Top all-time · #6164
Justin Domke
Personal blog · justindomke.wordpress.comaidatamath
Justin Domke's personal blog at justindomke.wordpress.com covers ai, data and math. Since 2008, 13 different Hacker News readers have linked it 30 times, most often “Automatic Differentiation: The most underused tool in the machine learning toolbox?”. The last post was in 2024, but the archive is still widely read.
Too few sampled comments (7) to say how often HN recommends it.?
Why HN links it
HN readers cite Justin Domke's posts as sources on compiler benchmarks (Stalin Scheme vs C, Julia vs C) and on automatic differentiation. The comments link them as evidence rather than judging the writing.
Summarised by Claude from a sample of 7 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Most-linked posts?
- Automatic Differentiation: The most underused tool in the machine learning toolbox?3 people · 11 mentions · first linked 2009-02-19
- The Stalin Compiler (2009)3 people · 3 mentions · first linked 2015-10-29
- Automatic Differentiation (2009)2 people · 5 mentions · first linked 2015-07-15
- Statistics – the rules of the game2 people · 2 mentions · first linked 2018-06-13
- /2009/02/23/the-stalin-compiler/2 people · 2 mentions · first linked 2013-08-26
What Hacker News says?
There is a Scheme compiler called Stalin[1] which apparently does some very sophisticated full-program and lifetime analysis. This of course leads to very long compile times, but the code is also very fast compared to other Schemes, and even C in some cases, from the benchmarks I've seen[2][3] etc. You can find more benchmarks for it, and similar compilers like MLton. Of course, I can't comment on whether or not it…
Comment linking /2009/02/23/the-stalin-compiler/
on “Random Acts of Optimization”Submitted The Human Regression Ensemble as a story · 53 points
HN discussionI've gotten pretty good results from PGO as well as JIT optimizations (which isn't conceptually much different), but even if I hadn't, your comment is still a fallacy on two accounts: 1) JITs were slow until they weren't. Poor results from existing experimental PGO compilers do not prove that the concept is flawed. 2) Profiling is a heuristic substitute for a real cost model. The most advanced optimizing compilers f…
Comment linking /2009/02/23/the-stalin-compiler/
on “Random Acts of Optimization”Submitted Automatic Differentiation: The most underused tool in the machine learning toolbox? as a story · 50 points
HN discussion> The downside was that it took several minutes to compile <100 line sources that produce <1000 instruction programs. This sounds like the Stalin Scheme compiler. Simple benchmark (vs C (gcc), Stalin wins): http://justindomke.wordpress.com/2009/02/23/the-stalin-compi... GUI addon: http://code.google.com/p/stalingui/ http://community.schemewiki.org/?Stalin Download link: https://engineering.purdue.edu/~qobi/software.…
Comment linking /2009/02/23/the-stalin-compiler/
on “On the Cleverness of Compilers”Submitted Automatic Differentiation (2009) as a story · 43 points
HN discussion@1) Is it? [0] seems to indicate that it is usually about twice as slow as C and sometimes fares much worse. Of course, that article is over a year old by now so maybe there are improvements, but if I can get a 2x performance gain ‘just’ by writing in C/C++, I will certainly do so. [0] http://justindomke.wordpress.com/2012/09/17/julia-matlab-and...
Comment linking /2012/09/17/julia-matlab-and-c/
on “Julia, the One Programming Language to Rule Them All”Submitted Confidence Games as a story · 34 points
HN discussionAccording to this blog post from 2009[0], it was slightly faster than C compiled with the GCC and almost 15x faster than Chicken Scheme. I have tried looking for more rigorous benchmarks but have been unable to find any. [0] http://justindomke.wordpress.com/2009/02/23/the-stalin-compi...
Comment linking /2009/02/23/the-stalin-compiler/
on “Stalin: a global optimizing compiler for Scheme”Submitted Automatic Differentiation: The most criminally underused tool [in ML] (2009) as a story · 5 points
HN discussionIt's also known as "automatic differentiation" -- it's quite different from numerical/symbolic differentiation. More information here: - https://justindomke.wordpress.com/2009/02/17/automatic-diffe... - https://wiki.haskell.org/Automatic_Differentiation The key idea is extending common operators (+, -, product, /, key mathematical functions) that usually operate on _real numbers_ to tuples of real numbers (x, dx) (t…
Comment linking /2009/02/17/automatic-differentiation-the-most-criminally-underused-tool-in-the-potential-machine-learning-toolbox/
on “Calculus on Computational Graphs: Backpropagation”Submitted The Stalin Compiler (2009) as a story · 4 points
HN discussionI've long been suspicious of this study from Ariely from 2003. It suggests that if you "prime" students by having them write down the last 2 digits of their student ID, then students with numbers like "93" will then be willing to pay 3.3x as much for a bottle of wine as students with numbers like "04". At first, some follow-up research seemed to replicate this, but then later work found no effect whatsoever. Very od…
Comment linking /2021/10/14/is-anchoring-a-reliable-cognitive-bias/
on “UCLA Professor Refuses to Cover for Dan Ariely in Issue of Data Provenance”Submitted An appalling and politically impossible way to improve voting outcomes. as a story · 4 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool [in ML] (2009) as a story · 3 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool (2009) [in ML] as a story · 2 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool (2009) [in ML] as a story · 2 points
HN discussionSubmitted Statistics – The Rules of the Game as a story · 2 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool [in ML] (2009) as a story · 2 points
HN discussionSubmitted Statistics – the rules of the game as a story · 2 points
HN discussionSubmitted Why does regularization work? (2018) as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool (2009) [in ML] as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool (2009) [in ML] as a story · 1 points
HN discussionSubmitted Sneaking up on Bayesian Inference (A fable in four acts) (2016) as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool (2009) [in ML] as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most underused tool in machine learning? (2009) as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool [in ML] (2009) as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool? as a story · 1 points
HN discussionSubmitted Automatic Differentiation: The most criminally underused tool [in ML] as a story · 1 points
HN discussionSubmitted Automatic Differentiation as a story · 1 points
HN discussion
