Lightnews — Scholar-powered news

Reposted by soodoku

Christopher Mims @mims.bsky.social · 13h

This is a REALLY good video that gets into the *specifics* of how AI is degrading our information environment and putting truth further out of reach for the average person / the population as a whole

🎃 🧟 Prof. Dr Beth Singler 👻 🎃 @bvlsingler.bsky.social · 19h

I walked into the lounge last night to find my teenage son watching this. We have long admired @kurzgesagt.org and found their videos to be extremely illuminating. But this one in particular I think is so important right now!
www.youtube.com/watch?v=_zfN...

AI Slop Is Killing Our Channel

YouTube video by Kurzgesagt – In a Nutshell

www.youtube.com

3 160 370

soodoku @soodoku.bsky.social · 14h

Trump Trumps All: Coverage of Presidents on Network Television News - By Gaurav in Social Science — 03 May 2020 Trump Trumps All: Coverage of Presidents on Network Television News Photo by Markus Spiske / Unsplash With Daniel Weitzel . The US government is a federal system with substan…

Read more →

Reposted by soodoku

soodoku @soodoku.bsky.social · 1d

a generative model for names trained on the florida voter registration data

github.com/appeler/naam...

GitHub - appeler/naamkaran: generative model for names

generative model for names. Contribute to appeler/naamkaran development by creating an account on GitHub.

github.com

1

Reposted by soodoku

soodoku @soodoku.bsky.social · 4d

Traveling Salesman - By Gaurav in Commentary — 17 Nov 2023 Traveling Salesman White-collar elites venerate travel, especially to exotic and far-away places. There is some justification for the fervor—traveling is pleasant. But veneration creates an umbra that hides some truths: Local t…

Read more →

1

Reposted by soodoku

soodoku @soodoku.bsky.social · 5d

The Story of Science: Storytelling Bias in Science - By Gaurav in Science — 07 Jun 2021 The Story of Science: Storytelling Bias in Science Photo by Etienne Girardet / Unsplash Often enough, scientists are left with the unenviable task of conducting an orchestra with out-of-tune instrum…

Read more →

1

Reposted by soodoku

soodoku @soodoku.bsky.social · 6d

Streaming Calibration - By Gaurav in ML/Statistics — 19 Jul 2025 Streaming Calibration Photo by Cam Adams / Unsplash Modern applications—from ad platforms calibrating click-through predictions to polling systems incorporating responses to ML algorithms adapting fairness thresholds—shar…

Read more →

1

Reposted by soodoku

soodoku @soodoku.bsky.social · 7d

Measuring Tree Stability - By Gaurav in ML/Statistics — 04 Jul 2025 Measuring Tree Stability Photo by billow926 / Unsplash Consider training a decision tree to predict customer churn. You split your data in half and build a model on one half. The tree identifies tenure, monthly charges…

Read more →

1

soodoku @soodoku.bsky.social · 1d

a generative model for names trained on the florida voter registration data

github.com/appeler/naam...

GitHub - appeler/naamkaran: generative model for names

generative model for names. Contribute to appeler/naamkaran development by creating an account on GitHub.

github.com

1

Reposted by soodoku

Kevin Collins @kwcollins.bsky.social · 1d

More great work from @gelliottmorris.com and the @verasight.bsky.social team on synthetic samples. The area where it is probably most enthusiastically embraced is market research ... turns out it doesn't work well for simulating consumer behavior either www.verasight.io/reports/coff...

1 11 38

soodoku @soodoku.bsky.social · 2d

see also: github.com/soodoku/wait

GitHub - soodoku/wait: Waiting Times at CA DMV

Waiting Times at CA DMV. Contribute to soodoku/wait development by creating an account on GitHub.

github.com

soodoku @soodoku.bsky.social · 2d

technically no one should have to wait. they should allow people to book slots based on forecasted volume/staff or provide a call back functionality

1

soodoku @soodoku.bsky.social · 2d

after extensive de-vibing!

github.com/finite-sampl...

GitHub - finite-sample/optimal_classification_cutoffs: Script for calculating the optimal cut-off for max. F1-score or accuracy

Script for calculating the optimal cut-off for max. F1-score or accuracy - finite-sample/optimal_classification_cutoffs

github.com

1

Reposted by soodoku

Jeremy Horpedahl @horp.bsky.social · 3d

The increase in life expectancy is not solely a function of reduced child mortality. Life expectancy at age 15 has also increased dramatically in the past 100-200 years.

