Naloxone helps mitigate the effects of an overdose, it doesn’t prevent overdoses (or at least isn’t used for that purpose). An overdose is not always lethal.

The research suggests that the observed reduction in overdose deaths was not primarily due to increased use of naloxone.

There are limits, but they are still too large and preparing an ERC grant application takes weeks of fulltime work.

I think you can ethically, if not bureaucratically, toss all the grant applications in the garbage and just look at the resumes and publication records, and you’d probably get better results in terms of ensuring the best scientists obtain funding.

This rise is attributable, in part, to the ease of writing applications with artificial-intelligence tools.

This is definitely one of the better uses of AI. Billions are wasted every year on utterly pointless grant application writing. I have long been in favour of laws severely restricting the maximum length of scientific grant applications, but now technology has made that unnecessary. In many cases scientists simply chose not to apply for grants they might have a chance of getting because the process is so cumbersome and bureaucratic, now they can just have an LLM write it so they can get back to their actual job.