@mdekstrand@hci.social
@mdekstrand@hci.social avatar

mdekstrand

@mdekstrand@hci.social

Asst. Prof IS @ Drexel. Trying to make information access (RecSys, IR, etc.) useful and equitable.

This profile is from a federated server and may be incomplete. Browse more on the original instance.

mdekstrand, to random
@mdekstrand@hci.social avatar

When I created my Data Science class, I wrote a guide to writing effective assignment solution notebooks in Jupyter. I've now revised it to be more course- and tech-independent and moved it to my website. https://md.ekstrandom.net/resources/writing-notebooks

mdekstrand, to random
@mdekstrand@hci.social avatar

New on my blog — a detailed writeup of my setup for Zoom calls, remote teaching, and video recording & production, along with recommendations at a variety of price points for improved quality. https://md.ekstrandom.net/blog/2024/01/zoom-lecture-studio

mdekstrand, to random
@mdekstrand@hci.social avatar

Frodo: "It looks like a form of shell. I can't quite make it out."

Gandalf: "The tongue is that of Windows batch files, which I will not utter here."

mdekstrand, to random
@mdekstrand@hci.social avatar

A few songs I learned will be 25 years old next year…

• I Want It That Way (Backstreet Boys)
• Higher (Creed)
• All Star (Smash Mouth)
• (You Drive Me) Crazy (Britney Spears)
• All The Small Things (Blink-182)
• Smooth (Santana)
• Learn to Fly (Foo Fighters)

mdekstrand, to random
@mdekstrand@hci.social avatar

As is my tradition the last few years, I've written up the current state of my computing toolbelt: https://md.ekstrandom.net/blog/2023/12/tools

mdekstrand, to random
@mdekstrand@hci.social avatar

This by @kissane is really good, and towards the end covers a solid non-religious argument grounded in demonstrable harms for second-degree defederation from Threads. https://erinkissane.com/untangling-threads

mdekstrand,
@mdekstrand@hci.social avatar

The visibility-of-harassment problems are a significant weakness of the ActivityPub federation model. They also look incredibly difficult to solve.

grimalkina, to random
@grimalkina@mastodon.social avatar

To revive (jk it's never gone just buried under administrivia and the weight of healing from general occupational hazards like researcher harassment) my love for science, I've been rereading the early and famous papers that helped us understand that smoking causes lung cancer. Famous beautiful examples of causal reasoning and difficult to remember how HARD IT WAS to tell this story back then. Profoundly poignant to see people use the weapons of evidence reasoning to create real health change.

mdekstrand,
@mdekstrand@hci.social avatar

@natematias @grimalkina tobacco, eugenics, he really knew how to pick his causes, didn’t he.

mbojan, to rstats
@mbojan@sciences.social avatar

users and developers, I'm looking for opinions, success stories, failure stories of using DVC https://dvc.org/ or Datalad https://www.datalad.org/ with -based projects. I tried both solutions myself in small projects and have mixed feelings. I'm curious about stories/blogposts/experiences etc. of others. gügling was not too successful. @rstats

mdekstrand,
@mdekstrand@hci.social avatar

@mbojan @rstats I haven't written anything up about it, but do have a bit of R (via Quarto) in my Book Data Tools: https://github.com/PIReTship/bookdata-tools/

It's worked mostly OK; currently using the irkernel backend, for some simplicity + committing analysis outputs in a GitHub-browsable format, but might change that.

mdekstrand, to random
@mdekstrand@hci.social avatar

XKCD on the money again. When I was teaching databases, I had an approx. 30-minute extended discourse on date and time and why it's harder than the students think even after they think they've accounted for everything. And another 20-30 minutes in a different lecture on text encodings, why we can't have nice things, and why if I had a time machine I would go to the 1950s or 60s and leave copies of the UTF-8 specification in the break rooms with the note "please save us".

https://xkcd.com/2867/

scheidegger, to random
@scheidegger@mastodon.social avatar

I dare you to read the comments section here: https://news.ycombinator.com/item?id=38629355

mdekstrand,
@mdekstrand@hci.social avatar

@scheidegger <dons hermetically-sealed lead-plated, asbestos-lined enviro-suit>

<wades in>

NOPE

anildash, to random
@anildash@me.dm avatar

Discogs shitting the bed right after Bandcamp is really too much “nothing pure can last” happening at once for music fans. https://www.stereogum.com/2241158/the-discontent-at-discogs/columns/sounding-board/

mdekstrand,
@mdekstrand@hci.social avatar

@anildash dammit. I'm not a heavy user, but use it to catalog my collection & wishlist + buy an album every month or two. Didn't realize things were getting so bad there 😭.

anildash, to random
@anildash@me.dm avatar

There were two ways to handle the Rapid Reset threat — lots of high-stress work to try to react in realtime, or you could anticipate it by months, address the challenge, and make it a non-issue for millions. Here's the story of how to do the latter. https://www.fastly.com/blog/how-fastly-protects-customers-from-ddos-including-rapid-reset-attack

mdekstrand,
@mdekstrand@hci.social avatar

@anildash Apache similarly decided in 2016 that unbounded client-controlled concurrency was a bad idea & designed a "scale up with good behavior" algorithm that made the core rapid reset attack moot (a later variant was able to exploit memory allocator optimizations). https://github.com/icing/blog/blob/main/h2-rapid-reset.md

upol, to mastodon
@upol@hci.social avatar

Can we get LLMs to separate emails with bad news and put 'em in the trash where they belong?

