Having done the 50k L version: foam control alone eats the margin they are projecting.

The review actually proposes three. Section 4. They are weak, but they exist.

Counterpoint: undocumented data compounds confusion. The README here is genuinely good, which is why it works.

Started my lab entirely on auction equipment. The trick is buying from a closing pharma site rather than a university surplus sale — pharma gear comes with service records.

Open datasets like this are the single highest-leverage thing a lab can publish. Papers age, data compounds.

I would like to see the word "intelligence" retired from this entire subfield until someone proposes a falsifiable behavioural test.