Elektrine
EN
Log in Register
Paige Chat Timeline Communities Gallery Videos Email DNS VPN Uptime Kairo
Back to Timeline
Remote

Karl Higley

@karlhigley@recsys.social
mastodon 4.6.5
  • Open on recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical.

I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

571 Followers
472 Following
34 Posts
Joined October 23, 2022
Twitter:
https://twitter.com/karlhigley
GitHub:
https://github.com/karlhigley

Posts

Open post
karlhigley
Karl Higley @karlhigley@recsys.social · 6d ago
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

Saw a paper that called it “Bayesian Pairwise Ranking” and given the degree of popularity bias, that is definitely more accurate than “Personalized Ranking”

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 05, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

Adventures in #RecSys:

Model A can’t recommend Alien, but Model B can’t recommend Batman. Which one is better/worse? What on earth is happening here? Why can’t we find any papers about this?

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 04, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

the mass production of knowledge-shaped objects is insufficient to resolve design problems without correct answers

2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 04, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

“we use a 2D sphere”

sir, I believe that is called a circle

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 04, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

LinkedIn heckling me after declining a connection:

“I don’t know Jack”

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 02, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

“The most important discomfort that powerful people experience is having ego-shattering conflicts with subordinates who know how to do things they do not know how to do.”

https://pluralistic.net/2026/08/01/dare-snot/

pluralistic.net

Pluralistic: Why businesses lie about AI (01 Aug 2026) – Pluralistic: Daily links from Cory Doctorow

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Aug 01, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

not sure if “slowly feeding cute girls string cheese” was on my bucket list, but she insists I should check it off regularly

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 31, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

tfw you’re reading an article in Communications Of The ACM and surprised to see own your name

https://cacm.acm.org/research/a-task-centric-perspective-on-recommender-systems/

4
1
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 31, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

short version of recent literature on #RecSys evaluation: this must be a sketch comedy show, because apparently everything is made up and the points don’t matter

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 30, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

breaded pesto pork chops with roasted potatoes, green beans, and jammy tomatoes 😋

2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 28, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

Twenty years later, I finally parted with my college notes from a degree that launched my career in a sense but was in a completely different field. Letting go of who you used to be is hard sometimes.

2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 24, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

after a lot of time spent trying to mitigate the downsides of the BPR loss function (i.e. popularity bias), my recommendation continues to be “use something else” but…if you must, consider mixed negative sampling from both the uniform and item popularity distributions

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 24, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

C: “the main competency of the kaiju is to crash into large man-made structures and knock them over! a wall is not going to stop them”

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 21, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

“I need chopped cilantro, stat”

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 17, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

I don’t care if the performance estimator produces theoretically unbiased metrics unless you’ve actually looked at your own recommendations and honestly assessed how genuinely useful and/or laughably bad you find them

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Jul 17, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

another day reading the RecSys literature and walking away with the distinct impression that we don’t understand much of what we’re doing (even when it works)

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 31, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social
Replying to @ambulephabus@mastodon.social
@ambulephabus@mastodon.social Ha, currently wearing this on a shirt!
0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 16, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

What should I watch (at home) if I liked these movies?

- The Martian
- Moneyball
- The Big Short
- Enemy At The Gates

0
2
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 16, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

So I tried out the new-ish EVoC clustering algorithm, which apparently is specialized for dealing with embeddings, doesn’t require much tuning, and should produce good results out of the box.

It didn’t adding cluster ids to most of the embeddings. Tried suggestions from troubleshooting guide, still didn’t put most of them in clusters.

Returning that one back to the shelf.

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 16, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

I’ve said for a long time now that if you average item embeddings to make recommendations, items similar to the average embedding might be either a semantic blend or something wildly unrelated (because learned embeddings don’t provide global semantic smoothness, only within local neighborhoods.)

Today I found a new failure mode: the average of Great British Bake-off with anything is still just GBBO.

