Profile

Cover photo
Ethan Miller
Works at Pure Storage
Attended University of California, Berkeley
263 followers|6,196 views
AboutPostsPhotos+1's

Stream

Ethan Miller

Shared publicly  - 
 
Age is only a curse if you don't keep up with advances in the field.  If your reaction to Python, Ruby, and Go is "I don't want to learn a new language - C/Java/Fortran is enough", and your reaction to cloud computing is "mainframes will never die", you deserve what you (don't) get.  But if you continue to acquire new skills and new knowledge, the combination of your background and experience will generate efficiency and quality that'll trump youthful enthusiasm.
1
Ethan Miller's profile photoDirk Grunwald's profile photo
3 comments
 
That might be, but there's a historical tacit agreement of income stability in salaried careers such as tech work that allows people to e.g. plan for families. This story is about how that's changing rapidly.

For example, faculty are normally judged and rewarded based on grants, publications, student production, etc. If you have an "off year", your unit doesn't cut your salary and hand it to the young hot-shot who had a better production record for that year than you did. You're not doing piece work. Is that for "experience"? What if that "experience" doesn't translate into production quickly? A pure libertarian would say to reward the young person instead.  That may happen over the long term through lay-offs, but if skills are the only thing of value (as opposed to skills+experience), the cycle for adjusting pay for output would be much shorter (e.g. more like piece-work).

If that's the new contract, then tech piece-work should be priced higher, to allow offsetting future down-times when you can't work at the pace or level of your late 20's.

To state that young people are "smarter" because they eschew families, cars or anything other than work is getting it precisely backwards or incredibly naive.
Add a comment...

Ethan Miller

Shared publicly  - 
 
It occurs to me that perhaps the best way to prevent hijackers from taking over airplanes is not to ban everything that could be used as a weapon (there's a lot of things that could be used to do that), but rather to provision each plane with several rolls of duct tape.  If a passenger acts up, the crew and other passengers will subdue him/her (it seems to happen on a regular basis these days), and the duct tape can be used to ensure that s/he remains firmly secured to the seat.

Duct tape is known for being strong enough to restrain people (think of all the "duct tape someone to a goalpost" pranks), and it's cheap.  Plus, you can't actually damage an airplane with it, other than the adhesive not coming off easily.  Of course, there's the chance that a restrained individual might have a potty accident if it's a long flight, but, hey, you shouldn't have acted up in the first place.

If we do this, perhaps we can be allowed to carry itty-bitty (2") pocketknives again.  Please?
1
H. Peter Anvin's profile photoEthan Miller's profile photoDavid Andersen's profile photo
4 comments
 
I'll yield to your clear experience edge in the matter of tying people to seats.  <big grin>  (But yes, that does make sense.)
Add a comment...

Ethan Miller

Shared publicly  - 
 
StreamScale has bullied USENIX into (temporarily) removing the paper that Jim Plank, Kevin Greenan, and I published at FAST 2013 (see http://www.prweb.com/releases/2013/7/prweb10954275.htm for StreamScale's press release).  They claim that, in 2011, they were the first to figure out how to use Intel SIMD instructions for Galois field arithmetic.  They're wrong.

In May 2009, H. Peter Anvin published a white paper on RAID6 (http://web.archive.org/web/20090807060018/http://www.kernel.org/pub/linux/kernel/people/hpa/raid6.pdf) that describes how to do 8-bit GF multiplies using PSHUFB (see page 8).  Our FAST paper cited Anvin's pioneering work, since it inspired some of the material in our FAST13 paper (and I have proof that I had a working version implemented in Sept 2010).

For those who are looking for our FAST13 paper, you can still download it from the SSRC web page (http://www.ssrc.ucsc.edu/Papers/plank-fast13.pdf).  Source code is still available from https://bitbucket.org/ethanmiller/gf-complete.  Feel free to use either the paper or the library with no fear of infringing StreamScale's IP with respect to the use of SSE instructions for Galois field arithmetic (the only thing the press release claims we "stole").  Anvin's paper publicly disclosed the basic technique of using PSHUFB (the only relevant technique in our gf-complete library) more than two years before StreamScale disclosed their approach to anyone.  If you're still worried, simply copy the inner loop code for GF multiply from Anvin's 2009 paper—it's nearly identical to what's in gf-complete.

