There's a long, detailed, often repeated answer to your open question in the article.
replyNamely, if you can't do it without the AI, you can't tell when it's given you plausible sounding bullshit.
So Bob just wasted everyone's time and money.
You can verify by running the code and see if it works.
replySeriously? The article is about scientists learning to do science, not programming.
replyI know we're not supposed to say RTFA, but your comment really takes the cake.
My mistake then. There's so much AI discussion here on HN from the developer perspective, that it's easy to get lost.
replyBut why wouldn't my comment apply to all science? You have to make experiments and gather evidence, otherwise it's not science, it's just academia.