Which brings up the question of writing, and reading, more generally. In short, I encounter people who cannot read what is written and understand it in such a way as to describe or explain without violating actual words written (_said_ in some cases). As a teacher I have found this, and as a consumer of the words others put out, this I find is almost universal.
As a curious sort (junkie?), I am disappointed if I try to find intelligent authorities with the goods I am after. YouTube was YouTube "university" for me a few years ago. Now it is slop (ref. House of El, no nonsense, sophisticated presentation, plus my own experience) and enshitified (ref. Doctorow, Cory, quite insightful this his coinage) hardly worth a go-to source for the true, beautiful, and good, i.e., useful. What's needed is curation, and that will not happen given the current zeitgeist.
Many say AI in one form or another is to blame. I say yes _But_. If one uses the current tools at hand, one can help craft the true, beautiful, and good--er useful. But having accomplished this, I believe, multiple times over the past several years (inter-linear.com), I believe I am among a minority of amateur technophiles working for the common good.
My latest foray (yesterday) resulted in a surprise and concomitant dilemma. Here is a sketch of that experience.
First question to ask for what AI could produce to reduce the pressure of my curiosity: "There are physical and biological effects of a singular orgasm on the male and the female. What are these and their contributions to human health?"
The second question, which the first response did not actually answer: "Anecdotal note: I find as a senior male that after ejaculation urine flow is stronger and evacuation more complete and quicker. Is there research about this hypothesis?"
Answer: Confirmed. I am right about that, without going into details personal or scientific.
I asked, with AI's help, a follow-up question to the first above, and here is where the surprise came, downloaded to my computer a few moments later. The question: Research and compare the specific prophylactic health benefits associated with regular orgasm in males versus females, specifically documenting the impact on prostate cancer risk in males and dysmenorrhea and vaginal atrophy in females. Include data on changes in sperm morphology, leukocyte counts, and menstrual cycle regulation, creating a structured comparative report that highlights these sex-specific physiological and preventative maintenance roles of orgasmic activity, based on clinical research findings." (I obviously had help beyond my pay grade in the wording of this.)
I got an artifact, I will call it, a PDF beautifully formatted and written for an informed but not specialist reader, entitled "Sex-Specific Preventive Associations of Ejaculation, Orgasm, and Sexual Activity: A Comparative Report."
Wow. Or WOW!
Here is the _But_: What is this document? and would it be useful if shared (other than writing here concerning its provenance)? It has some value, as a kind of academic/scientific report. There are 171 sources cited as relevant to its validity/credibility.
The rules for citation of such a work are evolving, and different publishers require and recommend slightly different styles and disclaimers. Here is a key point about the document I got from the AI engine.
_A key consideration from multiple university library guides: the report itself already cites its primary clinical sources (PubMed, PMC, Harvard Health, etc.) with full URLs. Those primary sources are what readers should cite for any specific claim. The AI citation above attributes the synthesis and compilation work — the research, structuring, and drafting — to Perplexity, while the underlying evidence remains attributable to the original peer-reviewed studies._
So the actual citation for this document is the copy provider's choice among the professional style manuals, and the disclaimer, something like, "This report was researched, compiled, and generated by Perplexity Computer (AI agent), Perplexity AI, August 2026. Primary clinical sources are cited inline throughout". This needs to be up front or otherwise clearly transparent provided along with the document itself.
Tentative conclusions. One, what amazing work can be produced given carefully worded prompts or questions to reputable AI solutions (e.g., Perplexity). Two, it seems a kind of new genre(?) for grappling with knowledge bases and producing and including in scholarly work efficiently summarized states of the art.
A copy of the PDF is here if interested.
