Hacker News Reader

Comment by jazzpush2 | Hacker News Reader

Shame, it's a cool idea, but the writing is so full of claudism's my mind skims over it completely.

Bit of an embarrassment to publish that as a published researcher, to be honest.

This is a fairly cynical way to look at it. Consider the scenarios:

Pre-LLM, most researchers wouldn't have bothered writing it up outside of a journal. They're extremely busy people, and it takes a lot of effort to write a well polished post. They'll either hand that task off to a grad student or (much more likely) simply not post it anywhere.

Thanks to LLMs, we know about it. I've often been critical of academia for not making their work accessible to the masses. This is a step in that direction.

Existing academics' writing is very poor. Did you try reading the average Academic Joe's essays in the past? Much poorer quality (with much poorer images). LLM-speak is a step up.

It’s a step down. This is not efficient writing. It would rank as very bad academic writing indeed.

It’s not even good for academia, journals are flooded with slop.

Academic writing (as in journals) is pretty bad. This is way better.

Asimov joked that when it was time to write his thesis, he worried that he wouldn't be able to write as poorly as is expected in academia given that he had spent most of the last decade trying to improve his writing.

I am a researcher. Researchers, now more than ever with AI, have plenty of time to write something. The lift of writing is never the hard part, anyways.

When I was in grad school at a top tier university, academics never spent time writing anything but grants (or, occasionally, lecture notes). If they really felt they needed to (e.g. pressure to publish in IEEE Spectrum), they got an underling to do it for them and simply edited it.

> they got an underling to do it for them and simply edited it.

This was not edited so that's already far more effort than was put into this

Reply on news.ycombinator.com