The author started off by relexifying Toki Pona (substituting the vocabulary) and then modified the grammar to make it less ambiguous, ending up with a new language, which is given the ungooglable name "Mini". The relexification makes the phonology rather more complex: 19 instead of 14 phonemes.

I feel that if you're going to have a minimal vocabulary of about 125 words then it's good to have a minimal phonology, too, and I rather like Toki Pona's simple phonology. In any case, that's rather subjective and independent of the more interesting question of the grammar. Something like Toki Pona but with manageable syntactic ambiguity sounds like an interesting idea.

I feel that ideally all syntactic ambiguity would be explicit and obvious: stuff like where a prepositional phrase is attached rather than words whose part of speech may vary. I've not looked at "Mini" hard enough to see if that has been achieved. A list of what kinds of ambiguity remain in the language would be interesting.

(It's amazing how slow the world of invented languages moves compared with programming languages, say. It seems to have taken about 19 years for this arguably fairly natural further development of Toki Pona to appear. Rust appeared 5 years after Toki Pona and is now one of the most popular programming languages. That's not a criticism or even a sensible comparison; perhaps the more interesting observation is that there is a kind of progress in the world of invented languages: people aren't just reinventing the wheel.)