<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>odds &amp; ends</title><link>https://nilaykumar.github.io/</link><description>Recent content on odds &amp; ends</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 31 Dec 2025 22:59:40 -0800</lastBuildDate><atom:link href="https://nilaykumar.github.io/index.xml" rel="self" type="application/rss+xml"/><item><title>emacs</title><link>https://nilaykumar.github.io/garden/emacs/</link><pubDate>Wed, 31 Dec 2025 00:00:00 -0800</pubDate><guid>https://nilaykumar.github.io/garden/emacs/</guid><description>&lt;h2 id="installation">installation&lt;/h2>
&lt;h3 id="macos--apple-silicon">macos (apple silicon)&lt;/h3>
&lt;p>I&amp;rsquo;ve recently been having trouble with &lt;a href="https://github.com/railwaycat/homebrew-emacsmacport">emacs-mac&lt;/a>, so I&amp;rsquo;ve switched over to
jimeh&amp;rsquo;s &lt;a href="https://github.com/jimeh/emacs-builds">emacs-builds&lt;/a>, which seems to be working alright. As an added bonus, I&amp;rsquo;m
now on emacs 30, and my site seems to generate via &lt;code>ox-hugo&lt;/code> without errors (in
the past I could only get it to work on my linux desktop).&lt;/p></description></item><item><title>december adventure 2025</title><link>https://nilaykumar.github.io/garden/december-adventure/</link><pubDate>Mon, 08 Dec 2025 00:00:00 -0800</pubDate><guid>https://nilaykumar.github.io/garden/december-adventure/</guid><description>&lt;p>Found out about the December adventure on Mastodon (likely from folks over at
&lt;a href="https://merveilles.town/">merveilles.town&lt;/a>). It&amp;rsquo;s already almost a good third of the way into December, but
better late than never. Here&amp;rsquo;s a bit &lt;a href="https://eli.li/december-adventure">about it&lt;/a>, but the idea is just to pick a
project or a few, work on them, and log your progress.&lt;/p>
&lt;p>Now that December&amp;rsquo;s over, I&amp;rsquo;ve written a short &lt;a href="#retrospective">retrospective&lt;/a> at the bottom of
the page.&lt;/p>
&lt;div class="ox-hugo-table calendar-table">
&lt;table>
 &lt;thead>
 &lt;tr>
 &lt;th>日&lt;/th>
 &lt;th>月&lt;/th>
 &lt;th>火&lt;/th>
 &lt;th>水&lt;/th>
 &lt;th>木&lt;/th>
 &lt;th>金&lt;/th>
 &lt;th>土&lt;/th>
 &lt;/tr>
 &lt;/thead>
 &lt;tbody>
 &lt;tr>
 &lt;td>&lt;/td>
 &lt;td>01&lt;/td>
 &lt;td>02&lt;/td>
 &lt;td>03&lt;/td>
 &lt;td>&lt;a href="#december-4">04&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-5">05&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-6">06&lt;/a>&lt;/td>
 &lt;/tr>
 &lt;tr>
 &lt;td>&lt;a href="#december-7">07&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-8">08&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-9">09&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-10">10&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-11">11&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-12">12&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-13">13&lt;/a>&lt;/td>
 &lt;/tr>
 &lt;tr>
 &lt;td>14&lt;/td>
 &lt;td>15&lt;/td>
 &lt;td>16&lt;/td>
 &lt;td>17&lt;/td>
 &lt;td>18&lt;/td>
 &lt;td>&lt;a href="#december-19">19&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-20">20&lt;/a>&lt;/td>
 &lt;/tr>
 &lt;tr>
 &lt;td>&lt;a href="#december-21">21&lt;/a>&lt;/td>
 &lt;td>22&lt;/td>
 &lt;td>23&lt;/td>
 &lt;td>24&lt;/td>
 &lt;td>25&lt;/td>
 &lt;td>26&lt;/td>
 &lt;td>27&lt;/td>
 &lt;/tr>
 &lt;tr>
 &lt;td>28&lt;/td>
 &lt;td>29&lt;/td>
 &lt;td>&lt;a href="#december-30">30&lt;/a>&lt;/td>
 &lt;td>&lt;a href="#december-31">31&lt;/a>&lt;/td>
 &lt;td>&lt;/td>
 &lt;td>&lt;/td>
 &lt;td>&lt;/td>
 &lt;/tr>
 &lt;/tbody>
&lt;/table>
&lt;/div>
&lt;h2 id="december-1">december 1&lt;/h2>
&lt;p>Traveling back from Tokyo. I&amp;rsquo;m not very good at getting things done on the
plane.&lt;/p></description></item><item><title>innovation</title><link>https://nilaykumar.github.io/garden/innovation/</link><pubDate>Mon, 17 Mar 2025 00:00:00 -0400</pubDate><guid>https://nilaykumar.github.io/garden/innovation/</guid><description>&lt;h2 id="what-is-innovation">what is innovation?&lt;/h2>
&lt;h3 id="political-and-economic-change">political and economic change&lt;/h3>
&lt;p>Xiaowei Wang writes in &lt;em>Blockchain Chicken Farm&lt;/em>&lt;/p>
&lt;blockquote>
&lt;p>It&amp;rsquo;s not clear what exactly innovation is, but whatever it is, there is
apparently a paucity of this golden resource everywhere except Silicon Valley&amp;hellip;
The word &amp;ldquo;innovation&amp;rdquo; was derogatory in the age of monarchs, as it referred to
