Ben Radford's Avatar

Ben Radford

@benradford.bsky.social

Computational social science, machine learning, Bayesian stats, political conflict, forecasting, cybersecurity, international relations. www.benradford.com

693 Followers  |  852 Following  |  96 Posts  |  Joined: 03.07.2023
Posts Following

Posts by Ben Radford (@benradford.bsky.social)

AI's one and only killer app is summarizing in text all of the 10 min how-to videos that people decided were a better idea than how-to articles about ten years ago

01.03.2026 18:58 — 👍 0    🔁 0    💬 0    📌 0

Call me a conspiracy theorist but there's got to be more to this Anthropic / DoW fight than we've been told

27.02.2026 23:44 — 👍 0    🔁 0    💬 0    📌 0
Video thumbnail

Kat Abughazaleh: "We need to, first off abolish and prosecute ICE and DHS and we return DHS' responsibilities to Departments of Justice, State, of Treasury, of Transportation. All of these responsibilities that were taken from those departments... We need to repeal the Patriot Act."

26.02.2026 02:09 — 👍 1265    🔁 172    💬 22    📌 31

Banana pudding is so much better than the sum of it's parts

26.02.2026 00:35 — 👍 2    🔁 0    💬 0    📌 0

"AI isn't ready to write a paper on its own yet - it's just not a very good writer. But I let it do all my coding and analysis for me it's great at that."

24.02.2026 22:11 — 👍 1    🔁 0    💬 0    📌 0

whichever part of writing a paper that you think AI can do better than humans is just the part of writing a paper that you are the worst at send tweet

24.02.2026 22:05 — 👍 10    🔁 1    💬 1    📌 0

The top scorer for Fri, Feb 20, 2026, UTC is: @benradford.bsky.social! They scored 360 points for #Wordle 1707 and solved it on row 1! That's better than 2,396 (~100%) other users. Never give up!

21.02.2026 00:03 — 👍 1    🔁 1    💬 0    📌 0

Wordle 1,707 1/6

🟩🟩🟩🟩🟩

Finally

20.02.2026 05:14 — 👍 0    🔁 0    💬 0    📌 0
Preview
a close up of a woman 's face with the words skeptical sighing above her ALT: a close up of a woman 's face with the words skeptical sighing above her
19.02.2026 23:50 — 👍 1    🔁 0    💬 0    📌 0

Did I miss something?

19.02.2026 23:46 — 👍 0    🔁 0    💬 1    📌 0

How so? If someone contributed language to the manuscript, they should be listed as an author, no?

16.02.2026 23:51 — 👍 1    🔁 0    💬 0    📌 0

Yikes 😬

16.02.2026 16:46 — 👍 1    🔁 0    💬 1    📌 0

Literally just heard someone on the phone in the elevator say "I'm calling to let senator fetterman know I'm alarmed and concerned..."

20.01.2026 16:11 — 👍 2    🔁 0    💬 0    📌 0

Baby bok choi is the cutest vegetable

16.01.2026 23:33 — 👍 0    🔁 0    💬 0    📌 0

When the US military said it could fight simultaneous major wars on two fronts, who'd have thought those would be against Iran and Greenland?

14.01.2026 23:52 — 👍 0    🔁 0    💬 1    📌 0
Preview
Pentagon bought device through undercover operation some investigators suspect is linked to Havana Syndrome | CNN Politics The Defense Department has spent more than a year testing a device purchased in an undercover operation that some investigators think could be the cause of a series of mysterious ailments impacting sp...

Talk about an interesting conspiracy/mystery. holy cow

www.cnn.com/2026/01/13/p...

14.01.2026 04:26 — 👍 0    🔁 0    💬 0    📌 0

Sorry scientists, but coding is part of the learning process and offloading it to LLMs is equivalent to reading a paper, not writing one

10.01.2026 17:02 — 👍 4    🔁 0    💬 0    📌 0

OMG yes I've been saying this! I would pay so much for an email service that charges the sender $0.05 per incoming message

04.10.2025 21:04 — 👍 1    🔁 0    💬 1    📌 0
Preview
Truncated or Censored Data

Is this not a right censored duration model problem? mc-stan.org/docs/stan-us...

27.09.2025 15:19 — 👍 1    🔁 0    💬 2    📌 0

Oops I meant Calley

20.09.2025 17:50 — 👍 1    🔁 0    💬 0    📌 0

I think soldiers under Callum were charged but not convicted for Mai Lai

20.09.2025 17:37 — 👍 2    🔁 0    💬 1    📌 0
Codebook LLMs: Evaluating LLMs as Measurement Tools for Political Science Concepts | Political Analysis | Cambridge Core Codebook LLMs: Evaluating LLMs as Measurement Tools for Political Science Concepts

Very excited that my paper with @katakeith.bsky.social is now out in @polanalysis.bsky.social. We investigate whether LLMs actually follow the instructions/definitions provided in codebooks, propose some diagnostics, and release a new evaluation dataset.
www.cambridge.org/core/journal...

19.09.2025 13:45 — 👍 30    🔁 14    💬 0    📌 2

what are the most widely read digital newspapers by web traffic across the US and EU? do you know of any current repositories?

16.09.2025 16:38 — 👍 16    🔁 21    💬 8    📌 1

NASA maybe finding life on Mars doesn't even make the cut, huh?

13.09.2025 01:41 — 👍 1    🔁 0    💬 1    📌 0

Yesssssss ❤️ Snow Tha Product

22.08.2025 06:51 — 👍 3    🔁 0    💬 0    📌 0

Congratulations!

22.08.2025 03:10 — 👍 1    🔁 0    💬 0    📌 0

Skills & tools that nobody teaches you but that pay for themselves many times over in social sciences:

1. regex
2. Linux
3. Selenium & beautifulsoup
4. Git
5. SQL

21.08.2025 18:13 — 👍 5    🔁 0    💬 0    📌 0

Learn to love regex. Don't let ChatGPT steal the joy that could be yours

15.08.2025 17:40 — 👍 1    🔁 0    💬 0    📌 0

"[editorial team reading] largely corroborates the reviewers' assessment both positive and critical. We found merit in your paper but share some of the concerns raised by the reviewers" 🙄

31.07.2025 22:30 — 👍 1    🔁 0    💬 1    📌 0

Two out of two reviews are basically all "I just don't get LLMs" like this

31.07.2025 18:04 — 👍 1    🔁 0    💬 0    📌 0