Stephen Wild

I haven't read the book or seen the movie, so this reference is lost of me. To Gemini of an explanation! :-p

If anything I underestimate what they can do because I am cheap and refuse to pay for premium subscriptions

Either way, I still think what they can do is impressive, even if I think it premature to be declaring consequences like "the end of [X]"

Part of that investment may reflect the culture of AI companies, and part of it may reflect that these areas are places where it's easier to verify outputs, and part of it may be that's where the demand for them is. I'm very curious to see how they perform once we get outside that domain.

Interesting.

I am still relatively agnostic as to LLMs, Agents, and their abilities, and how I see them affecting different areas. This gives me something else to think about.

So far, the strongest "effect" of LLMs seems to be a compulsion to see this technology outside of history, and attributing pathologies to them that predate them. It's a bad way of thinking about this or any technology.

So it's nice to see a takes that come at it from different angles (apprenticeship and the normative role of professors)

It's interesting to see some of the takes on the article (yes, I succumbed to the bait)

In rough order:

1) AI doesn't do as good of job as humans
2) if you use AI, did you actually do the work?
3) AI is excellent and can replace [X] of the workflow
4) the publication system/science is doomed

Agreed. Maybes "advances" was not the right word. "Moves," maybe?

...the problem is in dealing with the quantity of output that advances knowledge little or not at all.

I disagree slightly here, in that I see science qua knowledge production as following a power law. Most of the time knowledge is advanced little or not at all, but some advances are massive. More knowledge production = more advancement (eventually)...

Hey #stats folk, I'm looking for some great (but concise) YouTube vids about likelihood functions & estimation (MLE, properties, Fisher information, std errs).

These would serve students taking my Bayesian course that didn't take the preceding likelihood course (that is taught by someone else).

Yes, I succumbed to the bait.

🤔

Interesting and useful perspective

But see for an alternstive:

www.cambridge.org/core/journal...

In Canada 5 year terms are the most popular by far

👀

bsky.app/profile/akhi...

Yes but

bsky.app/profile/step...

*academic science

Yes but

bsky.app/profile/step...

bsky.app/profile/step...

The Templin lectures are a nice intro to both Bayesian stats and psychometrics.

FWIW he (like me) doesn't like EFA.

bsky.app/profile/step...

Multidimensional Measurement Models Fall 2023 - YouTube Lectures and other videos for the Multilevel Measurement Models course from Fall Semester 2023

He also has a nice course on multidimensional measurement models:

youtube.com/playlist?lis...

This is a nice collection of lectures on Bayesian psychometrics by Jonathan Templin

One question in the poll, for example, asked our sample of 1,035 adults whether they supported “an amnesty program that would grant legal status to people who entered the country illegally.” The second asked about “providing a path to citizenship for undocumented immigrants who have resided in the United States for at least ten years, paid taxes, and passed a background check.”

Some of these questions aren't contradictory. Eg:

Sage Journals: Discover world-class research Subscription and open access journals from Sage, the world's leading independent academic publisher.

Six:

journals.sagepub.com/doi/abs/10.1...

Bayesian Item Response Modeling in R with brms and Stan by Paul-Christian Bürkner <p>Item response theory (IRT) is widely applied in the human sciences to model persons' responses on a set of items measuring one or more latent constructs. While several R packages have been develope...

Five:

www.jstatsoft.org/article/view...

Four:

medewitt.github.io/resources/st...