🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

WHAT HAPPENED IN 2024 - Both shifts in turnout AND shifts in voters’ partisan preferences contributed to the 2024 election outcome catalist.us/whathappened20…

It shouldn't be so hard to publish null results

The modern political information ecosystem is failing to support an informed citizenry. It's May 2025, and a major site's homepage shows you "stories" about Clinton, Harris, crime, and Trump's "good calls"

This is wrong. I talk to folks at the big AI labs all the time. Biggest errors they make, economically: 1) how long it takes to shift the structure of institutions and what this means for diffusion, 2) AI is acomplement for some tasks, not just a substitute, 3) prices adjust. 1/3

YouGov poll: Only 8% think permanent space living in 50 years is "Very likely." Over 45% are doubtful. Looks like the public's not quite inspired to pack their bags for Mars yet.

Confidence in national institutions, by partisanship, 1974-2024. The collapse in Democrats' Supreme Court confidence in 2022-2024 is the largest single year change seen in 50 years. Data from the GSS NORC.

A new bipartisan bill would require private insurers to fully cover the cost of childbirth with no co-pays or deductibles. Average out-of-pocket costs today is nearly $3,000. It’s a big deal, with an unusual origin story, and backed by an even more unusual alliance of orgs 🧵

So it’s actually a good time to be alive

A (rare) strike against motivated reasoning? Partisans on both sides were willing to indicate in 2023 that their likely nominee was non-ideal

Super-simple but apparently it bears repeating: “trust science” ideally means insights are derived from the process of scientific inquiry where what we think we know is provisional. It doesn't mean every single study the speaker chooses to cite is an immutable truth.

My advice to students - Listen for nuance: Does your teacher present findings as definitive proof, or as current best understanding subject to revision? Do they discuss the limitations of research?

Look at those numbers

"Think of your phone – all the logos on it – how many apps could you delete today, which do you rely on, and why? What happens if you lose your phone?" blogs.lse.ac.uk/parenting4digi…

“The alternative to chatting with GPT wasn’t chatting with an MD; it was being lost.”

Sci-fi author Ted Chiang famously likens ChatGPT to “a blurry JPEG of all the text on the web.” openquestionsblog.substack.com/p/why-arent-yo…

I guess that’s right. But not a W for the MAGA aligned media either though. They’re taking all of this very badly because however silly it makes some of the MSM coverage of Musk seem, it makes the fawning “coverage” of DOGE and the “dream team” look far dumber.

This is amazing - and also a case when suggestive evidence is more than enough for me

(looks at the calendar) Oh right, it's a new month, so of course it's time for a another impressive Yamil paper!