HNHacker News
TopNewBestAskShowJobs

ihnorton

1,471 karma · joined October 23, 2011

Software engineer at TileDB

isaiah plus hnhiring -at- tiledb -dot- com

submissionscomments
ihnorton··on Medical Science Has a Data Problem
There have been some changes [1] in recent years attempting to address this. Since 2007, all drug, biologic, and device trials in the US must be registered with ClinicalTrials.gov after phase 1 (which, at least in principle, addresses safety only - not efficacy), and "results" must be reported. There could certainly be some loopholes or limitations on the data release that I am unaware of.

However, it is true that most of the controversial, blockbuster drugs that are discussed in the original article were approved long before these requirements came into effect.

[1] http://clinicaltrials.gov/ct2/manage-recs/fdaaa

ihnorton··on Making Ruby Faster
Julia is really designed to feel dynamic while minimizing the run-time dynamicity the compiler must account for. Type inference and method specialization play a big role, and you can read the Julia papers (on julialang.org) for a discussion of many of the design choices.

Many modern JIT compilers use multiple techniques for aggressive specialization of code sections that meet optimization heuristics (whether tracing-based or otherwise), thus carefully written JavaScript (under V8 and others), Lua (under LuaJIT) and Python (under PyPy and others) can be quite fast. On these platforms there are usually implicit or explicit language subsets and coding styles required to get maximum performance out of the compiler (as with Julia: it is possible to write slow code in any language). For example, ASM.js is an optimization-friendly, explicit subset of JavaScript.

ihnorton··on The MBAs are fleeing, should SF be worried?
And what do all of those have in common? Strong unions, or legally-sanctioned licensing boards.
ihnorton··on Why Silicon Valley Works
Before you do that: become very familiar with the concept and consequences of RVUs (and, ideally, ask an active GP).
ihnorton··on The Berlin Wall fell 25 years ago, but Germany is still divided
President Obama made a major, iconically-staged speech in Cairo in 2009 regarding relations between Muslim-majority countries and the NATO (approximately) West:

http://en.wikipedia.org/wiki/A_New_Beginning

ihnorton··on Skunk Works Reveals Compact Fusion Reactor Details
For what it's worth, in the Google [X] talk last year, they briefly mentioned running actual plasma confinement experiments (although I guess not reaching fusion conditions or that would probably have been publicized).

https://www.youtube.com/watch?v=JAsRFVbcyUY

ihnorton··on Skunk Works Reveals Compact Fusion Reactor Details
Content-free article, but presumably it is this:

http://en.wikipedia.org/wiki/High_beta_fusion_reactor

This is not the first time they have gone public with this - Charles Chase gave a talk at Google X last year, recorded and publicly-available:

https://www.youtube.com/watch?v=JAsRFVbcyUY

ihnorton··on Can Celiac Disease Affect the Brain?
One of the quoted scientists, Dr. Fasano, discovered a protein called zonulin [1] responsible for mediating the permeability of cells in the gut, with possible implications for gut-involved autoimmune diseases. He authored a fascinating (so far, it's long!) review article on the subject, freely accessible [2].

The original article hints at one potential etiology for gluten's neurological effects:

antibodies directed at gliadin, a gluten protein, could cross-react with proteins in the brain. The finding suggests that, by some coincidence, certain proteins on neurons structurally resemble proteins in wheat. Meaning that, if your immune system attacks gluten, it might also inadvertently pursue brain tissue.

Interestingly, gliadin also activates the zonulin pathway. There is drug trial underway to target that pathway, now reportedly past the phase 2 (efficacy and safety) stage in 300 patients [3].

Another possible neurological tie-in is myelin. Myelin is a lipid sheath surrounding axons, very roughly analogous to cable insulation. De-myelination is involved in several autoimmune neurological diseases, including multiple-sclerosis. There are several proposed mechanisms referenced in [2], and recent work has suggested at least a weak association between celiac disease and multiple-sclerosis [4]. Since the original article also mentions schizophrenia, it is worth pointing out that de-myelination has also been suggested as a factor in some schizophrenic pathologies.

[1] http://en.wikipedia.org/wiki/Zonulin [2] http://physrev.physiology.org/content/91/1/151.long [3] http://www.prnewswire.com/news-releases/alba-therapeutics-an... [4] http://www.biomedcentral.com/1471-2377/11/31

ihnorton··on Box acquires MedXT
Interesting move, but: this, folks, is why hospitals buy from $BIGCORP (Epic, GE, Siemens).

edit: to clarify, by "this" I mean "stability". One path to build a medical software company has been to start by targeting individual practices, clinics, and small hospitals. That is still possible, and there are plenty of companies doing it. But that path is getting harder to take in the US because of medical industry consolidation. Even "independent" practices now usually have referring affiliations, and more often than not this means that they use the hospital network's existing software infrastructure.

ihnorton··on Help improve gcc
DragonEgg is an LLVM backend for GCC: http://dragonegg.llvm.org/

There was some discussion of joining forces in the early days of LLVM (ca. 2005 I believe), but for various reasons that did not go forward. Licensing is indeed a big issue: GCC has had the explicit goal of making proprietary plugins technically infeasible, which leads to avoidance of some modularization/decoupling opportunities.

ihnorton··on Are Ivy League schools overrated?
That is a great read. It has been posted several times but didn't get much traction, unfortunately.

