generate hydrogen and oxygen offers a no-carbon strategy to make an important fuel (Elvis Cao, Berkeley, Mar 2019; Joseph Romm, Montreal, Apr 2004; Amory Lovins, Toronto, Apr 2002). The problem is reaction efficiency. Researchers at the University of Arkansas are finding that iron–nickel nanocatalysts might provide the key to industrially viable water electrolysis.
Was Facebook’s day-long outage last week really just a “server configuration change” or an inevitable consequence of its move-fast-and-break-things mindset? (Roger McNamee and Jonathan Taplin, Los Angeles, Mar 2018; Jonathan Taplin, Boston, Apr 2017; Judith Estrin, Berkeley, Mar 2019; Sean Parker, San Francisco, Feb 2010)
By now, you likely know that we love all things space-related (e.g., Jeremiah Pate, Berkeley, Mar 2019; Jordi Puig-Suari, Boston, Apr 2017; Erika DeBenedictis, Anne Miglarese, Vienna, Jul 2013; Daniel Clancy, Phoenix, Dec 2003; Lynn Rothschild, San Jose, Sep 2000). This week President Trump announced an agreement with Brazil that allows U.S. companies to conduct space launches from Brazil. While Alcantara isn’t quite ready for prime time yet, Jim Cantrell (Berkeley, Mar 2019) was among those interested in the development for reasons of the fuel efficiency.
If this has whet your appetite for space, you can satisfy it right now, provided you opened this newsletter as soon as it hit your inbox. NASA is airing today’s spacewalk outside the International Space Station. If you missed it, catch the walks scheduled for Mar 29 and Apr 8. (Michael Schrage, Dallas, Dec 1999)
How did we miss this news? Last week, on Pi Day, Google announced one of its researchers has taken pi to 31.4 trillion digits, using 25 Google Cloud virtual machines (Ike Nassi, Washington D.C., Sep 2014; Michael Groner, Vishal Sikka and David Reed, Rich Wolski, San Diego, Feb 2009) over the course of four months. As the researcher, Emma Haruka Iwao, noted, only the first hundred or so are useful, but it’s still nice that we now know the next 31,415,926,535,797 as well, isn’t it?
"I finally convinced my cardiologist to look at data, and now he’s at Google."