The clear reason to pay for news is that you can buy higher quality news than what your social media shows you. But I did definitely carve out politically sensitive areas in my discussion for a reason.

> They even changed font to random sizes to have it appear unhinged

This caught my eye, but appears to be false: https://web.archive.org/web/20170805210606/https://gizmodo.com/exclusive-heres-the-full-10-page-anti-diversity-screed-1797564320 Has some weird formatting, presumably from copying it in from a Google doc, and presumably also why it lost the figures and URLs. The formatting doesn't look unhinged at all, just a bit awkward, though their summarizing the changes as removing "several" hyperlinks is terrible (it looks more like a couple dozen links in the original to me). Though, I would not have ever thought of Gizmodo as a being high tier journalism in the first place.

Good suggestion, though I don't know how to systematically assess that. I can't even think of what topics would be most likely to have this come up in.

Thanks for this example. I definitely see ridiculous headlines like that from less reputable places. Do you also have examples from the type of news media I'm talking about like WSJ? For example, searching "Washington Post AI robotics" I get headlines:
 

• "Humanoid robots were sci-fi. Suddenly they're everywhere" about companies investing in and demoing humanoid robots, which seems to be true.
• "Not ready for robots in homes? The maker of a friendly new humanoid thinks it might change your mind" about the product "Sprout" by Fauna Robotics, which seems okay unless Fauna is completely faking it
• "Russia's much-hyped humanoid robot face-plants onstage during debut" OK
• "Opinion | The Chinese robots are coming" hard to assess but not particularly hyperbolic
• "Robot smaller than grain of salt can 'sense, think and act'" subtitled "With solar cells and its own propulsion system, the device is a step toward sending robots into the human body". This seems closest to what you're talking about. Here's a press release for it: https://www.seas.upenn.edu/stories/penn-and-umich-create-worlds-smallest-programmable-autonomous-robots/ I'm sure this is an optimistic take on a research project but it seems fairly reasonable

(I realize now that "robotics" wasn't really in your original statement, I guess I extrapolated that from your drone example.)

Drug approvals have gone up in recent years: https://pmc.ncbi.nlm.nih.gov/articles/PMC10856271/ (figure 1). Of course most of those are not ones that you'll encounter in day-to-day life. Meanwhile, some of the most commonly used over-the-counter drugs from previous decades have been pulled from the market or made harder to get (cold medicine particularly: phenylpropanolamine due to rare side effects in 2000, oral phenylephrine due to lack of effect last year, and pseudoephedrine restricted to behind the counter due to use in meth a decade ago or so).

I was going to say that you should still have the kid checked due to "secondary drowning", but apparently that's largely a myth: https://www.redcross.org/take-a-class/resources/articles/dry-or-delayed-secondary-drowning According to the Red Cross, there's no record of anyone nearly drowning, completely returning to normal, and then dying afterwards. If the person had shown symptoms like confusion or coughing, they'd be at risk for later dying despite rescue, but not if they completely and quickly recovered after the incident.

I'm not as concerned about your points because there are a number of projects already doing something similar and (if you believe them) succeeding at it. Here's a paper comparing some of them: https://www.biorxiv.org/content/10.1101/2025.02.11.637758v2.full

ML arguments can take more data as input. In particular, the genomic sequence is not a predictor used in LASSO regression models: the variants are just arbitrarily coded as 0,1, or 2 alternative allele count. The LASSO models have limited ability to pool information across variants or across data modes. ML models like this one can (in theory) predict effects of variants just based off their sequence on data like RNA-sequencing (which shows which genes are actively being transcribed). That information is effectively pooled across variants and ties genomic sequence to another data type (RNA-seq). If you include that information into a disease-effect prediction model, you might improve upon the LASSO regression model. There are a lot of papers claiming to do that now, for example the BRCA1 supervised experiment in the EVO-2 paper. Of course, a supervised disease-effect prediction layer could be LASSO itself and just include some additional features derived from the ML model.

This is a lovely little problem, so thank you for sharing it. I thought at first it would be [a different problem](https://www.wolfram.com/mathematica/new-in-9/markov-chains-and-queues/coin-flip-sequences.html) that's similarly paradoxical.

Again, why wouldn't you want to read things addressed to other sorts of audiences if you thought altering public opinion on that topic was important? Maybe you don't care about altering public opinion but a large number of people here say they do care.

He's influential and it's worth knowing what his opinion is because it will become the opinion of many of his readers. Hes also representative of what a lot of other people are (independently) thinking.

What's Scott Alexander qualified to comment on? Should we not care about the opinion of Joe Biden because he has no particular knowledge about AI? Sure, I'm doubt we learn anything from rebutting his arguments, but once upon a time LW cared about changing the public opinion on this matter and so should absolutely care about reading that public opinion.

Honestly, I embarrassed for us that this needs to be said.