> Admissions to Ivy league schools is not as objective as it should be. There is too much focus on fluff and not ability/potential (e.g., meaningless "volunteer work").

To expand on this a little bit: he describes the current admissions systems as "eye-of-newt-wing-of-bat mysticism that jerks teenagers and their moms around and conceals unknown mischief" and he laments that "test-based selection used to be the enlightened policy among liberals and progressives, since it can level a hereditary caste system by favoring the Jenny Cavilleris (poor and smart) over the Oliver Barretts (rich and stupid)."

(aside: I don't understand why the parent story was modded into oblivion over the course of 10 minutes)

ihnorton··on Simple Language Implementation Techniques for the 21st Century
It really depends what libraries you need. You might be interested in: http://pkg.julialang.org/pulse.html (also a searchable package list).
ihnorton··on Simple Language Implementation Techniques for the 21st Century
This paper was an interesting read! And thank you for the exceptionally well-selected references: I think I will enjoy reading several of those.

I'm not sure about the others, but as far as Julia goes, the main parts of the language implementation are: parser and lowering in about 5000 lines of Scheme; type inference is 3k lines of Julia; and codegen to LLVM is about 6000 SLOC of C++. (and then there are a few tens of thousands of LOC of runtime and library code in C and Julia). I suspect it would be possible to implement a nice, smallish, LLVM-backed DSL using the Ocaml bindings in well under 10k lines, but I am likewise unaware of such an experiment. On the other hand, implementing Julia in Truffle or RPython would be a neat project.

ihnorton··on Juno: A free environment for the Julia language
Yeah, it was exciting. The Julia ecosystem is growing quickly. A few months ago I could more-or-less keep up with all of the issues on most of the important packages, but doing so now feels impossible. A lot of new packages are still at the "it works for me" stage, but there is an amazing range of things in development and maturity will come with time.

Beyond the main LightTable presentation and the generative art you mentioned, there were also lightning talks on:

- calling JITed C++ from the Julia REPL with Clang [1].

- spectrograms, audio playback, and tone generation with interactive manipulation from IJulia [2].

- Kalman filters [3]

[1] https://github.com/Keno/Cxx.jl [2] https://github.com/ssfrr/AudioIO.jl [3] https://github.com/wkearn/Kalman.jl

ihnorton··on Juno: A free environment for the Julia language
Unfortunately not. There was a Mathematica-in-Julia-in-LightTable cameo which you might have found amusing. Maybe we can record in the future, but having everything be informal and low pressure gave it a really fun feeling.
ihnorton··on Juno: A free environment for the Julia language
Mike Innes gave a talk/demo of this at the Boston/Cambridge Julia meetup [1] last night and it was really impressive. Some highlights:

- interactive sliders to manipulate input parameters

- support for inline plots (like IJulia)

- very nice code completion

[1] http://www.meetup.com/julia-cajun/

ihnorton··on PNNL Smartphone Microscope
Some previous discussions of similar techniques (from different groups):

https://news.ycombinator.com/item?id=7591573

https://news.ycombinator.com/item?id=8129934

ihnorton··on Artificial sweeteners linked to glucose intolerance
The paper contains some preliminary human data as well. From the New Scientist writeup:

But can the results in mice be extrapolated to humans? To find out, the team examined data from 381 people. They found an association between glucose intolerance and general sweetener use.

and:

To explore this, the team asked seven healthy people who don't normally consume sweeteners to eat the FDA's maximum daily allowance of saccharin. ... By day five, four of the seven people had a significant decrease in their glucose tolerance, while three saw no change. Sequencing showed that those who responded to the sweetener started out with different gut bacteria to those who didn't respond. What's more, the gut bacteria of the four responders changed significantly after consuming sweeteners, while the non-responders' barely changed.

ihnorton··on Minix 3.3.0
> i felt that all Microsoft had to do for twenty years, was expose functions form the NT kernel, to stay ahead in terms of features.

I got some hints of that feeling while poking around http://undocumented.ntinternals.net/ ... They seem to have still forgotten about ZwCreateProcess (copy-on-write forking).

ihnorton··on SIMD Vectorization in Julia
Shared-memory multiprocessing is supported in the latest release on linux and mac. Multi-threading is in development and has seen some significant progress over the past few months.

> But I'm sure, it's on the right track.