political and economic change that could bring down empires, threatening the
status of kings and elites.&lt;/p>
&lt;div class="attribution">
&lt;p>Wang, &lt;em>Blockchain Chicken Farm&lt;/em>, page 123&lt;/p></description></item><item><title>python</title><link>https://nilaykumar.github.io/garden/python/</link><pubDate>Wed, 26 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/python/</guid><description>&lt;h2 id="package-management">package management&lt;/h2>
&lt;h3 id="uv">uv&lt;/h3>
&lt;p>I&amp;rsquo;m using &lt;a href="https://docs.astral.sh/uv/">uv&lt;/a> to manage python versions and per-project virtual environments.&lt;/p>
&lt;h4 id="uvx-uv-tool">uvx/uv tool&lt;/h4>
&lt;p>One interesting feature of &lt;code>uv&lt;/code> is the ability to run tools without installing
them, e.g. &lt;code>uvx ruff&lt;/code> (which is just an alias for &lt;code>uv tool run ruff&lt;/code>) to run the
code format without requiring it in the project. This will install the tool into
a temporary, isolated environment and then run it. Note that this workflow
therefore requires an internet connection, and does not work offline. Instead we
can run &lt;code>uv tool install ruff&lt;/code> to remove the need to re-install the tool every
time we run it. The tool should then be visible in the &lt;code>PATH&lt;/code>, but its modules
will not be available for use in the project.&lt;/p></description></item><item><title>yojijukugo</title><link>https://nilaykumar.github.io/garden/yojijukugo/</link><pubDate>Sun, 23 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/yojijukugo/</guid><description>&lt;p>In Japanese, &lt;a href="https://en.wikipedia.org/wiki/yojijukugo">yojijukugo&lt;/a> (四字熟語) are four-character kanji compounds that
sometimes have idiomatic meanings. Here are a few I&amp;rsquo;ve learned so far that I
found interesting.&lt;/p>
&lt;ul>
&lt;li>前途多難（ぜんとたなん）: having many difficulties in store; grim prospects&lt;/li>
&lt;li>悪因悪果（あくいんあっか）: evil sowed, evil reaped&lt;/li>
&lt;li>善因善果（ぜんいんぜんか）: good actions lead to good rewards; one good turn
deserves another&lt;/li>
&lt;li>一石二鳥（いっせきにちょう）: killing two birds with one stone&lt;/li>
&lt;/ul></description></item><item><title>alienation</title><link>https://nilaykumar.github.io/garden/alienation/</link><pubDate>Tue, 18 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/alienation/</guid><description>&lt;h2 id="marx-s-definitions">marx&amp;rsquo;s definitions&lt;/h2>
&lt;h3 id="from-the-product-of-one-s-labor">from the product of one&amp;rsquo;s labor&lt;/h3>
&lt;h3 id="in-the-labor-process">in the labor process&lt;/h3>
&lt;h3 id="from-species-being">from species-being&lt;/h3>
&lt;h3 id="from-others">from others&lt;/h3></description></item><item><title>commodity fetishism</title><link>https://nilaykumar.github.io/garden/commodity-fetishism/</link><pubDate>Tue, 18 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/commodity-fetishism/</guid><description>&lt;h2 id="marx-s-definition">marx&amp;rsquo;s definition&lt;/h2>
&lt;p>Marx introduces the notion of commodity fetishism quite early in &lt;a href="https://www.marxists.org/archive/marx/works/1867-c1/ch01.htm#S4">Capital&lt;/a>.&lt;/p>
&lt;h2 id="technology">technology&lt;/h2>
&lt;p>The material objects underlying the technologies with which we interact present
themselves as commodities, products of social labor:&lt;/p>
&lt;blockquote>
&lt;p>The problem comes when we fetishize technology by only seeing the object in
front of us without considering any of the broader networks of social,
political, and economic relationships that the object is embedded within. In
other words, all the things that are behind what technology gets made, why it&amp;rsquo;s
made, where it&amp;rsquo;s made, who it&amp;rsquo;s made for, and all the consequences that follow
from its being made.&lt;/p></description></item><item><title>digital gardens</title><link>https://nilaykumar.github.io/garden/digital-gardens/</link><pubDate>Mon, 17 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/digital-gardens/</guid><description>&lt;p>I don&amp;rsquo;t remember exactly when it was that I stumbled upon &lt;a href="https://gwern.net/">gwern.net&lt;/a>, perhaps
high school or early college. It&amp;rsquo;s tempting to make the comparison to Alice
tumbling down a rabbit hole, but my hazy recollection is that I found the
content less interesting than the shape and form of the website itself. Less of
a rabbit hole and more of a garden, albeit one somewhat mysterious in its
