An experiment with phonetic writing

[Jorge_Stolfi]

There may not be a recipe. If there isn’t one, I’ll provide the text I used. Well, most likely it won’t, since AI gave it to me.

I ran my program on my Shennong Bencao file.  It found exactly one recipe with a similar pattern -- namely, the second word is repeated two more times, at indices [n-3] and [n-9]:

蜂子主风头除蛊毒补虚羸伤中久服令人光泽好颜色不老大黄蜂子主心腹胀满痛轻身益气土蜂子主痈肿

This entry is for Bee Larva.  The Gemini lalamo gave me this translation:
  (A)   |Bee larva:
  (A1)  |··[Nature] sweet, neutral.
  (A3)  |··[Mainly for]
  (A31) |····wind-caused headache,
  (A32) |····eliminating magic toxins,
  (A33) |····repairing depletion and emaciation,
  (A34) |····injury to digestive system.
  (A4)  |··[Prolonged consumption]
  (A41) |····makes one's skin lustrous,
  (A42) |····improves complexion,
  (A43) |····prevents aging.
  (B)   |Giant hornet larva:
  (B3)  |··[Mainly for]
  (B31) |····distention, fullness, and pain in torso,
  (B32) |····lightens the body,
  (B33) |····boosts the vital energy
  ©   |Ground-nesting wasp larva:
  (C3)  |··[Mainly for]
  (C31) |····carbuncles,
  (C32) |····abscesses.

That entry apparently is lacking the usual final fields [Another name] and [Provenance].  Si that repeated character is not 生 ("provenance", saang¹ in Cantonese) but 主 ("mainly for", zyu² in Cantonese, daiin in Voynichese...)

According to the Gemini, the Cantonese reading of that recipe is

fung1 zi2 zyu2 fung1 tau4 coi4 gu2 duk6 bou2 heoi1 leoi4 soeng1 zung1 gau2 fuk6 ling6 jan4 gwong1 zaak6 hou2 ngaan4 sik1 bat1 lou5 hing1 san1 jik1 hei3 tou2 fung1 zi2 zyu2 jung1 zung2  

But of course "AI responses may include mistakes. Learn more".  And indeed there is something wrong with that readin, hang on...

All the best, --stolfi