Yes, and it is worth pointing out that both this article and the @simd feature were written by Arch Robison - who is also the architect of Intel's Threading Building Blocks.

ihnorton··on The Mathematics of Ebola Trigger Stark Warnings
Doctors Without Borders/Médecins Sans Frontières (MSF) is on the ground and is one of the most effective ways to directly support efforts to aid Ebola victims and stop the spread (disclosure: I just donated):

https://donate.doctorswithoutborders.org

The health systems in these countries were already overwhelmed and under-resourced long before Ebola broke out, and the international governmental response has so far been alarmingly tepid.

Recent AmA with an MSF doctor:

http://www.reddit.com/r/IAmA/comments/2g79ip/i_work_for_doct...

Note that MSF has an excellent efficiency rating (~87% of revenue goes to program expenses, 11% to fundraising):

http://www.charitynavigator.org/index.cfm?bay=search.summary...

ihnorton··on Playboy Interview: Steve Jobs (1985)
On the other hand, how long did it take Microsoft to figure that out - remember the Windows 95 TCP/IP stack?

edit: note, Windows 95 originally didn't have a TCP/IP stack! The point is, plenty of smart people missed this boat for years, and I don't think it was at all obvious in 1985 that getting on a network would be the driving use case for widespread computer adoption by "non-techie" people.

ihnorton··on We’re All Nerds Now
As jacalata pointed out, President Obama could be cited as a counter-example, and under the same criteria I would add at least Wilson to the list.

Being perceived as an intellectual is not strictly exclusionary, but opponents will certainly try to hold it against a candidate - much like any other characteristic that can be played up for factionalism.

ihnorton··on Peter Thiel AMA
By productivity, I certainly did not intend to imply volume! Perhaps impact would have been better. Although I'll leave aside the philosophical question of historical inevitability vs. singular genius :)

Do you think that there is a widespread failure by the scientific community to identify and support superstars?

ihnorton··on Peter Thiel AMA
In physics at least, isn't that because most of the low-hanging (low-energy..) fruit is gone? CMS or Atlas at the Large Hadron Collider are feats of engineering, not to mention the physics, and doing all of that takes a large organization. In theoretical physics, people seem to still be doing reasonably well with the lone-genius thing (although I suppose Peter Woit disagrees).

Biology is interesting because small, high-impact groups become arguably more sustainable every year due to cheaper, smaller instruments and the rise of contract services (e.g. DNA sequencing). On the other hand, many of the human problems biology is attempting to address are hard to inferentially study on a small-sample level, again necessitating large organizations.

ihnorton··on Peter Thiel AMA
Thinking about this, I wonder: has anyone tried to systematically measure "eccentricity" vs. scientific productivity?

Also, I wonder if there a risk that this characterization just perpetuates the pointy-egghead stereotype - that good scientists don't have social skills? The discussion of "holistic" evaluation in Steven Pinker's recent essay on Harvard admission comes to mind [1].

Perhaps the funding agencies need a separate, high-prestige "individual contributor" track similar to corporate Principal Engineer or fellowship tracks, to allow non-political but brilliant scientists to work hands-on and avoid the grant scramble?

HHMI tries to provide something like this already, giving longer grants to early-career faculty to reduce grant pressure, and I guess NSF and NIH have also been trying to do this with CAREER awards and New Innovator grants...

[1] Very interesting piece: https://news.ycombinator.com/item?id=8277941

ihnorton··on Peter Thiel AMA
Interesting point on science funding practices:

I think there's been a Gresham's Law in science funding in this country, as the political people who are nimble in the art of writing government grants have gradually displaced the eccentric and idiosyncratic people who typically make the best scientists. The eccentric university professor is a species that is going extinct fast.

ihnorton··on CLion, the new C/C++ IDE from JetBrains
Try these one-click installers; the experience is much better.

mingw-builds to install gcc:

http://sourceforge.net/projects/mingwbuilds/

msys2 to install the rest of the build environment:

http://msys2.github.io/

(as well as a growing number of packages, managed with a port of pacman from Arch)

ihnorton··on Many in West Africa Are Thought to Be Immune, but Finding Them Will Not Be Easy
The hypothesis is that immunity is conferred by the presence of ebola antibodies developed after very-low-dose exposure from eating fruits visited by infected bats.

> His teams took 4,349 blood samples in 220 randomly selected villages. They found that 15 percent of Gabon’s population had antibodies. But it varied widely: near the coast, only 3 percent did; in some jungle villages near the Congo border, up to 34 percent did.

>Also, their antibody levels varied widely, and what level is protective is roughly known for lab monkeys, but not for humans.

ihnorton··on Rhine – A Lisp on LLVM
Julia has some nice functionality for introspecting both lowered and type-inferred ASTs, the corresponding LLVM IR, and finally the resulting assembly code. (several Lisp systems have this kind of thing too)
← PreviousPage 8 of 15Next →