1 4

soodoku @soodoku.bsky.social · 4d

Traveling Salesman - By Gaurav in Commentary — 17 Nov 2023 Traveling Salesman White-collar elites venerate travel, especially to exotic and far-away places. There is some justification for the fervor—traveling is pleasant. But veneration creates an umbra that hides some truths: Local t…

Read more →

1

soodoku @soodoku.bsky.social · 5d

The Story of Science: Storytelling Bias in Science - By Gaurav in Science — 07 Jun 2021 The Story of Science: Storytelling Bias in Science Photo by Etienne Girardet / Unsplash Often enough, scientists are left with the unenviable task of conducting an orchestra with out-of-tune instrum…

Read more →

1

soodoku @soodoku.bsky.social · 6d

Streaming Calibration - By Gaurav in ML/Statistics — 19 Jul 2025 Streaming Calibration Photo by Cam Adams / Unsplash Modern applications—from ad platforms calibrating click-through predictions to polling systems incorporating responses to ML algorithms adapting fairness thresholds—shar…

Read more →

1

soodoku @soodoku.bsky.social · 7d

Measuring Tree Stability - By Gaurav in ML/Statistics — 04 Jul 2025 Measuring Tree Stability Photo by billow926 / Unsplash Consider training a decision tree to predict customer churn. You split your data in half and build a model on one half. The tree identifies tenure, monthly charges…

Read more →

1

Reposted by soodoku

Christopher Mims @mims.bsky.social · 8d

I've long skimmed the comment section of pieces at WSJ to find out where political fault lines are -- readers there actually debate each other

Don't think I've ever seen one like this. There is hardly any debate. Comments are overwhelmingly negative. Hardly anyone is trying to rationalize it

headline: Trump Tells Generals the Military Will Be Used to Fight ‘Enemy Within’

President says some U.S. cities he considers dangerous should become training grounds for American troops

2 15 70

soodoku @soodoku.bsky.social · 8d

The never heard of numbers are amazing ...

3

Reposted by soodoku

apoorva lal @apoorvalal.com · 8d

apoorvalal.github.io/lalgorithms/...
new linklog - ecological inference, adversarial estimation, bargaining games with The Agent

2025-09-29

Apologies for the interruption in coverage; I have a good excuse, henceforth known as The Agent (of both the ‘defying the Principal’ variety and the fast-learning natural/artificial variety).

apoorvalal.github.io

1 4

soodoku @soodoku.bsky.social · 8d

Cell Mates: Extracting Useful Information From Tables for LLMs - By Gaurav in ML/Statistics — 06 May 2025 Cell Mates: Extracting Useful Information From Tables for LLMs We seem to have cracked the art of distilling information in words and images using large machine learning models. Ho…

Read more →

Reposted by soodoku

Julia M. Rohrer @dingdingpeng.the100.ci · 9d

Looks like my favorite paper on the age trajectory of happiness is finally out!!! So happy for the authors. Go check it out, it’s great.

7 18 110

Reposted by soodoku

Matthew Yglesias @mattyglesias.bsky.social · 9d

Office buildings and the municipal governments who depend on them for tax revenue are still not back

www.slowboring.com/p/municipali...

5 6 47

soodoku @soodoku.bsky.social · 9d

github.com/finite-sampl...

Maximize F1, precision, recall, accuracy, or custom cost-sensitive utilities with algorithms designed for piecewise‑constant classification metrics.

GitHub - finite-sample/optimal_classification_cutoffs: Script for calculating the optimal cut-off for max. F1-score or accuracy

Script for calculating the optimal cut-off for max. F1-score or accuracy - finite-sample/optimal_classification_cutoffs

github.com

1

soodoku @soodoku.bsky.social · 9d

the going rate is < $1 for 1000: 2captcha.com (i have used this for some indian gov. websites)

Captcha Solver: reCAPTCHA solver and captcha solving service. Bypass captchas using the best auto captcha solver online API - 2Captcha

Fastest online captcha solving service starting at just $1 for 1,000 captchas. Service supports APIs including PHP, Python, C++, JAVA, C#, and JavaScript, ensuring seamless integration with your appli...

2captcha.com

1