StreamScale should be ashamed of themselves.  Companies that use unsubstantiated claims to coerce publishers to remove academic papers should be soundly punished when their claims are shown to be false.
8
4
faried nawaz's profile photoZoran Dimitrijevic's profile photoThouis Jones's profile photoDavid Andersen's profile photo
12 comments
 
...and of course, apologies, pshufb was ssse3 not sse3
Add a comment...

Ethan Miller

Shared publicly  - 
 
Amazing video of surfing at Steamer Lane in Santa Cruz, shot from a quadcopter "drone".  Really brings a new perspective to a scene I've watched many times.
1
Add a comment...

Ethan Miller

Shared publicly  - 
 
Louis Farrakhan was able to speak in Berkeley on Saturday, despite his racist and anti-Semitic diatribes. He was invited by the Black Student Union. He was denounced by many student groups and an online petition, yet he was allowed to give his speech without disruption by protesters.

Benjamin Netanyahu was unable to speak in Berkeley on November 28th (see http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2000/11/29/MN44311.DTL). Hundreds of protesters blocked the entrance to where he was to speak, forcing the cancellation of his lecture. One said "I don't believe in free speech for war criminals."

It seems that the folks in Berkeley believe in free speech for virulent racists and anti-Semites, but not for the former leader of a democratic nation with whose policies they disagree. For all their bluster about freedom and rights, "liberals" (at least in Berkeley) are only willing to give those rights to people with whom they agree. Shame on them.
1
Add a comment...
Have him in circles
263 people
Nitin Nagpal's profile photo
Fred Douglis's profile photo
Wes Mertes's profile photo
Mike Chudzik's profile photo
Tim Finin's profile photo

Ethan Miller

Shared publicly  - 
 
I don't use pot myself, but I did try it a few times in college, though. (And, no, I didn't inhale - I couldn't hold the smoke in my lungs - but there's more than one way to get it into your system.  Mmm, brownies...). 

Having said that, I completely agree with this article.  I see no reason that we continue to demonize a substance that's less addictive and less harmful than tobacco, and that has numerous medical uses.  Why not legalize weed?  If we still feel it necessary, we can show ads about its "negative" effects, just as we do for tobacco.
2
David Fulton's profile photoDan Wallach's profile photo
2 comments
 
Having read the article. If we set all unfully tested substances as a class one prohibited drug, the drug companies and the chemical companies would be howling. I see no reason that it should not be part of the range of prescription options. Sadly because there is little money in plant grown drugs big medicine has no return on researching it.

Personally I have never found pleasure with the stuff when I have had the misfortune to walk in to its smoke. Headaches that knock me out for the rest of the day are no fun.
Add a comment...

Ethan Miller

Shared publicly  - 
 
In response to yesterday's debacle of an AMBER alert, I turned them off on my iPhone.  The first alert came at 1:31AM, and, despite my "Do Not Disturb" setting, blared the 3-tone emergency alert sound at full volume.  Since I was in bed and nowhere near a road, waking me to tell me to be on the lookout for a blue Nissan Versa did no good.  (And, for those who say to move the phone, I use it as my alarm clock, as do many people.)

The next 7 (!) alerts came during the day, all accompanied by the same tones, and all with the same description and license plate number as the first.  No new information, just a jarring sound that caused me to jerk the steering wheel during the alert that came when I was in the car.  Fortunately, I was in the Baskin parking lot when that happened, so I wasn't going fast and didn't hit anything.

I have no objection to AMBER alerts being sent to many people, including me.  But the "emergency" sound is unnecessary, and the alerts should respect the "do not disturb" setting on the phone.  If I'm asleep or giving a talk, I won't be looking for that car.  If I'm driving, do you really want me looking at the phone for the description?  If this were an imminent threat to everyone (severe weather warning, earthquake, etc.), that'd be different (you might want people to pull over and seek shelter), but a message about which license plate to look out for isn't so high priority that I need to look at it RIGHT NOW.  This is doubly true when it's been 15+ hours since the first alert and the driver could be anywhere in a 900 mile radius.  Besides, the description and license plate number were on every highway sign in the Bay Area, and those I could read while driving.

Sadly, all this will do is encourage a lot of people to completely disable AMBER alerts when all they want to do is stop them from making so much (uncontrollable) loud noise.
1
Avani Gadani's profile photoDavid Fulton's profile photo
2 comments
 
We have a campus alert system. Such systems need to be used well and appropriately as you commented. 
Add a comment...

