I used to cycle along that path. Along time ago. I met the guys painting it once as well. They said it was a nightmare to get correct. At some point someone to develop a way to check they didn't mae an error. A camera under a bike and come fancy imaga analysis maybe.

I think the microphones are noise cancelling, and you loose the sense of just how loud it is in the chamber. That long a in maa-dam is the chancelor trying to be heard over the din.

These exist.
en.wikipedia.org/wiki/Thermoe...
Just expensive and inefficient.

I read the cartoon as microsoft coming in like an "Angry Bird" wrecking ball to demolish the whole stack. Or at least the bit being made unstable by the AI jack.

I'm sure I remember, although I can find no trace of it, a website around year 2000 called something like "communications from elsewhere" that procedurally generated nonsence humanities papers. Maybe one was submitted to a jounal in a Sokal type hoax. A lot of this stuff isn't actually new.

So I googled "Ordered Monoid" and "Kripke Semantics". Wow. Sending thoughts and prayers.

The agony of having other people see the mess I made of signing up to a group social event in google sheets was enough for me.

Funny graphic. But I read the intent of "faster than the universe" to mean something like "complete computation in a reasonable time frame that would take conventional compute capability longer than the expected lifetime of the universe."

Take n pairs of twins; separate them from their parents and each other at birth; randomly allocate one twin to a bilingual couple the other to a monolingual couple; follow up over 90 years. Costs include bribes for ethics comittee.

I mean, there may be other issues with the paper! Psyc research being what it is. It's just the problem wasn't power.

Also, doesn't kin selection really only apply (in the sense of increasing population frequency over time) to the actual genes that confer kin selecting altruism behaviours. All other genes are just along for the ride.

Not power though. The original paper made a type-I error, false rejection of null hypothesis. Power is sensitivity of experiment to correctly reject null hypothesis, or avoidance of type-2 error. Type-I errors will occure 5% of the time at p-value 0.05. Independed replication provides protection.

Since yoy mention excel...
For ggplot2 graphs you can get data out with ggplot_build(), and then output the actual data for the plot to excel with openxlsx.
ggplot2.tidyverse.org/reference/gg...
cran.r-project.org/web/packages...

Arguably both trends are driven by some weird stuff going on 1985-95. That's when computers became available, but researchers maybe didn't know how to use them yet, data entry was manual and error prone and software wasn't well developed.