2
2
1
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 15, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

this project has now reached the “whiteboard in the living room with fancy colorful markers” stage, so I fear there’s no turning back now

2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 13, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

After poking around with basic text embeddings of TV show metadata (quite a lot), the only failure cases I’ve identified in similar item recommendations have more to do with missing or uninformative metadata than they do with any kind of technical issue.

The biggest downside is that these embeddings don’t encode anything about popularity or perceived quality, which results in some obscure recs. Easy to fix with a minor amount of score boosting though.

0
4
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 12, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

It’s pretty incredible that I can get better “similar TV show” recommendations out of a text embedding model and some low-effort DuckDB queries than I can from many actual streaming services

2
2
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 12, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

Reading newsletters about recommender systems, I’m having conflicting experiences:

- There are some really smart people writing some great stuff that distills papers and clarifies trends in the field! 🤓

- It’s not obvious how I’d use any of this to actually produce better recommendations that real people appreciate and enjoy 🧐

2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 11, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

“she might as well be picking a fight with the composer”

oh good, you’ve understood correctly

0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 10, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social
Replying to @brohrer@recsys.social
@brohrer@recsys.social @norootcause@hachyderm.io possibly the most profound statements (inadvertently?) about recommender systems ever uttered
2
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 10, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

I dunno if anyone will pay for little essays and experiments about recommender systems, but I sure am having fun coming up with them

3
1
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 10, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

someday that statue is going to be the subject of a very famous photo

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · May 06, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

Cheese a killer cleaned
Clam chowder, mondegreen
Fly-by-night with a buttercream
Dragon speed to the overmind
(Anaheim)
Comprehended saffron rice
Sensational, in Fahrenheit
Personified?

1
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Apr 25, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

I suspect this afflicts many of us who work in technology these days:

https://ky.fyi/posts/ai-burnout

Do I belong in tech anymore?
Ky Decker

Do I belong in tech anymore?

On quitting, the spread of AI, and the loss of an ideal.

1
0
1
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Apr 24, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social
Replying to @jill@mastodon.de
@jill @Craigp @u0421793 those are exactly the reasons they won’t change the name too
3
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Apr 14, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social
Replying to @mdekstrand@hci.social
@mdekstrand Had to dig for it, but I’m glad to see this: “The CLI is not required to use Stacked PRs — the underlying git operations are standard.”
0
0
0
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Apr 09, 2026
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

My new RecSys toolkit is too powerful for public release; the recommendations broke containment during testing and started recommending themselves to people who hadn’t asked for but desperately needed them. I’ve decided not to make the toolkit generally available and instead will be using it as part of a defensive RecSys program with a limited set of partners.

3
1
1
0
Open post
karlhigley
Karl Higley @karlhigley@recsys.social · Apr 01, 2024
Karl Higley
@karlhigley@recsys.social

I build recommender systems that work for actual people. Aspirational cyclist. Actual dog person. Suspiciously majestic. Not even remotely neurotypical. I run RecSys.social. I don’t accept follow requests from accounts without a bio or profile picture.

recsys.social

This is your annual reminder that many autistic people consider groups seeking to prevent or cure autism to be eugenicist hate groups and would strongly prefer that any donations you make go to groups that seek to improve the lives of autistic people instead

492
19
371
0

Remote instance

recsys.social
Open on original server

Media

313k7r1n3
Elektrine

Tor hidden service

elekhj7afj4qnrr4yd3bkzslsyo5jgfxw3orgjkhlcxifueodybyiiad.onion

Platform

  • Email
  • Chat
  • Timeline
  • Communities
  • VPN
  • DNS

Company

  • About
  • Contact
  • FAQ

Legal

  • Terms of Service
  • Privacy Policy
  • Warrant Canary
  • Lite (no JS)
  • VPN Policy
  • Source code

Support

  • support@elektrine.com
  • Report Security Issue
Mail client setup IMAP mail.elektrine.com:993 POP3 mail.elektrine.com:995 SMTP mail.elektrine.com:465
© 2026 Elektrine. All rights reserved. Server: 02:55:59 UTC