Ethan Miller

Shared publicly  - 
 
OK, so this is an old article on MOOCs.  I read it over, and it's an interesting discussion of online courses.

I have one important question, though: the article mentions a "15-year old student in Mongolia" who "received a perfect score on the final exam" and is now applying to MIT and UC Berkeley.  Why?  If a MOOC is as good as the "real thing", why bother going halfway around the world and paying a lot of money (or, more likely, in this case, getting a lot of scholarships) for the "same thing"?  Either this student is a fool (which I doubt) or a MOOC from Berkeley isn't as good, for some reason, as being there in person.
1
1
Robert Rich, Ph.D.'s profile photo
Add a comment...

Ethan Miller

Shared publicly  - 
 
I'm astounded (but happy) to see that CNN published this opinion piece, which points out that Europe tolerates a lot of anti-Semitism, especially as a reaction to issues with the Palestinians.
"At times, human rights activists seem to have no problem with anti-Semitism -- even of the genocidal variety -- condemning it forcefully only if it is accompanied by anti-immigrant or anti-Muslim sentiment."
2
Add a comment...

Ethan Miller

Shared publicly  - 
 
This guy made a video in 1994 showcasing a prototype that looks a lot like an iPad. So what? I know of video made in 1988 showing the iPad concept, and it was broadcast widely on TV. Star Trek: TNG had devices that looked a lot like the iPad too. Sure, they didn't work, but getting the device to work is a matter of technology catching up with fiction.
2
Add a comment...
People
Have him in circles
263 people
Nitin Nagpal's profile photo
Fred Douglis's profile photo
Wes Mertes's profile photo
Mike Chudzik's profile photo
Tim Finin's profile photo
Work
Occupation
Computer scientist & professor
Employment
  • Pure Storage
    Research Scientist, 2009 - present
  • University of California, Santa Cruz
    Professor, 2000 - present
  • University of Maryland, Baltimore County
    Asst. Professor, 1994 - 2000
Basic Information
Gender
Male
Story
Introduction
I'm a professor of computer science at the University of California, Santa Cruz.  I'm also a consultant at Pure Storage, a startup in Mountain View.
Education
  • University of California, Berkeley
    computer science, 1988 - 1995
  • Brown University
    computer science, 1983 - 1987
Ethan Miller's +1's are the things they like, agree with, or want to recommend.
Santa Cruz to be split into congressional districts; Dividing line weave...
www.santacruzsentinel.com

The line ambles through downtown Santa Cruz like a drunken party guest, sort of following the river, stumbling between familiar friends and

Center for Research in Intelligent Storage
www.cris.soe.ucsc.edu

The growth rate of data has been rapidly increasing over the past decade. This growth has pushed data quantities to a point where users and

Transactional memory going mainstream with Intel Haswell
arstechnica.com

Transactional memory is a promising technique for making the development of high-performance parallel programs easier and safer. It has larg

U.S. trade policy: America's unequal trade relationship with China - Los...
articles.latimes.com

The American economy has been in trouble for more than a decade, and no amount of right-wing tax cuts or left-wing fiscal stimuli will solve

The Signal: Digital Preservation
blogs.loc.gov

Digitization is Different than Digital Preservation: Help Prevent Digital Orphans! July 15, 2011 by Bill LeFurgy. USB-drive Graveyard, by Be

How Much Data Will Humans Create &amp; Store This Year? [INFOGRAPHIC]
mashable.com

How much is 1.8 zettabytes? We all know it's a big number, but just how big? The mind-blowing comparisons in this infographic will put it

A History of College Grade Inflation
economix.blogs.nytimes.com

A new report on college grade inflation finds that about 43 percent of all letter grades given are A's, an increase of 28 percentage points

Letter to UC President Mark Yudof from the Jewish Community Survey
www.surveymonkey.com

1. Dear President Yudof, We are Jewish community members of California – rabbis, principals, teachers, alumni, parents, grandparents, family

Archival Storage @ SSRC
www.ssrc.ucsc.edu

Archival Storage. Scalable File System Indexing. Deduplication. HECURA: Scalable Data Management. Petabyte-Scale Storage. Reliable Storage.