Закреплённый твит
Raman Shah
4.4K posts

Raman Shah
@rs_datascience
I improve operations by measuring them.
Providence, RI, USA Присоединился Mart 2019
890 Подписки510 Подписчики

@lauriewired @corewarrior This promtped a very brief literature search. Prompted and "ahh gotcha"/"ugh blech" - lead-210 with its 22.2 year half-life. Wonder if you knew that lead-210 is also responsible for a fair % of the lung-cancer burden of smoking.
English

@corewarrior it wasn't accidentally contaminated, in fact, IBM actually used low alpha solder! it wasn't enough though, and subsequent systems used ULA solder (ultra-low-alpha solder)
English

How do you kill a Supercomputer?
(Accidentally) using radioactive solder is a good way.
IBM’s Blue Gene/L frequently crashed when running simulations at LLNL.
Turned out that alpha particles from the lead solder in the board carrier were slamming the L1 cache with bit flips. It’s a particularly nasty issue. Unlike DRAM /w ECC, there wasn’t a transparent way to correct it.
The fix was brutal. IBM recommended scientists reprogram critical workloads to write-through mode.
Basically, every store in L1 would also travel down the cache hierarchy immediately. If L1 got corrupted, the kernel would invalidate the whole cacheline and force a refill from slower pools.
It worked…but had a massive performance hit. As much as 20-30%!


English

@mayukh_panja @rahulj51 I'd counter, based on the reports of my many friends involved in the rubber-meets-the-road mundane operations of the world, that the world is, indeed, coming to a standstill. It's quite frightening.
English

One thing academia does extremely well and startups and companies massively screw up is hiring. Hear me out!
A friend of mine, PhD in Astrophysics, solved a tough problem for their PhD: when light from stars travels through the Earth's atmosphere, the turbulence and density fluctuations cause the light rays to become "squiggly" instead of straight and the resulting image you get from a telescope becomes blurry.
So he had to model atmospheric turbulence and then write a piece of software in C++ that inverts this problem to get de-blurred images. This involved understanding physics, maths, computation, a bit of ML and writing production-level code in C++.
When he tried to look for an industry job he simply couldn't find any. It was also hard to just get interviews.
The first problem is that recruiters, who are often deeply non-technical, look for specific keywords in CVs and they just don't know how to parse a non-standard CV. This is a guaranteed way of missing out on outlier candidates.
Second, a lot of hiring managers over index on niche knowledge about a specific tool/framework/language and the ability to remember syntax off the top of your head. A solid researcher sees programming languages, machine learning, physics, maths etc as tools that are at their disposal and may not know/remember very specific information or every little detail about arbitrary technical things.
The whole process essentially becomes a lottery.
This was how we hired at our Max Planck Institute: the candidate would be given a paper a week before the interview and the interviewer and the candidate would discuss it together. A second interview would entail asking the candidate about THEIR past work and checking if they deeply understood what they did.
This interview format doesn't require the candidate to memorizes stuff beforehand and is pretty much independent of the whims, fancies and "taste" of the hiring manager.
A lot of stuff is wrong with academia but this is an area where they do much much better than startups/companies.
English

@JEBistline Oo did you read @rlmcelreath Statistical Rethinking? 😀
English

@McMasterDaily You got me for thirty seconds there. Well played.
English

@stevesi Oooo this winter we ran into a contraband lightbulb issue when the overhead light went out in our kitchen. Non-contraband substitute did not fit and I ended up having to change out the ceiling fan entirely.
English

@AlexKontorovich @3rdMoment I’ve heard Caltech has also gone this away after generations of bona fide honor code. Irredeemable shark jumpage.
English

Very sad indeed. When I moved from Princeton to Columbia for grad school, I was *shocked* when I was told that I actually had to proctor my calc exams, and just not trust their "honor code".
At Princeton, the honor code genuinely meant something. In the first week of freshman year, we had to write an essay explaining in detail what the consequences of cheating were, why it didn't serve our long term interests, and how even if we weren't the ones cheating, if we knew that others were and didn't report it, we would be just as guilty. (I remember vividly, because my first attempt at such an essay was rejected as insufficiently detailed! I had to write a much longer version.) As a result, people really didn't cheat (as far as I knew; every year there was ~1 student kicked out of school for cheating).
It was something really special, it turned out; that's not how it works elsewhere. Sad to see further deterioration of the culture at Princeton.
Steve McGuire@sfmcguire79
Exams at Princeton have been unproctored under an Honor Code since 1893. “Students pledge both to refrain from infractions of academic dishonesty and to report any breaches of the Constitution they witness.” But AI has led to an increase in academic dishonesty cases, so:
English

