[gjm11]

The following is perfectly compatible with everything in the article being correct and important (though I share the suspicion that there may be some reverse-causality going on), but the author seems to be extremely fishy in some ways.

He's posting about two substantial (~10-30 pages) papers per week to the arXiv, in various areas of mathematics. He claims to have developed "Alpay Algebra: A recursive language for thought" and what he's written about it looks to me like (1) it was actually written by an LLM and (2) it's basically word/symbol salad. (He has some papers about it on the arXiv. The first isn't particularly bullshitty, but uses a great deal of formalism to say almost nothing. Later ones look like grandiose AI-written slop. And he has some web pages that are just grandiose AI-written slop.)

I repeat: none of this needs to cast any particular doubt on what he writes about loneliness. He might be an AI-driven mathematical crank who also has wise thoughts about loneliness. I might just be wrong about his being an AI-driven mathematical crank, though I'd be pretty surprised. Or, for that matter, his post about loneliness might be AI-written but none the less correct. (It's the kind of thing I would expect an LLM to be able to write something reasonable about -- though in that case it would be advisable to double-check the references.) But it doesn't inspire confidence.

[tootie]

I didn't dig in to every cite but a few seemed suspicious. One linked on digital interventions specifically said it showed no positive results. Another on pet ownership having 100% success rate was sponsored by the pet industry. I think the author is tripping over a lot of data and cherry picking. We know that loneliness can cause problems and interventions can help but this post seems like it is indeed a rambling jumble and not a cogent argument for anything in particular.