monochromatic darkness. To find a hypertext garden in the age of Facebook&amp;rsquo;s
explosive popularity felt like a breath of fresh air.&lt;/p></description></item><item><title>test</title><link>https://nilaykumar.github.io/garden/test/</link><pubDate>Sun, 16 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/garden/test/</guid><description>&lt;p>This is a test file in the garden.&lt;/p></description></item><item><title>tracking damage in Gaza using open-access SAR data</title><link>https://nilaykumar.github.io/blog/insar-gaza/</link><pubDate>Mon, 10 Feb 2025 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/blog/insar-gaza/</guid><description>&lt;blockquote>
&lt;p>At the threshold of detectability, both the surface of the negative and that of
the thing it represents must be studied as both material objects and as media
representations. In other words, this condition forces us to remember that the
negative is not only an image representing reality, but that it is itself a
material thing, simultaneously both representation and presence.&lt;/p>
&lt;div class="attribution">
&lt;p>Weizman, &lt;em>Forensic architecture: violence at the threshold of detectability&lt;/em> [&lt;a href="#citeproc_bib_item_1">1&lt;/a>]&lt;/p></description></item><item><title>mnist and the idx file format</title><link>https://nilaykumar.github.io/blog/mnist-data/</link><pubDate>Sat, 05 Mar 2022 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/blog/mnist-data/</guid><description>&lt;p>Boser, Guyon, and Vapnik (BGV) introduce support vector machines (SVMs) in
[&lt;a href="#citeproc_bib_item_1">1&lt;/a>] and their immediate application is that of optical character
recognition (OCR). More specifically, given images of handwritten numerals, they
train a classifier to categorize a given image by digit. BGV show that SVMs can
achieve performance comparable to that of a more sophisticated five-layer neural
network classifier. I&amp;rsquo;ve been working through their paper to learn about SVMs,
and I thought it might be fun to try to follow along. There is a fairly large
dataset of handwritten numerals due to LeCun, Cortes, and Burges [&lt;a href="#citeproc_bib_item_2">2&lt;/a>],
called the MNIST database that&amp;rsquo;s publicly available. What follows are my notes
(mostly for my future self) on how to get started working with this dataset.&lt;/p></description></item><item><title>one weird kernel trick</title><link>https://nilaykumar.github.io/blog/one-weird-kernel-trick/</link><pubDate>Thu, 03 Feb 2022 00:00:00 -0500</pubDate><guid>https://nilaykumar.github.io/blog/one-weird-kernel-trick/</guid><description>&lt;p>I recently stumbled upon a programming interview question and the first solution
that sprang to mind used a sprinkling of algebraic topology. I thought this
might be a fun little pedagogical way to introduce the idea of homology, so here
goes!&lt;sup id="fnref:1">&lt;a href="#fn:1" class="footnote-ref" role="doc-noteref">1&lt;/a>&lt;/sup> I&amp;rsquo;ll assume familiarity with matrices, the basics of
linear algebra, and a bit of python.&lt;/p>
&lt;p>Say we&amp;rsquo;re given a 2-dimensional matrix where each cell consists of either a
forward slash, a backward slash, or a space. We&amp;rsquo;ll write a program to compute
the number of regions in the resulting space, treating the matrix edges as
boundaries. Given, for example, the 3-by-6 matrix,&lt;/p></description></item><item><title>the twisted eilenberg zilber theorem</title><link>https://nilaykumar.github.io/blog/the-twisted-eilenberg-zilber-theorem/</link><pubDate>Fri, 13 Aug 2021 00:00:00 -0400</pubDate><guid>https://nilaykumar.github.io/blog/the-twisted-eilenberg-zilber-theorem/</guid><description>&lt;p>Recently I was curious about the origins and initial applications of the
homological perturbation lemma (more on that below). Turns out it&amp;rsquo;s a beautiful
story involving the homology of simplicial fibrations (due to [&lt;a href="#citeproc_bib_item_1">1&lt;/a>] and
[&lt;a href="#citeproc_bib_item_2">2&lt;/a>]) that I had somehow managed to avoid learning about. In this brief
note I want to sketch this arc of ideas.&lt;/p>
&lt;p>For spaces \(B\) and \(F\), the Eilenberg-Zilber theorem [&lt;a href="#citeproc_bib_item_3">3&lt;/a>] constructs
chain equivalences&lt;/p>
&lt;p>\begin{equation*}
\begin{xy}
\xymatrix{
C_*(B\times F) \ar@(ur, ul)[r]^{} &amp;amp; C_*(B) \otimes C_*(F) \ar@(dl, dr)[l]^{}
}
\end{xy}
\end{equation*}&lt;/p></description></item></channel></rss>