Particularly when the only source is a friend of the author, posting on a blog named "AI Clambake" about "A weekly, human-powered newsletter for advertising folks who want to stay on top of the AI mayhem" and not a publication with any credibility in linguistics.

None of that means it can't be true, but some basic skepticism is warranted here. Otherwise we end up in a situation like the LK99 room temperature superconductor where a lot of HN commenters were also upset at the cynical "downers" who just couldn't root for a good thing/progress.

1) Many preprints are bad, incredible bad. I read a lot of posts about ivermectine during 2020 and the errors were obvious. Like no control groups, the control group is a bunch of unrelated guys in another city, and a weird articles that split the 20+20 cases in 10 bins with 2+2 cases in each. They had a lot of error that were easy to spot without being a medical doctor. (Ctrl+F exclusions, you may get a surprise.) (And don't get me started with Chlorine Dioxide.)

2) Perpetual mobile and mass less drive reappear every few years. I definetively can read most of them. The most interesting part is the totally broken explanation of why this new version does not break the laws of physics.

3) HN has a lot of users specialized in niche topic. A few weeks ago I wrote a comment with a joke: "the list of text transformation to allow a Spanish speaker to read German enters in a napkin" (for example v->f and w->v and a few more). Someone was surprised because s/he knows that German has more phonemes than English that has more phonemes than Spanish. There is someone wandering here that really knows about phonetics.

So, I want to see a preprint. Perhaps I can read it, perhaps someone else can read it, perhaps we have to wait a few days until someone writes a nice blog post and debunks it, perhaps it's correct.