Asking for my friend, my enemy, my frenemy, and myself.

mdekstrand,
@mdekstrand@hci.social avatar

@upol better yet, autoreply.

"Dear editor,

We regret to inform you that my client cannot accept your rejection of his paper at this time. He has received many excellent rejections, and but unfortunately space limitations restrict the number of rejections that can be accepted."

mdekstrand, to random
@mdekstrand@hci.social avatar

I took another swing at writing up my student recruiting page — wanted to highlight more directly why a student might want to work with me & what makes a good fit, with less “please don't do X or Y”. I'm recruiting for Fall 2024! https://md.ekstrandom.net/students/prospective

SecurityWriter, to random

My cat is great at catching spiders. There are however a few caveats:

  1. Spiders must be sufficiently enormous to be fun. Somewhere between “Oh lawd he coming” and Shelob.

  2. Spider catching only happens at night, and only in the bedroom.

  3. Cat will make a few chirping noises to alert us that a spider is being hunted.

  4. All spiders must be delivered as gifts onto the chest or face of one of us.

  5. Note I only said great at ‘catching’. Delivered spiders are very much alive and very fucking irate about the current state of affairs.

mdekstrand,
@mdekstrand@hci.social avatar

@SecurityWriter

A gif from The Good Place. One demon asks “why did you do that?”, to which Shawn replies “I’m a demon. It’s fun.”

shriramk, to random
@shriramk@mastodon.social avatar

People who complain "CS students today don't even take X!": do you know how much CS has grown? I pulled up the @BrownCSDept course list and marked in yellow all courses that did NOT exist ~20y ago (when many grumblers graduated). YMMV, but there are way, WAY more things to learn.

About half yellow.
Fully yellow.
Fully yellow.

mdekstrand, (edited )
@mdekstrand@hci.social avatar

@jbigham @neilernst @shriramk brb proposing a course on why TCL is Good, Actually, and the things that beat it out (except maybe python) are worse in crucial ways until idk typescript or something.

mdekstrand, to random
@mdekstrand@hci.social avatar

BE/NL IR friends — the Dutch-Belgian IR workshop is next month, CFP now up: https://dir2023.github.io/DIR2023/

timnitGebru, to random
@timnitGebru@dair-community.social avatar

You have rights under the EU AI act (except when you're are a refugee and cease to be a human).
https://www.article19.org/resources/eu-court-denies-full-transparency-about-emotion-recognition/

mdekstrand,
@mdekstrand@hci.social avatar

@timnitGebru 😭🤬

ct_bergstrom, (edited ) to random
@ct_bergstrom@fediscience.org avatar

AITA, academic publishing edition.

Journal sends a review back to me because as reviewer I did not run the code and replicate the results.

My reply: "As an unpaid anonymous peer reviewer who handles probably 30 manuscripts a year, I am absolutely not in the position of running and evaluating code any more than I am in the position of running gels to evaluate lab results. If this is important to you, I suspect you can find a consultant who will do it at a reasonable rate."

Am I the asshole?

mdekstrand,
@mdekstrand@hci.social avatar

@ct_bergstrom Nah. If a journal expects that, they need to set up a dedicated artifact review process like some computer science conferences have (and those usually have a team of Ph.D students run the supplied artifacts independent of the scientific peer review process for the paper content itself; I don't think they usually affect acceptance, but papers that pass get a badge for having a usable, validated artifact).

mdekstrand, to random
@mdekstrand@hci.social avatar

@deilann tired: ACID compliance.

Wired: postgreSQUIRREL, the ADHD-compliant database. Forgets what it’s doing halfway through a transaction! Commits twice because it isn’t sure it did the first time! Replies to queries with paranthetical bonus content!

mdekstrand, to random
@mdekstrand@hci.social avatar

still p. sad that Microsoft bought & shuttered Wunderlist. It was a really great piece of software, and I haven't found anything else quite like it.

scheidegger, to random
@scheidegger@mastodon.social avatar

It's very weird to cross an arbitrary, dumb threshold for a goal that you set yourself a decade ago for a job you haven't done in more than a year.

This is going to come across as bragging, but what I mean by it is that I let that goal dictate decisions that were probably overall negative for my actual previous job, my mental health, and that of my students.

And it really doesn't mean much at all. Goodhart's Law gets you coming and it gets you going, it turns out.

mdekstrand,
@mdekstrand@hci.social avatar

@scheidegger CSrankings is so toxic. But hey, I don’t count for it anymore! (At least until I go through the process to get an affiliate appointment in computer science.)

histoftech, to random
@histoftech@mastodon.social avatar

After more than a decade in Chicago I’m leaving for greener professional pastures: as of this fall I am a tenured associate professor in the new data science school at UVA. I’ll still be doing the same critical and historical work, just under a different disciplinary umbrella. I look forward to making connections with folks in multiple departments at UVA, particularly history, STS, & gender studies. If you’re there, hmu!

mdekstrand,
@mdekstrand@hci.social avatar

@histoftech many many congrats!!

mdekstrand, to random
@mdekstrand@hci.social avatar

Savage Chickens on AI…

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • GTA5RPClips
  • DreamBathrooms
  • InstantRegret
  • magazineikmin
  • osvaldo12
  • Youngstown
  • ngwrru68w68
  • slotface
  • everett
  • rosin
  • thenastyranch
  • kavyap
  • tacticalgear
  • megavids
  • modclub
  • normalnudes
  • cubers
  • ethstaker
  • mdbf
  • Durango
  • khanakhh
  • tester
  • provamag3
  • cisconetworking
  • Leos
  • anitta
  • lostlight
  • All magazines
    •