CA23 comments on Introducing the WeirdML Benchmark