How does this compare with state-of-the-art LLMs? Does this mean that the bottom 20% of 15 years old in Denmark, which has free education[^1] are only good for playing Call of Duty and taking plumber jobs? What does that mean for the future?

[^1]: At least for citizens. Immigrants there face a lot of hurdles, in part because Danish people have turned somewhat xenophobic, and I wouldn't be surprised if access to education for people with a foreign background is a complicated matter.

I couldn't find anything about this 24% figure.

There was a PISA report from 2022 though, that showed 19% of Danish 15-year-olds are below 'Level 2 reading'. But this bar is considerably higher than 'understand a simple text'.

> Some 81% of students in Denmark attained Level 2 or higher in reading (OECD average: 74%). At a minimum, these students can identify the main idea in a text of moderate length, find information based on explicit, though sometimes complex criteria, and can reflect on the purpose and form of texts when explicitly directed to do so. The share of 15-year-old students who attained minimum levels of proficiency in reading (Level 2 or higher) varied from 89% in Singapore to 8% in Cambodia.

So yeah, I doubt that claim. Journalism at its best, as always

In 2015 (or some other year), they calculated the bottom quintile (or maybe quartile) score of students on some reading assessment, declared this to be the lowest level of reading competency, and wrote a helpful explanation that people who score at the lowest level of reading competency struggle with simple texts. But there's no single simple text that you could point to where people below the threshold completely fail to understand it and people above the threshold easily breeze through.

Because reading competency isn't binary like that. Reading competency is binary in that complete illiterates cannot read at all and people who aren't complete illiterates typically can read quite a lot, but there probably aren't enough actual illiterates in Denmark to produce scary-sounding statistics. So instead you get arbitrary thresholds that don't mean what they claim to mean.

I could imagine a significant portion of the students tested can't be arsed to make the effort of reading carefully, especially since the test result wouldn't even impact them, I'm not sure how they can control for that.