@AngelicaOung I often feel this way about how I played violin at 16.
English

In college I had a friend who figure skated competitively until she was 15 or something.
It was like she coulda been a contender when she was 14 but knew as soon as she turned 15 that it wasn’t going to be her. All of a sudden the thing that was her life became not a thing.
She would watch the fuzzy VHS tapes of herself executing super clean doubles on the ice. And she’ll say to me wistfully, yet with a sense of wonder…I’ll never ever be as good at anything else in life, ever again…
jean’s glucose monitor@mirrenelle
i found out she exists, her name is midori ito and she’s 51
English

@Foil_Island @solarsystern You summer in Nantucket; I winter in despair. Potato potahto.
English

@solarsystern Wait until you start dealing with people who refer to seasons as verbs
English

@dieworkwear Hahaha for a while I had a little hobby of recommending your “The Springboard Wardrobe” on a looksmaxxing subreddit using an anon account. It was genuinely rewarding to convince youngsters that what they were born with was fine and had opportunities to style themselves better.
English

@llm_bender @fute_nukem This is like the one legitimate application of a LLM agent I can think of 😂
English

@fute_nukem Ha! But I don't want to engage with him or he'll never stop bothering me
English

@TheCatholicEngr Not a chem eng but a Ph.D. chemist. These fields do not have particularly many jobs in the most developed countries. And even more critically, graduate programs are not sized to train a workforce, they ARE the workforce. That is to say, overproduction of talent is rampant.
English

@McMasterDaily Obviously as a 12 year former Chicagoan I am partial to the city. But Elmhurst is quite nice, with good commuter rail connectivity. Can’t help you re the work visas tho 😭
English

@lauriewired 100% agreed. The tape recorder and its modern equivalents are also vastly valuable practice aids for musicians and other performers. I used mine way too little when I was a serious violinist.
English

@reboplim As a scientist by love and former training who was extensively dragged into technology, this resonates.
I totally love how SU(2) is a double cover of SO(3); my first paper in my Ph.D. used this fact ❤️
English

When I did low-level programming, I always felt unsatisfied after learning why computers worked the way they do. Many design decisions felt like arbitrary choices made by some guy in the 1980s, and we had to deal with the ramifications.
E.g., every modern OS has to enable the A20 address line to access even-numbered MBs in memory. This is because programmers used undefined behavior in the Intel 8086, where any address above 1MB "wrapped around" to zero. On newer processors, instead of letting things break, Intel pandered to the programmers and created a latch that set the A20 line to zero by default.
I've never felt this way in physics. Every detail in the theories I've learnt felt intentional, and it always led me down rabbit holes to deeper concepts. I'm drafting a thread on one of these rabbit holes - the Bloch sphere representation of a two-level system and SU(2) being a double cover of SO(3)!
English

@CyborgLavery @_Stocko_ An interesting analog to this problem is what goes into an AVPN-certified Neapolitan style pizza (baked at ~900 F), and the dough formulation mods needed to simulate such a pizza in a cooler domestic oven.
English

@CyborgLavery @_Stocko_ I believe this is the right answer. The moisture transport needed to get the nice crispy browning sets a time scale.
English

@thoughtlesslabs @ThePrimeagen Sometimes orgs profit by producing even less, manufacturing a shortage - through either brand differentiation (Ferrari) or collusion (beef packers). 100% agree with your first paragraph but the "profit seeking = more quantity, less quality" rule does not seem ironclad.
English

i fundamentally believe it comes down to care and respect for a craft versus a desire to profit off of the craft.
there is nothing inherently wrong with wanting to profit of something, but it always leads to a decline in quality while quantity increases.
if anyone has a counter example here i'm all ears
English

i am convinced that software devs have a speed problem
they think the #1 issues is writing code faster... its not. its fixing the code that is already there to stop being utter garbage (as a garbage code connoisseur)
quality is really lacking these days, yet quantity has never been higher
English













